ReviewJanuary 28, 2026

Graph Neural Networks for Polymer Characterization and Property Prediction: Opportunities and Challenges
Click to copy article linkArticle link copied!

Hector Medina*
Hector Medina
School of Engineering, Liberty University, Lynchburg, Virginia 24515, United States
*Email: [email protected]
More by Hector Medina
https://orcid.org/0000-0003-1014-2275
Rachel Drake
Rachel Drake
School of Engineering, Liberty University, Lynchburg, Virginia 24515, United States
More by Rachel Drake

Open PDF

Journal of Chemical Information and Modeling

Cite this: J. Chem. Inf. Model. 2026, 66, 3, 1316–1336

Click to copy citationCitation copied!

https://doi.org/10.1021/acs.jcim.5c02421

Published January 28, 2026

CC-BY 4.0 .

Abstract

Click to copy section linkSection link copied!

Using machine learning to accelerate the characterization and prediction of properties of many-molecule systems, such as polymers, is appealing, yet challenging. Polymers are large, complex molecules that have unique properties and potential applications in a wide range of industries. Their potential in advancing fields such as ion-transport polymer for energy storage, lightweighting of structural materials, bioinspired multifunctional materials, etc., provide enough impetus for accelerating the discovery of novel polymeric materials. However, mathematical mapping and the consequent manipulation of polymer structures are still challenging tasks due to their complex configuration and the smorgasbord of motifs encountered naturally and in engineering materials. Traditional methods of polymer structure mapping and property prediction at multiscale domains can include approaches such as Density Functional Theory, Molecular Dynamics, and Finite Element Analysis, which can be time-consuming and computationally expensive. The promise of machine learning to accelerate these tasks is appealing, and currently, researchers are pursuing the development of architectures and composition approaches to accomplish this. Here we discuss the current state of the knowledge on the use of Graph Neural Networks, and related architectures, being developed and/or used for the characterization and prediction of properties of polymers. Many challenges still exist such as the lack of sufficient and comprehensive data sets. To address these issues, efforts are being pursued─such as the so-called CRIPT (Community Resource for Innovation in Polymer Technology) led by a lab consortium that includes representations from private industry, academia, government, and others. We conclude that even though this field is young it has both momentum and promise. The current challenges that must be overcome are also addressed.

This publication is licensed under

CC-BY 4.0 .

License Summary*
You are free to share(copy and redistribute) this article in any medium or format and to adapt(remix, transform, and build upon) the material for any purpose, even commercially within the parameters below:
- Creative Commons (CC): This is a Creative Commons license.
- Attribution (BY): Credit must be given to the creator.
View full license
*Disclaimer
This summary highlights only some of the key features and terms of the actual license. It is not a license and has no legal value. Carefully review the actual license before using these materials.
License Summary*
You are free to share(copy and redistribute) this article in any medium or format and to adapt(remix, transform, and build upon) the material for any purpose, even commercially within the parameters below:
- Creative Commons (CC): This is a Creative Commons license.
- Attribution (BY): Credit must be given to the creator.
View full license
*Disclaimer
This summary highlights only some of the key features and terms of the actual license. It is not a license and has no legal value. Carefully review the actual license before using these materials.

Subjects

Keywords

1. Introduction

Click to copy section linkSection link copied!

Polymers─topologically complex compositions of chains of repeating molecular units─play an important role in materials, electronics, and biomedical devices due to their unique properties and versatility. Understanding and predicting their broad chemical structures is vital for the creation of applications in energy storage, lightweight structural materials, and bioinspired multifunctional materials. (1−13) However, the complexity of polymers with repeat units, diverse topologies, and multiscale interactions presents challenges for property prediction and material design. (14−19) Traditional methods such as Density Functional Theory (DFT), Molecular Dynamics (MD), and Finite Element Analysis (FEA) have been used for polymer characterization. (20−27) These are computational approaches ranging from first-principle methods like DFT to atomistic simulations such as MD and continuum scale techniques like FEA to characterize polymer behavior. DFT offers information on electronic structures; (28−30) however, it is computationally demanding for large polymer systems. (31−35) MD simulations give thorough dynamics behavior, but are limited by system size and simulation times. (18,36−38) FEA excels at modeling the mechanical behavior of materials, but struggles with the geometrical intricacy of polymers. (39−41) Although accurate, these methods face scalability issues due to computational requirements to model large polymer systems. (42−44)

Machine learning (ML) is a promising alternative for polymer characterization using data-driven insights to predict properties. (45−47) Random forests (RF), (48−50) support vector machines (SVM), (51−54) and artificial neural networks (ANN) (55−58) are some techniques that have shown promise for the identification of patterns and relationships in polymer data. (59−63) However, these methods often require tedious feature engineering and may not fully encode the relational or hierarchical nature of polymer structures. (64−67)

Graph Neural Networks (GNNs) are a step forward in polymer science using ML. Rather than replacing physics based methods such as DFT, MD, or FEA, GNNs function as data driven surrogate models trained on their outputs, extending these simulations to larger chemical spaces. GNNs can naturally handle the graph-like structures of polymers by representing molecular structures as graphs with atoms as nodes and bonds as edges. (68−71) GNNs extend traditional neural networks (NNs) and convolutional neural networks (CNNs) to operate on graph-structured data that can learn the local neighborhood information on each node through message-passing mechanisms. (72−77) Graph-based inputs allow these models to capture structural patterns across multiple scales, from monomer chemistry to polymer-chain architecture, providing a unified way to learn relationships embedded within the topology. In contrast to traditional ML approaches, GNNs are invariant to atom ordering and can handle the complex connectivity of polymer chains. (69,78,79) GNNs have shown promise in many areas, including social network analysis, bioinformatics, and materials science, since they can capture complex structural information in graphs. (80−89) GNN architectures such as Graph Convolutional Networks (GCNs), Message Passing Neural Networks (MPNNs), and Graph Attention Networks (GATs) have been modified to improve polymer property prediction. (90−97) (We provide a somewhat comprehensive overview of the GNN architectures in Section 2.)

The tasks of modeling polymer structures and predicting their properties using GNN are not without challenges. One of them is the lack of suitable data sets that match the smorgasbord variety of polymer’s chemistry, structure, and properties. (98−100) Furthermore, even if data mining is carried out to consolidate information from various sources, other issues arise, such variation of levels of accuracy, distinct experimental setups, various synthesis methods, etc., from different data sets. (101) Projects such as the Community Resource for Innovation in Polymer Technology (CRIPT) are attempting to overcome some of those challenges. (In Section 3, we provide a summary of some existing data sets along with some remaining challenges.) A further unresolved challenge in polymer informatics is that most current GNN models operate solely on chemical structure and largely neglect morphology, processing history, and environmental conditions, factors that critically shape real polymer behavior. Highlighting this gap helps motivate the need for future GNN architectures capable of encoding multiscale and processing-dependent information.

Keys to overcoming challenges related to the proper multiscale representation of polymers is the proper composition─and corresponding embedding─of groups of atoms into larger “pseudo components”, a concept referred to as “coarse graining” (CG). (102−105) More recent advances in polymer-specific GNN architectures have incorporated coarse-grained graph representations to connect atomistic level information with macroscopic polymer properties. (106−108) Several challenges still remain with CG, especially related to the trade-off of accuracy versus computational cost. A nonexhaustive list of CG models with their corresponding key features are summarized in Section 4.

Applications of GNNs in polymer research span a wide range of property prediction tasks. These include fundamental thermophysical properties such as glass transition temperature, melting point, solubility, and density, as well as performance-relevant attributes like dielectric constant, mechanical strength, permeability, and viscosity. (109−111) Beyond bulk properties, GNNs have been applied in specialized domains such as gas separation membranes, polymer electrolytes for batteries, and functional polymers for drug delivery or organic electronics. (106) Importantly, GNN-based models have demonstrated the ability to generalize across diverse polymer classes, learning transferable chemical representations that outperform conventional machine learning approaches on large databases. Some advances in applying GNNs to polymer property prediction are also reviewed here, both their strengths and limitations. Specific polymer property prediction tasks and associated challenges are discussed in Section 5.

Finally, we provide, under the Discussion and Overlook section, an overview of some interpretability methods used and highlight the key innovations that define emerging trends and future opportunities in the general area of polymer informatics. While several existing reviews discuss ML for polymers or GNNs for molecules and materials more broadly, fewer works explicitly examine how polymer-specific graph representations, data limitations, and multiscale considerations shape the use of GNNs in polymer informatics. This review synthesizes polymer-focused GNN works by overviewing representation choices (repeat-unit/periodic and coarse-grained graphs), data set and infrastructure constraints, and application-driven modeling across key polymer property domains, while highlighting open challenges in data quality/coverage and in capturing morphology and processing effects. To our knowledge, no other review has focused entirely on GNN-based polymer informatics, as outlined.

2. GNN Architectures

Click to copy section linkSection link copied!

In this section, we review some of the common GNN architectures used to model complex polymer systems. The models are grouped according to their core architectural classes, MPNNs, GCNs, and GATs. Each class reflects different strategies for encoding graph input, message propagation, and handling polymer-specific topologies. Table 1 provides an overview of representative GNN types and models, including their training data, polymer families, and target properties. Compared with traditional neural networks that operate on fixed-size vectors, GNNs naturally process variable graph structures and therefore support edge-, node-, and graph-level prediction tasks. (68,69,87,112) Unlike traditional NNs that process fixed-size vectors, GNNs can handle data with diverse structures, making them suitable for edge, node, and graph prediction. (72,113−115) Across these architectures, message passing layers update node and edge features based on local neighborhoods, while readout layers aggregate node representations into a graph-level embedding suitable for property prediction. (87,116,117) Figure 1 provides a schematic overview of the basic GNN architectures, illustrating their shared message-passing and readout components. GNNs can capture features such as repeat units, ring structures, and periodic graph patterns supporting improved generalization in the prediction of polymer properties.

Figure 1. General architecture of a GNN for molecular property prediction. The input molecular graph is processed through multiple message passing layers, where the node embeddings are iteratively updated by propagating information from neighboring atoms. These updated representations are then combined by a readout function (e.g., average or sum) to produce a graph-level embedding, which is fed into a downstream prediction layer.

Table 1. GNN Architectures

model	type	training data	polymer family	target properties
PU-gn-exp (106)	MPNN	697 polymers, collected from literature	conjugated semiconducting polymers	mobility, HOMO, LUMO
wD-MPNN (121)	MPNN	42,966 copolymers, computational (xTB/DFT)	conjugated copolymers	IP, EA
GH-GNN (124)	MPNN	2500–2800 polymer–solvent pairs, experimental	homopolymer–solvent mixtures	activity coefficient at infinite dilution
polyGNN (126)	MPNN	13,388 polymers, computational and experimental	diverse synthetic polymers	multiproperty: electronic, optical, thermal, mechanical, solubility
St. John MPNN (127)	MPNN	91,000 molecules and oligomers, computational (DFT)	OPV candidates	HOMO, LUMO, excitation energy
Periodic D-MPNN (123)	MPNN	15,219 data points, experimental and DFT	homopolymers	atomization energy, bandgap, EA, dielectric, T_g
GNN-A/B	MPNN	372 experimental polyimides (8,205,096 virtual candidates screened)	polyimides	T_g
sGNN (108)	MPNN	20,000 MP2 conformations, computational	PEG, PE fragments	bonding potential energy
Multitask GNN (147)	GCN	876 polymers (5 ns MD), 117 polymers (50 ns MD)	polymer electrolytes	ionic conductivity, diffusivity
Chem-DAGNN (148)	GCN	687 polyimides, experimental	polyimides	T_g
Park GCN (149)	GCN	2687 Organic Polyamides	polyamides	T_g, T_m, density, elastic modulus
Hu GCN (133)	GCN	300–600 polymers, experimental	homopolymers	T_g
Hickey GCN (134)	GCN	7558 polymers, experimental	general polymers	T_g
Zeng GCNN (110)	GCN	1073 polymers, computational (DFT)	organic polymers	dielectric constant, bandgap
SweetNet (150)	GCN	19,775 glycans, experimental	glycans	taxonomy, immunogenicity, pathogenicity, viral glycan bonding
Kimmig GCNN (135)	GCN	2813 nanoparticle measurements, experimental	poly(methacrylates)	nanoparticle size
Volgin GCNN (151)	GCN	6,726,950 synthetic (QSPR) and 214 experimental polyimides	polyimides	T_g
GATBoost (144)	GAT	235 polymers, experimental	acid containing polymers	T_g
POLYMERGNN (111)	GAT	296 polymers, experimental	polyesters	T_g, intrinsic viscosity
GNNs (107)	MPNN, GAT, GCN	1313 glycans, 15,778 peptides, experimental	glycans, peptides	immunogenicity, taxonomy, minimum inhibitory concentration

This section outlines the core computational structure and polymer-specific adaptations of GNN models, with performance results deferred to a later section. For each class, we organize the contributions by their input representation, GNN design choices, and output strategies.

2.1. Message Passing Neural Networks

Message Passing Neural Networks (MPNNs) operate on molecular graphs by iteratively exchanging information between nodes based on both node and edge features. (91,118−120) Given a graph G = (V, E), where V is the set of nodes and E is the set of edges, MPNNs proceed in two phases: a message-passing phase and a readout phase. During the message passing phase each node v has a feature vector h_v^t in iteration t. For each neighbor w ∈ N(v), the edge message from w to v is computed by

m_{v, w}^{t} = M^{t} (h_{w}^{t}, e_{v, w})

(1)

then all incoming messages are aggregated at v via

{\tilde{m}}_{v}^{t} = \sum_{w \in N (v)} m_{v, w}^{t}

(2)

and v’s hidden state is updated

h_{v}^{t + 1} = U^{t} (h_{v}^{t}, {\tilde{m}}_{v}^{t})

(3)

A graph-level feature vector is obtained by applying a readout (e.g., summation or averaging) over h_vv∈V^T, for T iterations. All learnable parameters in functions M^t, update functions U^t, and the final readout stage are trained from the beginning to the end to minimize loss in a set of labeled graphs. MPNNs serve as a flexible backbone in polymer prediction tasks due to their customizable message and update functions, which are useful in capturing polymer-specific periodicity, stoichiometry, and topological motifs.

The following models demonstrate how the MPNN framework has been extended to predict a wide range of polymer properties including glass transition temperature (T_g), ionic conductivity, diffusivity, electronic structure descriptors, activity coefficients, and force field energy terms. Although detailed benchmarking and interpretability results are deferred to later sections, this section emphasizes architecture design and graph inputs.

MPNNs are adapted in many ways to handle the structural complexity inherent to polymer systems. Several approaches incorporate directed and weighted propagation, where weights encode the frequency or importance of connections within input graphs. (121−123) MPNNs have also been extended to multigraph systems, allowing interactions between multiple molecular components. (124,125) In these architectures, each graph is initialized with its own node, edge, and global attributes. Separate message-passing layers are first applied to each graph, and their intermediate embeddings are pooled and passed to a mixture-level GNN, which demonstrates how hierarchical message passing scales naturally to multicomponent systems. Periodicity, localized subgraphs, and repeat units are also addressed by constructing periodic graph inputs and adding virtual edges between terminal nodes, enabling the MPNN to preserve local environments, as demonstrated in Figure 2. (108,126) Multitask is further implemented by concatenating the pooled embeddings with a task selector vector, one-hot identifiers that condition a shared multilayer perceptron (MLP) to produce multiple outputs from a single model. CG graph representations use higher-level structural subunits rather than individual nodes using preprocessing to extract and store these units. (106) Each subunit is encoded as a lower-dimensional node embedding, and the resulting graph is passed through an MPNN based on the gn-exp framework. (125) Additional architectural refinements focus on node update mechanisms such as replacing simple feed-forward updated with GRU-based recurrent updates, that enables iterative refinement of node states during message passing in which information accumulates across multiple message-passing steps. (127−129) Finally, physics guided adaptations have been developed. One approach constructs localized bond-centered subgraphs defined over internal coordinates, enabling the model to decouple bonding energies from nonbonded interactions and scale linearly with system size. (108)

Figure 2. Two graph augmentations in polymer GNN architectures: a virtual edges connecting distant nodes to encode periodic or long-range interactions (left), and a global node connected to all nodes to capture system-level or graph-wide information (right).

MPNN-based architectures have been adapted to retrieve polymer-specific features such as periodicity, stoichiometry, and chain-level structure. These adaptations include the use of directed and weighted edges, repeat-unit graphs, virtual periodic bonds, PU inputs, and message-passing schemes optimized for high-throughput 2D inputs. Some models also introduce input augmentation for translation invariance of repeat units and support multitask learning via property-specific conditioning. Though the basic mechanism of message-passing remains unchanged, these enhancements demonstrate the adaptability of MPNNs to model local chemical properties as well as global polymer topology.

2.2. Graph Convolutional Networks (GCN)

Graph Convolutional Networks (GCNs) are NNs that operate on graph-structured inputs. (90,130−132) Given a graph G = (V, E), a GCN learns the representations of nodes by aggregating information from the node’s neighbors. Let Ã = A + I be the adjacency matrix of the graph G with added self-loops, and let D̃ be the corresponding degree matrix. Let X be the initial input node feature matrix. The first layer of a GCN is represented as

H^{(1)} = σ ({\tilde{D}}^{- 1 / 2} \tilde{A} {\tilde{D}}^{- 1 / 2} XW)

(4)

Here, W is a learnable weight matrix, and σ is a nonlinear activation function. The output of the k-th layer output of the GCN is given by

H^{(k)} = σ ({\tilde{D}}^{- 1 / 2} \tilde{A} {\tilde{D}}^{- 1 / 2} H^{(k - 1)} W^{(k)})

(5)

For K layers, node embeddings are aggregated using a readout function to produce a graph-level feature vector. All parameters are trained from start to finish, minimizing task-specific loss on a set of labeled graphs.

GCNs are used in polymer informatics due to their efficiency and flexibility. This section groups GCN-based models by representational and architectural adaptations: monomer-level graphs, crystalline input graphs, data-augmented GCNs, attention-based variants, and transfer learning frameworks.

GCNs encode node information and consist of multiple graph convolutional layers passed through a linear regression model, a FCNN for property prediction, or fed through a global average pooling layer. (133−135) Models can be trained using stochastic gradient descent (SGD), which separates hyperparameter optimization for each property to minimize error. (110,136) Additionally, architectures can include a boom layer, a fully connected expansion layer introduced to help the model escape local minima and improve convergence (137) pass the graph level embeddings through a fully connected layers with multisample dropout scheme (138) to produce the final prediction. Further derivations of GCNs consist of two sequential blocks of gated graph convolutional layers (GCLs). The first block processes the standard molecular graph using gated recurrent units (GRUs) for node updates, (139) while edge-aware message passing is performed via a learnable transformation matrix computed by a two-layer MLP. After three GCLs, the graph is transformed into a 2-GNN format, where nodes represent atom pairs to capture second-order structural dependencies, as inspired by Grohe et al. (140) Two additional GCLs are applied to this higher-order graph. The final node embeddings are aggregated using sum pooling and passed through an MLP with ReLU activation and a linear output layer. This two-stage design allows the model to integrate both local and pairwise interactions, enhancing its ability to represent the structural complexity of subunits during pretraining and fine-tuning. Furthermore, GCNs have implemented global nodes to connect every single node in a graph via imaginary edges that participate in message passing to collect global molecular information to retain more information in the overall graph structure, as illustrated in Figure 2. (109)

GCNs provide a simple and flexible framework for polymer property prediction. Across the studies, researchers have adapted the baseline GCN architecture to accommodate data augmentation, crystalline input formats, graph attention, and higher-order message passing. These innovations allow GCN to encode complex topological structure.

2.3. GAT and Attention-Based Architectures

Graph Attention Networks (GATs) attempt to learn node embeddings via attention mechanisms. (92,141−143) Given a graph G = (V, E) with node features X, GATs compute node embeddings h_i via attention-weighted aggregation over each node’s neighborhood

h_{i} = σ (\sum_{j \in N_{i}} α_{ij} W x_{j})

(6)

where W is a learnable weight matrix, σ is an activation function, N_i is the neighborhood of node i, and α_ij is the attention coefficient computed as

α_{ij} = \frac{\exp (LeakyReLU (a^{T} [W x_{i} ∥ W x_{j}]))}{\sum_{k \in N_{i}} \exp (LeakyReLU (a^{T} [W x_{i} ∥ W x_{k}]))}

(7)

GATs generate node-level embeddings {h_i}, which can be used for downstream tasks such as node classification. For graph-level prediction, a readout function such as a summation, mean, or attention-based pooling over the set of node embeddings before passing them through an output layer (Figure 3).

Figure 3. An illustration of GAT operations. The left portion of the diagram shows how attention coefficients α_ij are computed: the transformed node features Wh_i and Wh_j are concatenated and passed through a shared attention mechanism parametrized by the vector a, followed by a LeakyReLU nonlinearity and a softmax normalization over all neighbors j ∈ N_i. The resulting coefficient α_ij encodes the learned importance of neighbor j when updating node i. The right portion illustrates multihead attention for a specific node (node 1 in the drawing): each neighbor contributes messages through multiple attention heads (depicted by the green, blue and purple lines), producing head-specific weights α_1j^(k) and transformed messages W^(k)h_j. The index k refers to the attention head number, where each head has its own learnable weight matrix W^(k) and produces its own attention score. These per-head aggregations are combined via concatenation (for intermediate layers) or averaging (for final layers) to yield the updated node embedding h_i. The diagram highlights both key components of GATs: attention-based neighborhood weighting and multihead message aggregation.

The attention mechanisms weigh neighbor contribution during message passing. An attention mechanism is a neural-network operation that allows the model to assign different importance weights to inputs (or neighboring nodes, in graph settings) when aggregating information. In the context of GNNs, this means that each node learns to focus more on its most relevant neighbors, improving representation flexibility and interpretability compared with uniform aggregation schemes.

Attention mechanisms have been incorporated in polymer graph models in several ways. One applies a GAT-based encoder coupled with a boosted decision-tree regressor. In this setup, the input is represented as a molecular graph and processed using a multihead graph attention layer. (144) Multihead attention weights highlight the relative importance of neighboring nodes and refine local feature aggregation. The GAT serves as a feature extractor, the resulting graph-level embedding obtained through standard pooling over the node states is then passed to an XGBoost regressor, (145) which performs supervised prediction using an ensemble of gradient-boosted trees. This hybrid design leverages the representational capacity of attention-based GNNs while benefiting from the speed and tabular regression accuracy of boosted tree models. Another attention-augmented architecture extend this idea to structured, multi-input graph settings. In this model, each component of the system is represented as a separate graph, and a dedicated GNN encoder processes each graph independently. (111) The encoders combine a GAT layer with a GraphSAGE layer, (80,146) followed by a self-attention pooling mechanism that outputs a graph-level embedding for each input graph. Sets of related graphs are pooled to produce fixed-size representations, which are then concatenated with additional sample-level attributes to form a unified embedding. This embedding is processed through a fully connected layer to generate a central representation that feeds into one of several output heads, enabling single-task or multitask configurations. During multitask learning, weighted loss functions balance differences in scale across outputs. These architectures are designed to be permutation-invariant to the ordering of input graphs and robust to missing information, making them suitable for screening large combinatorial design spaces.

Although fewer in number, GAT-based architectures demonstrate strong potential for polymer property prediction through adaptive, attention-driven message passing. These models offer dynamic weighting of neighboring features, which enhances representational capacity in chemically diverse and topologically complex graphs. Both GATBoost and POLYMERGNN illustrate how attention mechanisms improve generalization in polymer discovery workflows.

3. Data Sets

Click to copy section linkSection link copied!

Polymer data resources range from traditional handbooks to modern ML-oriented databases. Sources such as PolyInfo and Polymer Handbook have validated properties in thousands of polymers, but are not inherently ML readable or designed for the metadata integration required for GNNs. (152−154) Recent initiatives such as PI1M and the community-driven CRIPT present structured data sets for predictive models. (155,156) These efforts are a step forward; however the scale and chemical diversity of polymer data sets remains a critical challenge for GNN applications. ML models require a large variety and size of training data. (63,157) The need for diverse chemistries, accurate structure–property mappings and standardized experimental data further compounds these challenges. Despite the gradual increase in resources, both data scarcity and complexity in data mining in heterogeneous data sets such as OMol25 remain critical bottlenecks for applying GNNs in polymer informatics. (158)

The table below highlights some of the databases available that are used for ML models. First there is a lot of literature that contains information on polymers but here we only highlight two. There is an extensive list of books that contains polymer information, but many are difficult to mine. The Handbook of Polymers contains polymer data on major plastics and other branches in the chemical industry. The data needs to be mined from the book, which contains over 200 types of polymers for research and development. (100) The Polymer Data Handbook is another book that contains 217 polymers across many modern polymer applications. (159)

Online resources provide easier access to search and find polymer structure-properties, but some are more built for ML models than others. Here we note some of the most common data sets. PubChem is a free resource for searching molecular structures and properties with literature citations. A portion of their data set includes polymers that can be found using their specific names. (98) MatWeb is another searchable database containing material properties that include data sheets for polymers such as polyester, polycarbonate, polyethelene, and polypropylene. The properties need to be extracted from the platform and reformatted to fit the needs of ML models. (99) RadonPy is an open source polymer database, developed to process all-atom MD simulations, and 15 properties were calculated for over 1000 amorphous polymers and this data is freely available and in a ML readable form. (160) PolyInfo provides data for polymeric material design where most of their information is sources from literature on polymers. Their data set contains over 20,000 polymer samples spanning 100 different properties in thermal electrical and mechanical properties across diverse numbers of homopolymers, copolymers, and literature data. Their data set allows for polymer searches, and is free to access but prohibits mass downloading of data. (152) Polymers: A Property Database provides data for properties and manufacturing processes for over 1000 synthetic and natural polymers. It is a book that requires data mining before implementation in ML studies. (161) P1IM is a freely accessible benchmark data set containing over 1 million polymers for ML researchers, the model is trained from 12,000 polymers collected from PolyInfo and then trained on a ML generative model for a larger, hypothetical polymer data set to serve as a large benchmark for ML models. (155) A Polymer Data set for Property Prediction is mainly focused on polyesters, containing 1073 polymers developed from first principle calculations. (162) CRIPT, the Community Resource for Innovation in Polymer Technology, is a new but growing platform built to easily store and search for polymer materials based on their chemistry and capture relationships between the materials, processes, and data to be used for use in machine learning. (156) It is a community driven effort for polymeric materials accessibility and designed to be open access for all ML researchers to allow for easy data sharing, and integration through academia, industry and government.

Although specialized literature references such as the Handbook of Polymers and the Polymer Data Handbook provide detailed resources of polymers and the processes that created them, they require manual extraction. Online resources are more searchable, but many are not directly designed for ML research. Table 2 summarizes existing polymer databases and data resources, highlighting the current landscape of available data sets. However, the development of polymer data sets goes beyond open access and format. Polymers are structurally complex, and their representation can differ across platforms which makes cross comparisons difficult. Another issue is imbalance in polymer data sets commonly including industrial polymers, and novel polymers making up a very small portion of the data sets. This leads to an under representations of the diversity needed in a large data set for effective ML learning. Additionally, experimental conditions and property measurements are not standardized in most instances, which can lead to inconsistencies in ML models, and a lack of standardization in validation of data sets can also lead to that same challenge. In order to take the next steps in ML models for polymers, much larger, accessible, and standardized models are needed.

Figure 4. Conceptual depiction of CG in polymer informatics. The top bar represents the (abstract) spectrum of CG strategies from macroscopic, geometry-based top-down strategies to atomistic, energetics-based bottom-up approaches. The molecular structure is overlaid with color-coded regions representing multiple levels of abstraction: atoms (gray), monomer/repeat units (red), substructures, functional groups or motifs (blue), small domains (purple), and large domains (yellow). This hierarchy also illustrates the conceptual transition from atomistic to increasingly coarse-grained graph representations, by grouping chemically or structurally important substructures so that scalable and interpretable ML pipelines are achievable.

Table 2. Polymer Databases and Data Resources

database	types of polymers	size/scope
Handbook of Polymers (100)	polymers used by plastics, electronics, and medical fields	over 200 different types of polymers
Polymer Data Handbook (159)	data on physical properties of polymers	217 polymers
PubChem (98)	wide range of polymers and properties	millions of chemical structures, only a portion are polymers
MatWeb (99)	wide range of material properties, including polymers	185,000 materials
RadonPy (160)	15 properties	over 1000 amorphous polymers
PolyInfo (152)	various data for polymeric material design	over 20,000 polymers
Polymers: A Property Database (161)	conducting polymers, hydrogels, nanopolymers and biomaterials	over 1000 polymers
PI1M (155)	polymers for tasks in density, glass transition temperature, melting temperature, and dielectric constants	1 million generated polymers
Polymer Data set for Property Prediction (162)	for the design of high dielectric constant polymers	1073 polymers
CRIPT (156)	polymer data structures stored as graphs containing metadata on properties	size and scope under development
Various small data sets (162−167)	atomization energy, bandgap, dielectric constant, unit cell volume	various sizes

4. Coarse-Graining in Polymer Informatics

Click to copy section linkSection link copied!

The direct use of GNN in small molecular systems, where nodes represent atoms and edges bonds have been widely utilized. (79,168,169) In polymer GNN applications, however, CG is introduced primarily as a graph-representation strategy to control graph size, information content, and learning scalability, rather than as a general molecular-simulation technique (see Figure 4). However, for large systems such as polymers, grouping atoms into higher-level pseudocomponents becomes necessary. In general, CG models in molecular science simplify the representation of complex systems by reducing the level of atomistic detail while retaining critical physical properties, a concept that maps naturally onto the abstraction of nodes and edges in polymer GNN graphs. (170−174) These models achieve this by grouping atoms into larger pseudoatoms or beads to reduce the number of degrees of freedom. (175−177) CG enables simulations of larger systems over longer time scales and is computationally more effective overall compared to fully atomistic approaches. (174,178) This scalability is beneficial for polymer systems in studying mesoscale effects such as polymer diffusion, viscoelasticity, and entanglement. (105,179,180) However, an inherent limitation of CG models is the loss of detailed atomistic information and the need for careful model parametrization. (102,180,181) From a GNN perspective, these trade-offs directly determine which chemical, topological, or mesoscale features can be learned at a given graph granularity, and which fine-scale interactions are necessarily discarded. Bead size, mapping of atomistic details and potential energy function accuracy also dictate the accuracy and reliability of the CG model. (103,182,183) Despite such limitations, CG methods are useful tools for GNNs in property prediction with scalable input representations maintaining useful structural and dynamic information. (184,185) The integration of CG models with GNNs offers a connection between molecular simulations and real polymer behavior, enabling efficient, structure-aware learning at scale. This is a promising field for polymer informatics, especially in applications in which high-throughput screening and long-time scale behavior are needed.

In polymer GNN literature, hierarchical graph representations directly parallel classical CG concepts, where graph nodes function as CG beads (e.g., monomers, repeat units, substructures, etc.), and graph edges encode the connectivity or interaction topology. The following GNN architectures implement CG in different ways, for example monomers as node graphs, bond centered subgraphs, and repeat unit graphs each correspond to specific CG choices about how atoms are grouped and which interactions are retained. Mohapatra et al. (107) introduced a CG representation where each monomer was depicted as a node and chemical bonds as edges such that similarity computation, supervised learning, and interpretability could be achieved for glycans and polymers. The approach was effectively used in classification and regression tasks with GNN models in both biological and synthetic systems, with accuracy only slightly lower than all-atom models. Wang et al. (108) employed a bond-centered subgraph GNN (sGNN), which encoded internal coordinates to describe local environments and outperformed conventional force field methods for large organic molecules. Zhang et al. (106) introduced the Polymer Unit Graph (PU Graph), which is used for polymer repeat units identified from SMILES using PURS. The model was predictive with accuracy and were interpretable for structure–property relationships. They collectively show CG based GNNs enhance scalability, domain-awareness, and explainability benefits.

There are two major bottlenecks that limit the full integration and application of CG models into polymer GNN pipelines. The first is the challenge of selecting an optimal CG resolution across diverse and complex polymer systems. The accuracy of a CG model depends on how atomic structures are mapped to CG beads, and how this mapping affects the polymer’s structural and dynamical behavior. (174,186) For some polymers, CG at the monomer scale may be sufficient, while others with complex groups, stereochemical constraints, or cross-linking architectures may require nuanced or hierarchical resolutions CG strategies. CG models are typically tailored to specific polymer chemistries, architectures, or targeted property spaces, model selection must be system dependent and often empirical. The second bottleneck is the lack of robust, high-quality databases at intermediate (mesoscale) resolutions. (187) While databases exist for macroscopic properties such as mechanical or optical behavior, and atomistic data can be obtained through high-fidelity quantum methods like density functional theory (DFT), (188) mesoscopic data sets remain sparse. (189) A comprehensive mesoscale data set would ideally include the polymer chemistry, processing conditions, and environmental variables. However, many available data sets lack reproducibility due to inconsistencies in concentration, thermal, or experimental methodology. Consequently, training and validating CG based models remains difficult due to irregular, incomplete, or underrepresented data. To address this gap experimental data reflecting chemical diversity and process history is essential to extend CG models into broader ML-driven polymer design.

Table 3 provides simply contextual background and terminology, rather than a comprehensive survey, of CG methods, categorized into top-down, bottom-up, and hybrid approaches. Top-down approaches derive coarse-grained potentials by fitting to experimental or macroscopic observables, prioritizing thermodynamic consistency and scalability. (190) Bottom-up approaches, in contrast, systematically reduce atomistic models by preserving microscopic forces or distributions, offering higher accuracy in reproducing molecular behavior at finer resolution. (191) Hybrid methods combine both philosophies-retaining atomistic detail in key regions while coarsening others, or integrating field-based and particle-based techniques. (192,193) While this is not exhaustive, this highlights classic methods, alongside newer ML techniques. These methods vary in their derivation of CG potentials or learned representations and differ in their suitability for modeling goals.

Table 3. Overview of Representative CG Models

CG model name	approach type	key features	references
Dissipative Particle Dynamics (DPD)	Top-down	particle-based CG method with soft forces and momentum conservation for mesoscale soft matter	(194−197)
Martini Model	Top-down	4-to-1 (variable) CG model tuned to reproduce experimental thermodynamic data	(198−200)
Self-Consistent Field Theory (SCFT)	Top-down	field-theoretic simulation framework for polymer phase behavior	(201,202)
TraPPE-CG	Top-down	typically for alkanes, parameters fitted to vapor–liquid coexistence	(203)
UNRES	Top-down	united-residue CG force field for protein structure and folding simulations	(204,205)
Multiscale CG (MS-CG)	Bottom-up	derives CG potentials by matching atomistic forces (force-matching)	(172)
Iterative Boltzmann Inversion (IBI)	Bottom-up	iteratively updates CG potentials to match atomistic pair distributions	(206,207)
Single-Chain Boltzmann Inversion	Bottom-up	uses ab initio torsions and intrachain conformational sampling	(208)
United Atom/Blobs	Bottom-up	groups atoms into beads; bonded and nonbonded interactions derived from atomistic structure	(209−211)
Slip-Spring Model	Bottom-up	represents entanglement effects via slip-link dynamics between chains	(212−214)
Bead–Rod Models	Bottom-up	captures chain stiffness using rigid or semiflexible rod-like segments	(215)
Kremer-Grest Model	Bottom-up	idealized bead–spring model for simulating entangled polymer melts	(216−219)
Adaptive Resolution Scheme (AdResS)	Hybrid	couples atomistic and coarse-grained regions dynamically in one simulation box	(220)
SIRAH	Hybrid	CG biomolecular force field combining atomistic mapping with empirical tuning with multiscaling available	(221)
Rosetta	Hybrid	fragment-based modeling using statistical potentials and structural bioinformatics data	(175,222,223)

5. Applications and Implementations

Click to copy section linkSection link copied!

Generalizing the architectural frameworks introduced in Section 2, this section reviews the applications of the models for a variety of polymer property prediction tasks. We organize the discussion by target property domains, such as electronic, thermal, and mechanical behavior, to emphasize how various GNN styles relate to specific challenges for polymer modeling.

GNNs have been adapted to a wide variety of polymer property prediction tasks, ranging from electronic and optical properties to thermal, transport, mechanical, and some specialized chemical behaviors. In this section, these architectures are reviewed with a property-oriented focus, organized by application field rather than type of GNN. Although MPNN-based architectures still lead in most type of properties, especially in electronic and thermodynamic tasks, GCN- and GAT-based models have demonstrated competitive or even superior performance in some scenarios. Data set sizes vary widely from a few hundred to tens of thousands of polymers, with sparsity issues of experimental data typically being mitigated using augmentation, transfer learning, or multitask modeling. Model quality is assessed via typical metrics such as RMSE, MAE, or R²; architectural variants detailed earlier Section 2 are briefly described here when relevant.

This next subsection provides an overview of the tasks and predictive accuracy of the GNN approaches.

5.1. Electronic and Optical Property Prediction

GNNs have been widely applied to predict polymer electronic and optical properties, including carrier mobility, orbital energies, bandgaps, and dielectric constants. This subsection summarizes key contributions based on model architecture, graph representation, and property type. While most models rely on MPNN backbones, several variants incorporate CG representations, periodic graphs, or multitask learning strategies to better capture polymer structure–function relationships.

Zhang et al. proposed the PU-gn-exp model, a MPNN architecture designed for interpretable predictions of optoelectronic properties in organic semiconductors (OSC). (106) The model replaces atomistic input graphs with polymer-unit (PU) graphs derived from repeat unit parsing. Each PU is encoded as a node, and edges capture interunit connectivity and topology. A gn-exp message-passing network is used with Layerwise Relevance Propagation (LRP) (224) to enable structure-attribution analysis. The model was trained on a data set of 697 polymers with SMILES inputs and DFT-calculated HOMO, LUMO, and carrier mobility values. PU-gn-exp achieved 81.96 and 88.20% classification accuracy for hole and electron mobility, respectively, while reducing training time by 98% compared to atom-level MPNNs. (91) This work highlights the interpretability and efficiency benefits of polymer-unit representations.

Aldeghi and Coley evaluated the wD-MPNN model on predicting ionization potential (IP) and electron affinity (EA) on a computationally generated test set of 42,966 copolymers. (121) The data set used to probe chemical space previously explored by Bai et al. (225) involved alterations of monomer composition, chain architecture (alternating, random, block), and stoichiometry (1:1, 1:3, 3:1). The EA and IP values were initially predicted using combined extended tight binding (xTB) and DFT calculations, with Boltzmann averaging over eight conformers and several sequences, and latterly calibrated against DFT (B3LYP/DZP) computations. (226−230) Benchmarked against were Chemprop’s D-MPNN (231) with disconnected monomer graphs, and fully connected neural networks and RF on ECFP fingerprints. (232,233) The wD-MPNN with chain architecture via weighted edges and stoichiometry via scaled atomic contributions significantly surpassed these baselines, achieving very low RMSE (0.03 eV) and nearly perfect R² = 1.00 on random splits. On a challenging monomer identity decomposition, it also performed superior with RMSEs of 0.10 eV for EA and 0.09 eV for IP. An ablation study pointed out architectural and stoichiometric encodings as being crucial to model correctness. Moreover, using their model to predict diblock copolymer phase behavior from Arora et al. (50) experiment data, they achieved a precision-recall curve (PRC) of 0.68, which is higher than that of D-MPNN (0.47) but lower than the RF baseline (0.71). Stoichiometry again proved to be the most important feature.

Zeng et al. applied a GCNN to predict bandgap and dielectric properties from crystal-derived polymer graphs. (110) They used a database of 1073 polymers compiled from the Polymergenome Project, (162,234,235) including experimentally synthesized polymers, COD structures, and computationally generated polymers. CIF file structural data included input polymer graphs for the GCNNs. The dielectric constant prediction model had a small mean absolute error (MAE) of 0.24 and performed better than existing Gaussian process regression benchmarks. (235) Bandgap prediction resulted in slightly greater MAE of 0.41 eV, owing to limited high-bandgap polymers. Notably, bandgap MAE decreased systematically with increasing data set size, demonstrating GCNN scalability. Tests against standard ML baselines (Kernel Ridge Regression, RF, Gradient Boosting, and baseline neural networks) from Matminer-generated descriptors (236) confirmed GCNN’s superior performance for both properties.

Gurnani et al. investigated the multitask polyGNN, a multitask MPNN that operates on periodic repeat unit graphs for electronic and optical properties. (126) Their data set was large and included polymer crystal and isolated chain band gaps, ionization energies, electron affinities, experimental and computed refractive indices, and frequency-dependent dielectric constants taken from DFT calculations, (162) literature, (237,238) and standard references. (153) PolyGNN performed better than a PG-MLP baseline trained over Polymer Genome (PG) fingerprints at every point of comparison, (235,239) specifically achieving smaller RMSE values for electronic properties such as Eg (0.445 eV) and EA (0.380 eV). The multitask setup preserved predictive power even in low-data scenarios, though performance limitations were uncovered for certain low-data dielectric tasks, perhaps owing to limited representation of larger structural motifs and potential over smoothing after only a few message-passing iterations. (240,241)

St. John et al. developed an MPNN model to efficiently screen for organic photovoltaic (OPV) materials and forecast optoelectronic properties directly from 2D molecular connectivity without requiring computational expense of 3D DFT geometry optimization. (127) Their data set contained approximately 91,000 unique OPV-related molecules with properties computed at both monomer and polymer levels by B3LYP/6–31g(d). (242) The single-task MPNN was accurate in prediction with mean absolute errors (MAEs) of 32.1 meV (HOMO) and 27.9 meV (LUMO). Polymer-level predictions were equally accurate for HOMO, LUMO, and optical properties. Although it suggested faster predictions, no concrete runtime benchmarking was provided by the authors. At scale, this work demonstrated that 2D connectivity is sufficient to capture important electronic properties of interest in OPV design.

Antonuik et al. compared three model architectures (RF on monomer descriptors, monomer-graph MPNN, and periodic polymer graph MPNN) on ten polymer properties including atomization energy, bandgap, and electron affinity from a data set of 15,219 computational and experimental data points. (123) Their periodic graph MPNN achieved the best performance on eight tasks with an average error reduction of 11% relative to monomer-level MPNNs, and 20% relative to descriptor-based RFs. Notable improvements were observed for atomization energy (29%), chain-level bandgap (19%), dielectric constant (17%), and glass transition temperature (6%). Periodic representation was confirmed to be the critical factor enhancing predictivity, with MPNN-generated descriptors being useful only when employed within RF models. Benchmark comparison with results from Kuenneth et al. (243) further emphasized periodic MPNN’s state-of-the-art predictive accuracy for some properties.

Overall, GNN-based models have demonstrated strong performance across a range of electronic and optical prediction tasks, particularly using MPNNs with polymer-specific graph inputs. Innovations such as adjacency-free architectures, (169) PU graphs, periodic edge representations, and multitask learning have enabled accurate property prediction. Table 4 provides a structured within-domain summary of the data sets, representation granularity, evaluation protocols, limitations, and main take-home messages for the electronic and optical property studies discussed above.

Table 4. Within-Domain Performance Synthesis for Electronic and Optical Polymer Property Prediction with GNNs

study	task	polymer representation	evaluation context	key takeaway/limitation
Zhang et al. (106)	mobility, HOMO/LUMO	polymer-unit graph	random split; classification	interpretability; parsing-dependent
Aldeghi and Coley (121)	IP/EA	stoichiometry- and architecture-aware monomer graph	random splits	weighted atomic contributions
Zeng et al. (110)	bandgap, dielectric	crystal-derived polymer graph	random split; MAE	structure-limited
Gurnani et al. (126)	electronic, optical, dielectric	periodic repeat-unit graph	RMSE	multitask; low-data dielectric limits
St. John et al. (127)	OPV optoelectronics	2D connectivity graph	MAE	scalable screening; runtime not quantified
Antoniuk et al. (123)	multiproperty electronics	periodic polymer graph	relative error reduction	periodicity dominant

5.2. Thermal and Mechanical Properties

GNNs have been widely adopted for predicting thermal and mechanical properties of polymers, including glass transition temperature (T_g), melting temperature (T_m), decomposition temperature (T_d) and elastic modulus. These tasks often rely on tailored input representations, such as monomer-based molecular graphs, periodic graphs, or augmented SMILES strings. GNN architectures in this domain range from standard GCNs and MPNNs to multitask or chemically informed frameworks, consistently improving over descriptor-based models. Common strategies to overcome data scarcity include data augmentation, transfer learning, and multitask learning.

Park et al. used GCN-based models to forecast thermal and mechanical properties of polyamides like glass transition temperature (T_g), melting temperature (T_m), density (ρ), and elastic modulus (E). (149) Their data set, sourced from the PoLyInfo database, (152) consisted of experimental property data for 1,388 polymers (T_g), 942 (T_m), 390 (ρ), and 306 (E). They tested two regression models, linear regression (GCN-LR) and fully connected neural network (GCN-NN), against baselines using ECFP descriptors together with LR and NN regressors. The GCN-NN model greatly surpassed ECFP-NN in every instance, particularly doing very well on properties strongly correlated with backbone rigidity. For instance, GCN-NN achieved an extremely high R² of 0.90 (RMSE: 29.98 K) for T_g and 0.76 (RMSE: 40.37 K) for T_m, outperforming ECFP-NN in both tasks. Density predictions were comparable across models, whereas elastic modulus predictions were influenced by noisy and sparse data. Nonlinear transformation using neural networks was identified as of vital importance to all properties. Lastly, transferability of latent representations learned by GCN was demonstrated by projecting 100,000 unlabeled polymers in the PI1M data set onto the same feature space, with uniform structural features as per polymer stiffness. The authors described reduced predictive power for less highly correlated properties and constraints that came from the use of monomer-level data in their model and absence of stereochemical or 3D information.

Gurnani et al. demonstrated polyGNN, a multitask MPNN, for predicting thermal and mechanical polymer properties like decomposition temperature (T_d), melting temperature (T_m), glass transition temperature (T_g), Young’s modulus, and tensile strength. (126) Their extensive data set consisted of experimentally measured and computationally derived property values, such as T_d (2200), T_m (3275), and T_g (3690), and mechanical properties such as Young’s modulus (412) and tensile strength (466). The multitask polyGNN improved over the PG-MLP baseline consistently for thermal properties with much lower RMSE values: T_d (58.7 K vs 59.3 K), T_m (45.0 K vs 47.2 K), and T_g (31.7 K vs 34.0 K). For mechanical properties, polyGNN performed well, with RMSE of 0.827 GPa (Young’s modulus) and 23.3 MPa (tensile strength), comparable to PG-MLP. Most importantly, polyGNN was capable of achieving positive R² on low-data tasks where individual-task models would perform poorly, highlighting the benefit of multitask learning with selector vector conditioning in capturing heterogeneous polymer behaviors.

Hu et al. used a GCN model with SMILES-based data augmentation to predict polymer glass transition temperatures (T_g) using augmented SMILES strings to transcend data set limitations. (133) On two data sets (D300 and D600), they employed SMILES enumeration to augment monomer representations, predictive performance significantly enhanced. For instance, augmentation enhanced D300 predictions from RMSE 19.4 K (R²: 0.88) to RMSE 7.4 K (R²: 0.97) and D600 predictions from RMSE 31.0 K (R²: 0.78) to RMSE 15.5 K (R²: 0.94). The GCN outperformed consistently descriptor-based baselines (Random Forest, MLP, SVR) and even MPNN models to some extent, especially at elevated augmentation levels, highlighting the power of data augmentation in polymer informatics.

Qiu et al. combined a GNN model and experimental validation to discover high T_g polyimides (PIs) for demanding aerospace and display applications. (109) Eight PIs were prepared, with the model estimating experimental T_g values (seven estimates within 30C). The model applied Atomic Contribution Maps (ACMs) to interpret structure–property correlations, stating that ether bonds and symmetrical methyl groups were undesirable, but aromatic and rigid motifs enhanced T_g. Screening over 8.2 million PI candidates uncovered numerous promising materials, of which 110 have a melting point over 400 C. They also constructed polyScreen, an accessible predictor that can perform fast T_g predictions. Even though computationally costly, their work suggests the promise of interpretable deep learning methods in polymer design.

Volgin et al. applied transfer learning to a GCNN pretrained on 6.7 million synthetic polyimide units (Askadskii’s QSPR-derived T_g) to predict experimentally measured T_g values. (151) Pretraining significantly reduced prediction errors compared to purely experimental data-trained models (MAE: 22.5 K vs 28.1 K). Conversely, general molecular pretrained models (QM9 data set) performed poorly (MAE: 40.6 K), emphasizing the importance of polymer-specific synthetic data. Their findings measure the “reality gap,” suggesting a minimum of 95% synthetic data content for peak performance, consistent with comparable results across other fields.

Hickey et al. compared a GCN trained on a diverse data set (7,558 polymer T_g values) and a quantum chemistry (QC)-based regression model trained on a small, detailed data set (83 polymers)─for predicting T_g. (134) The GCN achieved strong accuracy (RMSE: 38.1C, R²: 0.90), capturing broad topological diversity, whereas the QC model (RMSE: 34.5C, R²: 0.86) provided greater interpretability at lower data scales.

Qiu et al. constructed Chem-DAGNN, a chemically aware GNN using augmented data sets for high-accuracy T_g prediction with R² = 0.95 and RMSE = 28.0C. (148) The model outperformed regular GCNs, transformers, and conventional baselines. Screening more than a million polyimides identified new candidates with experimentally verified T_g above 450C. ACM interpretability also associated structural motifs with enhanced thermal performance, demonstrating the promise of interpretable, augmented data-driven GNNs for advanced polymer discovery.

POLYMERGNN, a multitask attention-based GNN, demonstrated multitask learning robustness for T_g and intrinsic viscosity (IV) prediction and outperformed kernel ridge regression and traditional descriptors Coulomb Matrices, Smooth overlap of atomic potentials, persistence images and many-body tensor representations. (244−248) The integration of molecular weight greatly improved IV prediction accuracy. The model’s capability to accurately conduct virtual screening proves its translational potential to guide polymer design with desirable physical properties.

Li et al.’s GATBoost model combined GAT with XGBoost to predict polymer T_g and identified important structural motifs influencing thermal properties. (144) Through massive data set augmentation (25 fold via SMILES enumeration), GATBoost attained significant accuracy gains (RMSE reduced to 2.20C, R² = 0.999), surpassing other graph-based and sequence-based models. This highlighted SMILES augmentation’s broad impact as well as the interpretability of GATBoost in sparse data regimes.

In summary, thermal and mechanical property prediction with GNNs benefits from data augmentation, transfer learning, and multitask architectures. While GCN and MPNN backbones remain common in this domain, chemically aware designs and hybrid architectures (e.g., GATBoost, Chem-DAGNN) have led to both accuracy improvements and enhanced model interpretability. Low data regimes are best handled using augmented inputs or multitask conditioning, and interpretability tools are increasingly leveraged to identify functional motifs for targeted materials design. Table 5 summarizes how representation choices and data strategies govern performance in thermal and mechanical property prediction, complementing the architecture-focused discussion in Section 2.

Table 5. Within-Domain Performance Synthesis for Thermal and Mechanical Polymer Property Prediction with GNNs

study	task	polymer representation	evaluation context	key takeaway/limitation
Park et al. (149)	T_g, T_m, ρ, E	monomer-level molecular graph	random split; RMSE/R²	backbone rigidity captured
Gurnani et al. (126)	T_g, T_m, T_d, modulus, strength	periodic repeat-unit graph	RMSE	multitask learning
Hu et al. (133)	T_g	monomer graph with SMILES augmentation	Random split; RMSE	SMILES augmentation
Qiu et al. (109)	High-T_g PI discovery	monomer graph	screening and experimental validation	interpretability
Volgin et al. (151)	T_g	monomer graph	transfer vs direct training	synthetic pretraining
Hickey et al. (134)	T_g	monomer graph	random split; RMSE/R²	small-scale interpretability
Qiu et al. (148)	high-T_g	chemically aware GNN	random split; RMSE/R²	data augmentation
POLYMERGNN (244)	T_g, intrinsic viscosity	attention-based graph	random split; RMSE	multitask coupling
Li et al. (GATBoost) (144)	T_g	SMILES graph	random split; RMSE/R²	augmentation boosts accuracy; realism unclear

5.3. Transport and Dynamic Properties

GNNs have also been used in predicting transport, diffusion, and dynamic properties in polymer systems, which can include time-dependent simulation, polymer–solvent interaction, or force field modeling. These tasks span thermodynamics, local atomic transitions, and nanoparticle formulation. Models in this domain frequently combine GNNs with domain-based information, such as MD, transfer learning, or physics-informed preprocessing, to address the multiscale nature of polymer systems.

Xie et al. developed a hybrid multitask GNN which was trained on short (5 ns) and long (50 ns) molecular dynamics (MD) simulations for the prediction of ionic conductivity (σ) and diffusivity (D_Li, D_TFSI, D_Poly) in polymer electrolytes. (147) Their data set was reduced from 53,362 polymer candidates to 876 polymers (5 ns MD) and 117 polymers (50 ns MD). The multitask GNN significantly improved predictions, yielding mean absolute errors (MAEs) of 0.076log₁₀(S/cm) on interpolated and 0.182 log₁₀(S/cm) on extrapolated cases, considerably outperforming short MD baselines and linear correction. The GNN model also performed better than a random forest trained on Morgan fingerprints and exhibited outstanding generalization to experimental data (31 polymers, external MAE: 0.093–0.120 log₁₀(S/cm)). This renders the multitask GNN a reliable and effective alternative to high-throughput virtual screening of lithium-ion battery polymer electrolytes.

Sanchez Medina et al. developed GH-GNN for predicting activity coefficients at infinite dilution Ω_ij^∞ in polymer solvent systems. (124) Their data set comprised of a curated collection of Ω_ij^∞ values sourced from various experimental studies using inverse gas chromatography (IGC), encompassing data from over 48 distinct homopolymers and 150 solvents. Various polymer representations including monomers, repeating units, periodic units, and oligomers were tested to determine the impact on model accuracy. The GH-GNN outperformed an RF model achieving a MAE of 0.15–0.26 and an R² of 0.64–0.92 on the data sets for interpolation and extrapolation. With transfer learning, the GNN model was pretrained on a data set from DECHEMA used in their previous work (249) containing 40,219 data points of smaller molecular systems. The GH-GNN improved further, reducing MAE by approximately 23.5% for the Mn data set, 13.3% for the Mw data set, and another 13.3% for combined Mn/Mw data set. The transfer-learned GH-GNN achieved MAEs near 0.13 and R² above 0.94 for interpolation and 0.15–0.20 MAE and R² above 0.89 extrapolation. The GH-GNN model was compared to additional thermodynamic models such as UNIFAC-ZM, (250) Entropic-FV, (251) and other group-contribution-based methods, (252) consistently outperforming them in both interpolation and extrapolation tasks across diverse polymer–solvent systems. The model demonstrates its strength in modeling complex polymer–solvent behavior and highlights the advantages of combining GNN architectures with transfer learning for mixture property prediction.

Wang et al. evaluated the sGNN, a subgraph-based GNN, to model intramolecular potential energy surfaces for flexible polymers like PEG and PE. (108) We note that sGNN is a GNN-based ML interatomic potential (MLIP), which is fundamentally different in purpose and scope from the property-prediction GNNs discussed elsewhere in this section. MLIPs for polymers remain relatively limited compared to property-focused GNNs, and a comprehensive survey lies beyond the scope of this review. The target property for learning was the intramolecular energy contribution obtained from second-order Møller–Plesset perturbation theory (MP2), with all nonbonding interactions subtracted using a separate physics-based model. (253) The training data were derived from NVT molecular dynamics simulations using OPLS-AA on methyl-capped polyethylene glycol (PEG) and polyethylene (PE). They generated 20,000 conformations per small molecule (10,000 each at 300 and 1000 K) and used 1000 randomly selected 300 K conformations for testing. For larger polymers such as PEG[8], only 1000 conformations at 300 K were generated for testing. The sGNN trained on PEG[2] and PEG[4] achieved root-mean-square errors (RMSE) of 0.020 kcal/mol/atom on PEG[4] and PEG[8], outperforming both the classical OPLS-AA force field and the TensorMol-1.0 ML potential, which yielded RMSEs ranging from 0.054–0.24 kcal/mol/atom. They also tested an ablated version of their model trained without removing nonbonding energy, which led to RMSEs of 0.025 kcal/mol/atom on PEG[4] and 0.047 kcal/mol/atom on PEG[8], demonstrating the importance of accurate nonbonding treatment. A message-passing–free variant (“sGNN-local”) showed slightly higher error (0.024 kcal/mol/atom), confirming the role of nonlocal coupling in bonded interactions.

Kimmig et al. applied a GNN to predict nanoparticle size from polymer structure for a data set containing 3753 nanoparticle formulations. (135) The data set included a diverse range of methacrylates synthesized via controlled radical polymerization techniques, such as reversible addition–fragmentation chain transfer (RAFT) polymerization, ensuring a broad representation of possible polymer structures and sizes. High-throughput nanoprecipitation methods were used to prepare nanoparticles, and dynamic light scattering (DLS) measurements provided precise particle size data. The data was preprocessed to remove outliers and artifacts. During training, the model achieved a mean absolute percentage error (MAPE) of 5.46% ± 0.65% on the training data and 15.12% ± 5.58% on the test data, indicating robust performance across diverse polymer structures and formulation conditions. The high accuracy in predictions for previously unseen polymers underscores the model’s generalizability and potential utility in nanoparticle formulation optimization, significantly reducing the need for extensive experimental trials.

In summary, transport and dynamic property predictions benefit from hybrid GNN approaches that incorporate domain-specific data. GH-GNN leveraged transfer learning for mixture thermodynamics, while sGNN addressed force-field behavior in polymer systems. Across tasks, GNNs demonstrated strong generalization and adaptability when tailored to the chemical or temporal structure of polymers. Table 6 highlights that reliable transport and dynamic property prediction typically requires hybrid GNN pipelines that integrate simulation, physics-based preprocessing, or transfer learning.

Table 6. Within-Domain Performance Synthesis for Transport and Dynamic Polymer Property Prediction with GNNs

study	task	polymer representation	evaluation context	key takeaway/limitation
Xie et al. (147)	ionic conductivity, diffusivity	polymer graph	MAE	MD–GNN hybrid
Sanchez Medina et al. (124)	activity coefficients at infinite dilution	polymer–solvent graphs	MAE/R²	transfer learning
Wang et al. (108)	intramolecular energy (MLIP)	subgraph-based polymer graph	conformational RMSE vs force fields	near MP2 accuracy achieved; limited to intramolecular PES
Kimmig et al. (135)	nanoparticle size	polymer graph	train/test MAPE	good generalization across chemistries

5.4. Biological Properties

GNNs have also been applied to adjacent macromolecular systems beyond classical synthetic polymers, including glycan classification, antimicrobial activity modeling, and viral binding prediction. We include these examples as a brief outlook to illustrate how polymer-inspired graph representations generalize to structurally modular biological macromolecules, rather than as core polymer informatics case studies. Mohapatra et al. developed a CG GNN framework for two supervised learning tasks: classification of glycans and regression of antimicrobial peptide (AMP) activity, both chosen to reflect biologically and chemically relevant macromolecular properties. (107) For classification, they curated a data set of 1313 labeled glycans from GlycoBase, spanning immunogenicity and eight taxonomic levels. Each glycan was parsed into a graph where monomers were nodes and linkages were edges, and features were generated using either ECFP fingerprints or one-hot encodings. Five GNN architectures were benchmarked: GCN, (90) Weave, (254) MPNN, (91) GAT, (92) and AttentiveFP, (143) all implemented via DGL-LifeSci. (255) All architectures achieved ROC-AUC > 0.95 for both immunogenicity and taxonomy prediction tasks, with AttentiveFP offering the most stable and interpretable attributions. For AMP regression, they used a curated data set of 15,778 peptides from DBAASP, focusing on minimum inhibitory concentration (MIC) predictions against E. coli and S. aureus. The Weave model yielded the best results among tested GNN, their framework matched or outperformed prior methods on four out of eight glycan classification tasks, demonstrating that even with coarse-grained inputs, structured monomer-level graphs can achieve high predictive accuracy.

Burkholz et al. introduced SweetNet, a GCN based model enhanced with a boom layer and designed for glycan species and classification. (150) The species prediction task was framed as a 581-class classification problem, and SweetNet achieved 44.3% accuracy using a GCN with a boom layer, an improvement of 7.79% over the previous method SweetTalk. (256) For immunogenicity prediction, the model reached an accuracy of 94.6%, exceeding SweetTalk’s 91.7%. In pathogenicity classification, the GCN with a boom layer achieved 91.9% accuracy, outperforming SweetTalk’s 89.1%. Benchmark comparisons across domain, kingdom, phylum, class, order, family, and genus levels showed consistent improvements with SweetNet, with the largest gains seen in highly granular predictions such as genus (11.09% gain over SweetTalk). In addition to taxonomy and functionality, SweetNet was used to predict viral glycan binding intensities an application relevant to molecular recognition. Using 126,894 glycan-binding measurements from glycan arrays of influenza virus hemagglutinin, the model achieved an MSE of 0.7352, outperforming a motif-count-based fully connected network (MSE = 0.8753) and a SweetTalk-based language model (MSE = 0.8726). They were able to capture nuanced chemical features, correctly prioritizing known sialic acid motifs, showing strong generalization, including to glycans not present in the training set.

Collectively, these examples highlight the conceptual transferability of polymer-aware GNN representations to biological macromolecules. While not the focus of (synthetic) polymer informatics, they motivate cross-domain method development and help contextualize why coarse-grained, monomer-level graphs are effective abstractions for structured macromolecular systems (Table 7).

Table 7. Within-Domain Performance Synthesis for Biological and Macromolecular Property Prediction with GNNs

study	task	polymer representation	evaluation context	key takeaway/limitation
Mohapatra et al. (107)	glycan classification	coarse-grained monomer graph	ROC–AUC	CG graphs effective for biology
Burkholz et al. (150)	glycan taxonomy, immunogenicity, binding	monosaccharide graph	MSE	interpretability

GNNs have shown solid performance on a wide range of polymer prediction tasks, particularly in electronic and thermal property domains. Multitask learning, CG representations, and data augmentation have worked best at solving data sparsity and generalizability. Message-passing models work better than traditional fingerprint-based models. Transport and mechanical properties are at times more difficult to predict due to noise, sparsity, and complex dynamic behavior opening up direction for future research. This difficulty may reflect architectural limitations as the GNNs reviewed here operate on static molecular graphs and cannot encode chain dynamics or processing history which strongly governs transport and mechanical behavior in real polymer systems. Across these property domains, clear cross-study patterns emerge. T_g appears repeatedly as a successful target for GCN- and MPNN-based models because it is a scalar, widely recorded experimental property and is well-captured by 2D monomer or repeat-unit graph inputs. This leads to broadly consistent performance across data sets, regardless of architectural variations. By contrast, properties with higher experimental noise, limited data, or more complex physical dependencies (e.g., mechanical moduli or diffusivities) show greater variability in GNN performance. This indicates that data quality and task definition often dominate over architectural choice. These trends provide context on why certain architectures succeed for specific polymer properties and highlight ongoing bottlenecks in low-data and high-noise regimes.

6. Discussion and Outlook

Click to copy section linkSection link copied!

GNNs have demonstrated versatility in predicting polymer properties throughout electronic, thermal, transport, and structural regimes. Apart from prediction ability, some recent studies have added interpretability strategies to GNN models, enabling researchers to move away from black-box performance and disclose structure–property relations. This section combines interpretability methods, common findings derived from them, and overall trends and challenges encountered among the surveyed works. It also outlines developing opportunities for polymer GNNs in model building and real-world materials discovery.

6.1. Understanding GNNs: Tools and Insights

Across various works, interpretability techniques were employed to move beyond black-box prediction and provide chemical or structural insight into discovered GNN behavior. Some of these included methods such as LRP, saliency maps, integrated gradients, and atomic contribution maps, along with chemically inspired model architectures to reason predictions in terms of structural motifs, functional groups, or topological properties. (257−260)

Zhang et al. (106) employed LRP with their PU-gn-exp which substitutes atom-level graphs with more CG based repeat-unit graphs to OSCs. Visual and statistical recognition of polymer substructures most responsible for carrier mobility, HOMO, and LUMO prediction was achievable. The interpretability framework was embedded directly into the PyTorch computation graph using automatic differentiation to render attribution both scalable and chemically interpretable.

Additionally, Hu et al. (133) employed a GCN on SMILES data augmented for the prediction of T_g to identify atomic features consistent with known thermodynamic behavior with saliency maps. Rigid backbones and bulky side groups were highlighted as influential, consistent with conventional polymer design insight.

Mohapatra et al. (107) utilized integrated gradients and input × gradient methods to study glycan immunogenicity. Their monomer-level plots show that chemically relevant motifs such as fucose and xylose consistently gave rise to classification issues. Out of several GNNs, AttentiveFP yielded the most stable and interpretable results, pointing to the significance of architecture choice in attribution stability.

Qiu et al. interpreted that the enhanced GNN had learned chemically meaningful rules like favoring rigid, aromatic, hydrogen-bonding and 3,4’-isomer structures that explain and predict high-T_g, heat-resistant polyimides. (148) Li et al. (144) then extended this further by combining a GAT and an XGBoost into their GATBoost model, where attention scores distinguished significant subgraphs and the boosting algorithm ordered their importance. This two-stage architecture allowed for explainable structure–property.

Across the literature, graph granularity is strongly coupled to property type: atom-level GNNs for local electronic descriptors (HOMO/LUMO, IP/EA), while monomer- or repeat-unit–level graphs perform for long-range or topology-dependent properties such as T_g, diffusivity, or activity coefficients. Data set size also plays a central role: small experimental data sets (<1000 polymers) typically favor architectures with stronger inductive biases, whereas large computational data sets (>10⁴ polymers) allow more expressive MPNNs to achieve their full representational capacity. Although only a few polymer GNN studies use multitask learning, some patterns emerge. When the predicted properties share strong physical coupling-such as thermal transitions (T_g, T_m, T_d) in polyGNN or correlated ionic transport coefficients models improve accuracy. In contrast, when the targets differ in data quality or underlying physics, as seen in the dielectric subtasks, multitask training provides little benefit or may degrade performance. These results suggest that the choice of which tasks to combine often matters more than the specific architecture. Overall, the literature suggests that successful polymer GNNs emerge when graph granularity, data, and multitask structures match the physics of the problem, sometimes exerting a stronger influence on accuracy than from architectural changes alone.

6.2. Coarse-Graining in Graph Design

A commonality among the most transferable and explainable models is that they use chemically informed graph representations that are more advanced than the simple atomic graph. In particular, Zhang et al.’s PU-GNN, (106) Mohapatra et al.’s glycan graphs, (107) and Wang et al.’s subgraph-based sGNN (108) all used CG based graph, grouping atoms into chemically functional subunits or repeat motifs. CG representations facilitate easier explanations by aligning model structure with human intuition and domain-specific abstraction. They also offer benefits in training efficiency, generalization, and visual attribution.

This renewed focus on graph-level CG holds promise for unifying representation conventions in polymer informatics. Atomistic graphs remain dominant in small-molecule ML, but polymers offer an inherent modularity-via repeat units, side chains, block architecture, and stereochemistry-that can be fully realized through customized CG representations. These enable researchers to reason over structure–property relationships at a chemically meaningful scale, and even accelerate high-throughput screening pipelines by reducing input complexity without compromising prediction accuracy.

6.3. Outlook: from Prediction to Design

As polymer GNNs continue to mature, their application is trending away from predictive accuracy toward active design, synthesis assistance, and discovery. The studies by Qiu et al., (148) Hu et al., (133) and Li et al. (144) illustrate this trend, where understandable models are coupled with synthetic verification or screening platforms. Models now can suggest candidates with high T_g, optimal mobility, or desired dielectric properties.

However, some challenges remain. Interpretability metrics vary across studies, and there is no baseline by which the quality of attribution in polymer GNNs can be judged. Morphological and processing effects are often not addressed by current models, even though it is well understood that these influence on polymer properties. Furthermore, CG graph representations contribute to explainability but continue to suffer from a lack of standardization in their definition, parsing, or encoding. At present, one of the biggest challenges in polymer ML is not just the use of existing data sets and their limitations, but also the intentional development of new ones. Polymers are structurally diverse with properties that depend strongly on experimental conditions, and there is currently little standardization across current sources. More direct efforts are needed, such as those led by CRIPT, to create a more uniform framework for polymer data sets. There is a clear need for a polymer-focused data set, where a task force of researchers can run millions of experiments under controlled conditions to build up a large, publicly available, and validated data set for ML models. This community effort should be transparent in the creation of the data, the maintenance of the data, and should avoid bias toward certain types polymers over others to ensure diversity in data sets which is critical for ML models that require that in their training data set to ensure generalizable models. Properties measured under different conditions should be carefully controlled, and computationally derived values should especially be validated before inclusion. Another challenge, is sustainability, many data sets are created for a singular project and then are not maintained or expanded. Overall, many challenges still remain in curating effective and validated, open source data sets that are maintained, balanced, large, and diverse enough for ML models.

To follow up on this work, this field should in the future:

Develop benchmark data sets that integrate property prediction with attribution labels;
Promote modular and hierarchical GNN architectures supporting hybrid input graphs and dynamic behavior predictions;
Enable interpretability with experimental pipelines, e.g., synthesis constraints and descriptors for applied knowledge extractability.

Overall, polymer GNNs have evolved from accurate black-box models to chemically aware and explainable tools. Merging CG graph representations with attribution techniques not only improves transparency but also enables rational polymer design. With increasingly diverse model architectures and larger data sets, interpretable GNNs, especially those based on CG representations are well poised to become the mainstay instruments in data-driven polymer science.

7. Conclusions

Click to copy section linkSection link copied!

GNNs have advanced polymer property prediction with adaptable architectures, scalable training, and the ability to model complex structure–property relationships from graph-based molecular representations directly. Across property domains such as electronic, thermal, transport, and beyond GNNs have shown superiority to traditional descriptors and ML pipelines, particularly when tailored to the unique challenges of polymers, such as chain repetition, stereochemistry, and topological heterogeneity.

This review not only highlights the predictive capabilities of MPNNs, GCNs, and multitask GNN variants, but also the increasing trend for polymer-specific innovation in model architecture. These include the use of periodic graphs, PU abstractions, and chemically pertinent edge features, all of which enhance the expressiveness and physical relevance of graph inputs. Furthermore, the integration of explainability tools such as LRP, saliency maps, and attention-based attribution will drive the field forward toward not just accurate but also chemically meaningful models for applied knowledge extractability.

Another direction for the future is the further development of CG graph representations, connecting high-level structure features and low-level graph connectivity. In the PU-GNNs and glycan modeling frameworks, CG graphs allow models to capture mesoscopic polymer properties such as repeat-unit behavior, block copolymer morphology, or motif-based interactions, while maintaining computational tractability and interpretability. Paired with scalable GNN architectures and hybrid learning paradigms (e.g., multitasking, transfer learning), CG representations can offer a promising path forward.

The success of GNNs in polymer informatics will depend on the integration of domain knowledge, standardized graph encodings, benchmarking data sets, and interpretable learning targets. Building on the innovations and challenges discussed here, future work can unleash the potential for graph-based deep learning to accelerate polymer discovery, design, and development.

Author Information

Click to copy section linkSection link copied!

Corresponding Author
- Hector Medina - School of Engineering, Liberty University, Lynchburg, Virginia 24515, United States; https://orcid.org/0000-0003-1014-2275; Email: [email protected]
Author
- Rachel Drake - School of Engineering, Liberty University, Lynchburg, Virginia 24515, United States
Notes
The authors declare no competing financial interest.

References

Click to copy section linkSection link copied!

This article references 260 other publications.

1
Liang, Y.; Yu, L. Development of semiconducting polymers for solar energy harvesting. Polym. Rev. 2010, 50 (4), 454– 473, DOI: 10.1080/15583724.2010.515765
Google Scholar
There is no corresponding record for this reference.
2
Baytekin, B.; Baytekin, H. T.; Grzybowski, B. A. Retrieving and converting energy from polymers: deployable technologies and emerging concepts. Energy Environ. Sci. 2013, 6 (12), 3467– 3482, DOI: 10.1039/c3ee41360h
Google Scholar
There is no corresponding record for this reference.
3
Abdelhamid, M. E.; O’Mullane, A. P.; Snook, G. A. Storing energy in plastics: a review on conducting polymers & their role in electrochemical energy storage. RSC Adv. 2015, 5 (15), 11611– 11626, DOI: 10.1039/C4RA15947K
Google Scholar
There is no corresponding record for this reference.
4
Fan, J.; Njuguna, J. An introduction to lightweight composite materials and their use in transport structures. In Lightweight Composite Structures in Transport; Elsevier, 2016; pp 3– 34.
Google Scholar
There is no corresponding record for this reference.
5
Singh, J.; Srivastawa, K.; Jana, S.; Dixit, C.; Ravichandran, S. Advancements in lightweight materials for aerospace structures: A comprehensive review. Acceleron Aerosp. J. 2024, 2 (3), 173– 183, DOI: 10.61359/11.2106-2409
Google Scholar
There is no corresponding record for this reference.
6
Yeo, S. J.; Oh, M. J.; Yoo, P. J. Structurally controlled cellular architectures for high-performance ultra-lightweight materials. Adv. Mater. 2019, 31 (34), 1803670 DOI: 10.1002/adma.201803670
Google Scholar
There is no corresponding record for this reference.
7
Cui, H.; Zhao, Q.; Zhang, L.; Du, X. Intelligent polymer-based bioinspired actuators: from monofunction to multifunction. Adv. Intell. Syst. 2020, 2 (11), 2000138 DOI: 10.1002/aisy.202000138
Google Scholar
There is no corresponding record for this reference.
8
Weng, M.; Ding, M.; Zhou, P.; Ye, Y.; Luo, Z.; Ye, X.; Guo, Q.; Chen, L. Multi-functional and integrated actuators made with bio-inspired cobweb carbon nanotube-polymer composites. Chem. Eng. J. 2023, 452, 139146 DOI: 10.1016/j.cej.2022.139146
Google Scholar
There is no corresponding record for this reference.
9
Lendlein, A.; Rehahn, M.; Buchmeiser, M. R.; Haag, R. Polymers in biomedicine and electronics. Macromol. Rapid Commun. 2010, 31 (17), 1487– 1491, DOI: 10.1002/marc.201000426
Google Scholar
There is no corresponding record for this reference.
10
Medina, H.; Child, N. A review of developments in carbon-based nanocomposite electrodes for noninvasive electroencephalography. Sensors 2025, 25 (7), 2274 DOI: 10.3390/s25072274
Google Scholar
There is no corresponding record for this reference.
11
Farmer, C.; Medina, H. Effects of electrostriction on the bifurcated electro-mechanical performance of conical dielectric elastomer actuators and sensors. Robotica 2023, 41 (1), 215– 235, DOI: 10.1017/S0263574722001254
Google Scholar
There is no corresponding record for this reference.
12
Medina, H.; Farmer, C.; Liu, I. Dielectric elastomer-based actuators: a modeling and control review for non-experts. In Actuators; MDPI, 2024; Vol. 13, p 151.
Google Scholar
There is no corresponding record for this reference.
13
Korn, D.; Farmer, C.; Medina, H. A detailed solution framework for the out-of-plane displacement of circular dielectric elastomer actuators. Eng. Rep. 2022, 4 (1), e12442 DOI: 10.1002/eng2.12442
Google Scholar
There is no corresponding record for this reference.
14
Peterson, G. I.; Choi, T.-L. Cascade polymerizations: recent developments in the formation of polymer repeat units by cascade reactions. Chem. Sci. 2020, 11 (19), 4843– 4854, DOI: 10.1039/D0SC01475C
Google Scholar
There is no corresponding record for this reference.
15
Cao, C.; Lin, Y. Correlation between the glass transition temperatures and repeating unit structure for high molecular weight polymers. J. Chem. Inf. Comput. Sci. 2003, 43 (2), 643– 650, DOI: 10.1021/ci0202990
Google Scholar
There is no corresponding record for this reference.
16
Tezuka, Y.; Oike, H. Topological polymer chemistry: systematic classification of nonlinear polymer topologies. J. Am. Chem. Soc. 2001, 123 (47), 11570– 11576, DOI: 10.1021/ja0114409
Google Scholar
There is no corresponding record for this reference.
17
Gu, Y.; Zhao, J.; Johnson, J. A. A (macro) molecular-level understanding of polymer network topology. Trends Chem. 2019, 1 (3), 318– 334, DOI: 10.1016/j.trechm.2019.02.017
Google Scholar
There is no corresponding record for this reference.
18
Li, Y.; Abberton, B. C.; Kröger, M.; Liu, W. K. Challenges in multiscale modeling of polymer dynamics. Polymers 2013, 5 (2), 751– 832, DOI: 10.3390/polym5020751
Google Scholar
There is no corresponding record for this reference.
19
Schmid, F. Understanding and modeling polymers: The challenge of multiple scales. ACS Polymers Au 2023, 3 (1), 28– 58, DOI: 10.1021/acspolymersau.2c00049
Google Scholar
There is no corresponding record for this reference.
20
Jones, R. Density functional theory: Its origins, rise to prominence, and future. Rev. Mod. Phys. 2015, 87 (3), 897, DOI: 10.1103/RevModPhys.87.897
Google Scholar
There is no corresponding record for this reference.
21
Parr, R. G. Density-Functional Theory of Atoms and Molecules; Oxford University Press, 1989.
Google Scholar
There is no corresponding record for this reference.
22
Karplus, M.; McCammon, J. A. Molecular dynamics simulations of biomolecules. Nat. Struct. Biol. 2002, 9 (9), 646– 652, DOI: 10.1038/nsb0902-646
Google Scholar
There is no corresponding record for this reference.
23
Hollingsworth, S. A.; Dror, R. O. Molecular dynamics simulation for all. Neuron 2018, 99 (6), 1129– 1143, DOI: 10.1016/j.neuron.2018.08.011
Google Scholar
There is no corresponding record for this reference.
24
Tuckerman, M. E.; Martyna, G. J. Understanding modern molecular dynamics: Techniques and applications. J. Phys. Chem. B 2000, 104, 159– 178, DOI: 10.1021/jp992433y
Google Scholar
There is no corresponding record for this reference.
25
Bhavikatti, S. Finite Element Analysis. New Age International, 2005.
Google Scholar
There is no corresponding record for this reference.
26
Kim, N.-H.; Sankar, B. V.; Kumar, A. V. Introduction to Finite Element Analysis and Design; John Wiley & Sons, 2018.
Google Scholar
There is no corresponding record for this reference.
27
Farmer, C.; Medina, H. Hyperelastics. jl: A julia package for hyperelastic material modelling with a large collection of models. J. Open Source Softw. 2024, 9 (96), 6314 DOI: 10.21105/joss.06314
Google Scholar
There is no corresponding record for this reference.
28
Farmer, C.; Medina, H. Impact-4ccs: Integrated modeling and prediction using ab initio and trained potentials for collision cross sections. J. Comput. Chem. 2025, 46 (11), e70106 DOI: 10.1002/jcc.70106
Google Scholar
There is no corresponding record for this reference.
29
Medina, H.; Farmer, C. Current challenges in monitoring low contaminant levels of per-and polyfluoroalkyl substances in water matrices in the field. Toxics 2024, 12 (8), 610 DOI: 10.3390/toxics12080610
Google Scholar
There is no corresponding record for this reference.
30
Farmer, C.; Medina, H. Universal interatomic potentials with dft for understanding orbital localization in polydimethylsiloxane-amorphous silica nanocomposites. ACS Omega 2025, 10 (49), 60812– 60818, DOI: 10.1021/acsomega.5c09273
Google Scholar
There is no corresponding record for this reference.
31
McCormick, T. M.; Bridges, C. R.; Carrera, E. I.; DiCarmine, P. M.; Gibson, G. L.; Hollinger, J.; Kozycz, L. M.; Seferos, D. S. Conjugated polymers: Evaluating dft methods for more accurate orbital energy modeling. Macromolecules 2013, 46 (10), 3879– 3886, DOI: 10.1021/ma4005023
Google Scholar
There is no corresponding record for this reference.
32
Adekoya, O. C.; Adekoya, G. J.; Sadiku, E. R.; Hamam, Y.; Ray, S. S. Application of dft calculations in designing polymer-based drug delivery systems: An overview. Pharmaceutics 2022, 14 (9), 1972, DOI: 10.3390/pharmaceutics14091972
Google Scholar
There is no corresponding record for this reference.
33
Cohen, A. J.; Mori-Sánchez, P.; Yang, W. Challenges for density functional theory. Chem. Rev. 2012, 112 (1), 289– 320, DOI: 10.1021/cr200107z
Google Scholar
There is no corresponding record for this reference.
34
Kohn, W.; Becke, A. D.; Parr, R. G. Density functional theory of electronic structure. J. Phys. Chem. A 1996, 100 (31), 12974– 12980, DOI: 10.1021/jp960669l
Google Scholar
There is no corresponding record for this reference.
35
Dawson, W.; Degomme, A.; Stella, M.; Nakajima, T.; Ratcliff, L. E.; Genovese, L. Density functional theory calculations of large systems: Interplay between fragments, observables, and computational complexity. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2022, 12 (3), e1574 DOI: 10.1002/wcms.1574
Google Scholar
There is no corresponding record for this reference.
36
Bishop, M.; Kalos, M.; Frisch, H. Molecular dynamics of polymeric systems. J. Chem. Phys. 1979, 70 (3), 1299– 1304, DOI: 10.1063/1.437567
Google Scholar
There is no corresponding record for this reference.
37
Han, J.; Gee, R. H.; Boyd, R. H. Glass transition temperatures of polymers from molecular dynamics simulations. Macromolecules 1994, 27 (26), 7781– 7784, DOI: 10.1021/ma00104a036
Google Scholar
There is no corresponding record for this reference.
38
Lau, D.; Jian, W.; Yu, Z.; Hui, D. Nano-engineering of construction materials using molecular dynamics simulations: Prospects and challenges. Compos. Part B 2018, 143, 282– 291, DOI: 10.1016/j.compositesb.2018.01.014
Google Scholar
There is no corresponding record for this reference.
39
Mackerle, J. Finite element analysis and simulation of polymersan addendum: a bibliography (1996–2002). Modell. Simul. Mater. Sci. Eng. 2003, 11 (2), 195 DOI: 10.1088/0965-0393/11/2/307
Google Scholar
There is no corresponding record for this reference.
40
Zeng, X.; Brown, L. P.; Endruweit, A.; Matveev, M.; Long, A. C. Geometrical modelling of 3d woven reinforcements for polymer composites: Prediction of fabric permeability and composite mechanical properties. Compos. Part A: Appl. Sci. Manuf. 2014, 56, 150– 160, DOI: 10.1016/j.compositesa.2013.10.004
Google Scholar
There is no corresponding record for this reference.
41
Alshahrani, H. Characterization and finite element modeling of coupled properties during polymer composites forming processes. Mech. Mater. 2020, 144, 103370 DOI: 10.1016/j.mechmat.2020.103370
Google Scholar
There is no corresponding record for this reference.
42
Thompson, A. P.; Aktulga, H. M.; Berger, R.; Bolintineanu, D. S.; Brown, W. M.; Crozier, P. S.; In’t Veld, P. J.; Kohlmeyer, A.; Moore, S. G.; Nguyen, T. D.; Shan, R.; Stevens, M. J.; Tranchida, J.; Trott, C.; Plimpton, S. J. Lammps-a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales. Comput. Phys. Commun. 2022, 271, 108171 DOI: 10.1016/j.cpc.2021.108171
Google Scholar
There is no corresponding record for this reference.
43
Ratcliff, L. E.; Mohr, S.; Huhs, G.; Deutsch, T.; Masella, M.; Genovese, L. Challenges in large scale quantum mechanical calculations. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2017, 7 (1), e1290 DOI: 10.1002/wcms.1290
Google Scholar
There is no corresponding record for this reference.
44
Fredrickson, G. H. Computational field theory of polymers: opportunities and challenges. Soft Matter 2007, 3 (11), 1329– 1334, DOI: 10.1039/b710604a
Google Scholar
There is no corresponding record for this reference.
45
Sindu, B.; Hamaekers, J. Feature-based prediction of properties of cross-linked epoxy polymers by molecular dynamics and machine learning techniques. Modell. Simul. Mater. Sci. Eng. 2025, 33 (6), 065010 DOI: 10.1088/1361-651X/adf56c
Google Scholar
There is no corresponding record for this reference.
46
Chew, A. K.; Afzal, M. A. F.; Chandrasekaran, A.; Kamps, J. H.; Ramakrishnan, V. Designing the next generation of polymers with machine learning and physics-based models. Mach. Learn. Sci. Technol. 2024, 5 (4), 045031 DOI: 10.1088/2632-2153/ad88d7
Google Scholar
There is no corresponding record for this reference.
47
He, J.; Rong, F. Prediction of molecularly imprinted polymer binding affinity based on graph neural networks. J. Phys.: Conference Series 2025, 3030, 012009
Google Scholar
There is no corresponding record for this reference.
48
Nagasawa, S.; Al-Naamani, E.; Saeki, A. Computer-aided screening of conjugated polymers for organic solar cell: classification by random forest. J. Phys. Chem. Lett. 2018, 9 (10), 2639– 2646, DOI: 10.1021/acs.jpclett.8b00635
Google Scholar
There is no corresponding record for this reference.
49
Kishino, M.; Matsumoto, K.; Kobayashi, Y.; Taguchi, R.; Akamatsu, N.; Shishido, A. Fatigue life prediction of bending polymer films using random forest. Int. J. Fatigue 2023, 166, 107230 DOI: 10.1016/j.ijfatigue.2022.107230
Google Scholar
There is no corresponding record for this reference.
50
Arora, A.; Lin, T.-S.; Rebello, N. J.; Av-Ron, S. H.; Mochigase, H.; Olsen, B. D. Random forest predictor for diblock copolymer phase behavior. ACS Macro Lett. 2021, 10 (11), 1339– 1345, DOI: 10.1021/acsmacrolett.1c00521
Google Scholar
There is no corresponding record for this reference.
51
Malashin, I.; Tynchenko, V.; Gantimurov, A.; Nelyub, V.; Borodulin, A. Support vector machines in polymer science: a review. Polymers 2025, 17 (4), 491, DOI: 10.3390/polym17040491
Google Scholar
There is no corresponding record for this reference.
52
Ziaee, H.; Hosseini, S. M.; Sharafpoor, A.; Fazavi, M.; Ghiasi, M. M.; Bahadori, A. Prediction of solubility of carbon dioxide in different polymers using support vector machine algorithm. J. Taiwan Inst. Chem. Eng. 2015, 46, 205– 213, DOI: 10.1016/j.jtice.2014.09.015
Google Scholar
There is no corresponding record for this reference.
53
Chen, F.-C. Virtual screening of conjugated polymers for organic photovoltaic devices using support vector machines and ensemble learning. Int. J. Polym. Sci. 2019, 2019 (1), 4538514 DOI: 10.1155/2019/4538514
Google Scholar
There is no corresponding record for this reference.
54
Brereton, R. G.; Lloyd, G. R. Support vector machines for classification and regression. Analyst 2010, 135 (2), 230– 267, DOI: 10.1039/B918972F
Google Scholar
There is no corresponding record for this reference.
55
Youshia, J.; Ali, M. E.; Lamprecht, A. Artificial neural network based particle size prediction of polymeric nanoparticles. Eur. J. Pharm. Biopharm. 2017, 119, 333– 342, DOI: 10.1016/j.ejpb.2017.06.030
Google Scholar
There is no corresponding record for this reference.
56
Zhang, Z.; Friedrich, K. Artificial neural networks applied to polymer composites: a review. Compos. Sci. Technol. 2003, 63 (14), 2029– 2044, DOI: 10.1016/S0266-3538(03)00106-4
Google Scholar
There is no corresponding record for this reference.
57
El Kadi, H. Modeling the mechanical behavior of fiber-reinforced polymeric composite materials using artificial neural networksa review. Compos. Struct. 2006, 73 (1), 1– 23, DOI: 10.1016/j.compstruct.2005.01.020
Google Scholar
There is no corresponding record for this reference.
58
Liu, W.; Cao, C. Artificial neural network prediction of glass transition temperature of polymers. Colloid Polym. Sci. 2009, 287 (7), 811– 818, DOI: 10.1007/s00396-009-2035-y
Google Scholar
There is no corresponding record for this reference.
59
Chen, L.; Pilania, G.; Batra, R.; Huan, T. D.; Kim, C.; Kuenneth, C.; Ramprasad, R. Polymer informatics: Current status and critical next steps. Mater. Sci. Eng. R 2021, 144, 100595 DOI: 10.1016/j.mser.2020.100595
Google Scholar
There is no corresponding record for this reference.
60
Butler, K. T.; Davies, D. W.; Cartwright, H.; Isayev, O.; Walsh, A. Machine learning for molecular and materials science. Nature 2018, 559, 547– 555, DOI: 10.1038/s41586-018-0337-2
Google Scholar
There is no corresponding record for this reference.
61
Schleder, G. R.; Padilha, A. C. M.; Acosta, C. M.; Costa, M.; Fazzio, A. From dft to machine learning: Recent approaches to materials science-a review. J. Chem. Inf. Model. 2019, 2 (4), 1347– 1356, DOI: 10.1088/2515-7639/ab084b
Google Scholar
There is no corresponding record for this reference.
62
Sha, W.; Li, Y.; Tang, S.; Tian, J.; Zhao, Y.; Guo, Y.; Zhang, W.; Zhang, X.; Lu, S.; Cao, Y.-C.; Cheng, S. Machine learning in polymer informatics. InfoMat 2021, 3 (4), 353– 361, DOI: 10.1002/inf2.12167
Google Scholar
There is no corresponding record for this reference.
63
Martin, T. B.; Audus, D. J. Emerging trends in machine learning: a polymer perspective. ACS Polymers Au 2023, 3 (3), 239– 258, DOI: 10.1021/acspolymersau.2c00053
Google Scholar
There is no corresponding record for this reference.
64
Ramprasad, R.; Batra, R.; Pilania, G.; Mannodi-Kanakkithodi, A.; Kim, C. Machine learning in materials informatics: recent applications and prospects. npj Comput. Mater. 2017, 3 (1), 54, DOI: 10.1038/s41524-017-0056-5
Google Scholar
There is no corresponding record for this reference.
65
Chen, G.; Shen, Z.; Iyer, A.; Ghumman, U. F.; Tang, S.; Bi, J.; Chen, W.; Li, Y. Machine-learning-assisted de novo design of organic molecules and polymers: opportunities and challenges. Polymers 2020, 12 (1), 163, DOI: 10.3390/polym12010163
Google Scholar
There is no corresponding record for this reference.
66
Wu, S.; Yamada, H.; Hayashi, Y.; Zamengo, M.; Yoshida, R. Potentials and challenges of polymer informatics: exploiting machine learning for polymer design, 2020, arXiv:2010.07683. arXiv.org e-Print archive https://arxiv.org/abs/2010.07683.
Google Scholar
There is no corresponding record for this reference.
67
Kumar, J. N.; Li, Q.; Jun, Y. Challenges and opportunities of polymer design with machine learning and high throughput experimentation. MRS Commun. 2019, 9 (2), 537– 544, DOI: 10.1557/mrc.2019.54
Google Scholar
There is no corresponding record for this reference.
68
Wieder, O.; Kohlbacher, S.; Kuenemann, M.; Garon, A.; Ducrot, P.; Seidel, T.; Langer, T. A compact review of molecular property prediction with graph neural networks. Drug Discovery Today: Technol. 2020, 37, 1– 12, DOI: 10.1016/j.ddtec.2020.11.009
Google Scholar
There is no corresponding record for this reference.
69
Wu, Z.; Pan, S.; Chen, F.; Long, G.; Zhang, C.; Philip, S. Y. A comprehensive survey on graph neural networks. IEEE Trans. Neural Netw. Learn. Syst. 2021, 32 (1), 4– 24, DOI: 10.1109/TNNLS.2020.2978386
Google Scholar
There is no corresponding record for this reference.
70
Gao, Q.; Dukker, T.; Schweidtmann, A. M.; Weber, J. M. Self-supervised graph neural networks for polymer property prediction. Mol. Syst. Des. Eng. 2024, 9 (11), 1130– 1143, DOI: 10.1039/D4ME00088A
Google Scholar
There is no corresponding record for this reference.
71
Inae, E.; Liu, Y.; Zhu, Y.; Xu, J.; Liu, G.; Zhang, R.; Luo, T.; Jiang, M. Modeling Polymers with Neural Networks; American Chemical Society, 2025.
Google Scholar
There is no corresponding record for this reference.
72
Gurney, K. An Introduction to Neural Networks; CRC Press, 2018.
Google Scholar
There is no corresponding record for this reference.
73
Haykin, S.; Network, N. A comprehensive foundation. Neural Netw. 2004, 2, 41
Google Scholar
There is no corresponding record for this reference.
74
LeCun, Y.; Bottou, L.; Bengio, Y.; Haffner, P. Gradient-based learning applied to document recognition. Proc. IEEE 1998, 86 (11), 2278– 2324, DOI: 10.1109/5.726791
Google Scholar
There is no corresponding record for this reference.
75
Li, Z.; Liu, F.; Yang, W.; Peng, S.; Zhou, J. A survey of convolutional neural networks: analysis, applications, and prospects. IEEE Trans. Neural Netw. Learn. Syst. 2022, 33 (12), 6999– 7019, DOI: 10.1109/TNNLS.2021.3084827
Google Scholar
There is no corresponding record for this reference.
76
Gu, J.; Wang, Z.; Kuen, J.; Ma, L.; Shahroudy, A.; Shuai, B.; Liu, T.; Wang, X.; Wang, G.; Cai, J.; Chen, T. Recent advances in convolutional neural networks. Pattern Recognit. 2018, 77, 354– 377, DOI: 10.1016/j.patcog.2017.10.013
Google Scholar
There is no corresponding record for this reference.
77
Ge, W.; De Silva, R.; Fan, Y.; Sisson, S. A.; Stenzel, M. H. Machine learning in polymer research. Adv. Mater. 2025, 37 (11), 2413695 DOI: 10.1002/adma.202413695
Google Scholar
There is no corresponding record for this reference.
78
Platonov, O.; Kuznedelev, D.; Diskin, M.; Babenko, A.; Prokhorenkova, L. A critical look at the evaluation of gnns under heterophily: Are we really making progress? 2023, arXiv:2302.11640. arXiv.org e-Print archive https://arxiv.org/abs/2302.11640.
Google Scholar
There is no corresponding record for this reference.
79
Wang, Y.; Li, Z.; Barati Farimani, A. Graph neural networks for molecules. In Machine Learning in Molecular Sciences; Springer, 2023; pp 21– 66.
Google Scholar
There is no corresponding record for this reference.
80
Hamilton, W.; Ying, Z.; Leskovec, J. Inductive representation learning on large graphs. In NIPS’17: Proceedings of the 31st International Conference on Neural Information Processing Systems; Department of Computer Science: Stanford, CA, 2017; pp 1025– 1035.
Google Scholar
There is no corresponding record for this reference.
81
Xu, K.; Hu, W.; Leskovec, J.; Jegelka, S. How powerful are graph neural networks? 2018, arXiv:1810.00826. arXiv.org e-Print archive https://arxiv.org/abs/1810.00826.
Google Scholar
There is no corresponding record for this reference.
82
Sharma, K.; Lee, Y.-C.; Nambi, S.; Salian, A.; Shah, S.; Kim, S.-W.; Kumar, S. A survey of graph neural networks for social recommender systems. ACM Computing Surveys 2024, 56 (10), 1– 34, DOI: 10.1145/3661821
Google Scholar
There is no corresponding record for this reference.
83
Kumar, S.; Mallik, A.; Khetarpal, A.; Panda, B. S. Influence maximization in social networks using graph embedding and graph neural network. Informat. Sci. 2022, 607, 1617– 1636, DOI: 10.1016/j.ins.2022.06.075
Google Scholar
There is no corresponding record for this reference.
84
Zhang, X.-M.; Liang, L.; Liu, L.; Tang, M.-J. Graph neural networks and their current applications in bioinformatics. Front. Genetics 2021, 12, 690049 DOI: 10.3389/fgene.2021.690049
Google Scholar
There is no corresponding record for this reference.
85
Malla, A. M.; Banka, A. A. A systematic review of deep graph neural networks: Challenges, classification, architectures, applications & potential utility in bioinformatics, 2023, arXiv:2311.02127. arXiv.org e-Print archive https://arxiv.org/abs/2311.02127.
Google Scholar
There is no corresponding record for this reference.
86
Zheng, P.; Wang, S.; Wang, X.; Zeng, X. Artificial intelligence in bioinformatics and drug repurposing: methods and applications. In Frontiers Research Topics , 2022.
Google Scholar
There is no corresponding record for this reference.
87
Reiser, P.; Neubert, M.; Eberhard, A.; Torresi, L.; Zhou, C.; Shao, C.; Metni, H.; van Hoesel, C.; Schopmans, H.; Sommer, T.; Friederich, P. Graph neural networks for materials science and chemistry. Commun. Mater. 2022, 3 (1), 93, DOI: 10.1038/s43246-022-00315-6
Google Scholar
There is no corresponding record for this reference.
88
Fung, V.; Zhang, J.; Juarez, E.; Sumpter, B. G. Benchmarking graph neural networks for materials chemistry. npj Computat. Mater. 2021, 7 (1), 84, DOI: 10.1038/s41524-021-00554-0
Google Scholar
There is no corresponding record for this reference.
89
Maurizi, M.; Gao, C.; Berto, F. Predicting stress, strain and deformation fields in materials and structures with graph neural networks. Sci. Rep. 2022, 12 (1), 21834 DOI: 10.1038/s41598-022-26424-3
Google Scholar
There is no corresponding record for this reference.
90
Kipf, T. N.; Welling, M. Semi-supervised classification with graph convolutional networks, 2016, arXiv:1609.02907. arXiv.org e-Print archive https://arxiv.org/abs/1609.02907.
Google Scholar
There is no corresponding record for this reference.
91
Gilmer, J.; Schoenholz, S. S.; Riley, P. F.; Vinyals, O.; Dahl, G. E. Neural message passing for quantum chemistry. In International Conference on Machine Learning; PMLR, 2017; pp 1263– 1272.
Google Scholar
There is no corresponding record for this reference.
92
Veličković, P.; Cucurull, G.; Casanova, A.; Romero, A.; Lio, P.; Bengio, Y. “ Graph attention networks, 2017, arXiv:1710.10903. arXiv.org e-Print archive https://arxiv.org/abs/1710.10903.
Google Scholar
There is no corresponding record for this reference.
93
Liu, G.; Inae, E.; Jiang, M. Deep learning for polymer property prediction. In Deep Learning for Polymer Discovery: Foundation and Advances; Springer, 2025; pp 17– 35.
Google Scholar
There is no corresponding record for this reference.
94
Dong, C.; Li, D.; Liu, J. Glass transition temperature prediction of polymers via graph reinforcement learning. Langmuir 2024, 40 (35), 18568– 18580, DOI: 10.1021/acs.langmuir.4c01906
Google Scholar
There is no corresponding record for this reference.
95
Wang, X.; Liao, L.; Huang, C.; Wu, X. Prediction of mechanical properties of cross-linked polymer interface by graph convolution network. Acta Mechanica Sinica 2026, 42 (3), 424627 DOI: 10.1007/s10409-024-24627-x
Google Scholar
There is no corresponding record for this reference.
96
Safari, H.; Bavarian, M. Enhancing polymer reaction engineering through the power of machine learning. In Systems and Control Transactions , 2024; p 157792.
Google Scholar
There is no corresponding record for this reference.
97
Liao, H.-C.; Lin, Y.-H.; Peng, C.-H.; Li, Y.-P. Directed message passing neural networks for accurate prediction of polymer-solvent interaction parameters. ACS Engineering Au 2025, 5, 530, DOI: 10.1021/acsengineeringau.5c00027
Google Scholar
There is no corresponding record for this reference.
98
Kim, S.; Chen, J.; Cheng, T.; Gindulyte, A.; He, J.; He, S.; Li, Q.; Shoemaker, B. A.; Thiessen, P. A.; Yu, B. Pubchem 2023 update. Nucleic Acids Res. 2023, 51 (D1), D1373– D1380, DOI: 10.1093/nar/gkac956
Google Scholar
There is no corresponding record for this reference.
99
Gao, Z.-Y.; Liu, G.-Q. Recent progress of web-enable material database and a case study of nims and matweb. J. Mater. Eng. 2013, 3 (11), 89– 96, DOI: 10.3969/j.issn.1001-4381.2013.06.001
Google Scholar
There is no corresponding record for this reference.
100
Wypych, G. Handbook of Polymers; Elsevier, 2022.
Google Scholar
There is no corresponding record for this reference.
101
Patra, A.; Batra, R.; Chandrasekaran, A.; Kim, C.; Huan, T. D.; Ramprasad, R. A multi-fidelity information-fusion approach to machine learn and predict polymer bandgap. Comput. Mater. Sci. 2020, 172, 109286 DOI: 10.1016/j.commatsci.2019.109286
Google Scholar
There is no corresponding record for this reference.
102
Dhamankar, S.; Webb, M. A. Chemically specific coarse-graining of polymers: methods and prospects. J. Polym. Sci. 2021, 59 (22), 2613– 2643, DOI: 10.1002/pol.20210555
Google Scholar
There is no corresponding record for this reference.
103
Müller-Plathe, F. Coarse-graining in polymer simulation: from the atomistic to the mesoscopic scale and back. ChemPhysChem 2002, 3 (9), 754– 769, DOI: 10.1002/1439-7641(20020916)3:9<754::AID-CPHC754>3.0.CO;2-U
Google Scholar
There is no corresponding record for this reference.
104
Harmandaris, V. A.; Reith, D.; Van der Vegt, N. F.; Kremer, K. Comparison between coarse-graining models for polymer systems: Two mapping schemes for polystyrene. Macromol. Chem. Phys. 2007, 208 (19–20), 2109– 2120, DOI: 10.1002/macp.200700245
Google Scholar
There is no corresponding record for this reference.
105
Padding, J. T.; Briels, W. J. Systematic coarse-graining of the dynamics of entangled polymer melts: the road from chemistry to rheology. J. Phys.: Condens. Matter 2011, 23 (23), 233101 DOI: 10.1088/0953-8984/23/23/233101
Google Scholar
There is no corresponding record for this reference.
106
Zhang, X.; Sheng, Y.; Liu, X.; Yang, J.; Goddard, W. A., III; Ye, C.; Zhang, W. Polymer-unit graph: Advancing interpretability in graph neural network machine learning for organic polymer semiconductor materials. J. Chem. Theory Comput. 2024, 20 (7), 2908– 2920, DOI: 10.1021/acs.jctc.3c01385
Google Scholar
There is no corresponding record for this reference.
107
Mohapatra, S.; An, J.; Gómez-Bombarelli, R. Chemistry-informed macromolecule graph representation for similarity computation, unsupervised and supervised learning. Machine Learning: Sci. Technol. 2022, 3 (1), 015028 DOI: 10.1088/2632-2153/ac545e
Google Scholar
There is no corresponding record for this reference.
108
Wang, X.; Xu, Y.; Zheng, H.; Yu, K. A scalable graph neural network method for developing an accurate force field of large flexible organic molecules. J. Phys. Chem. Lett. 2021, 12 (33), 7982– 7987, DOI: 10.1021/acs.jpclett.1c02214
Google Scholar
There is no corresponding record for this reference.
109
Qiu, H.; Qiu, X.; Dai, X.; Sun, Z.-Y. Design of polyimides with targeted glass transition temperature using a graph neural network. J. Mater. Chem. C 2023, 11 (8), 2930– 2940, DOI: 10.1039/D2TC05174E
Google Scholar
There is no corresponding record for this reference.
110
Zeng, M.; Kumar, J. N.; Zeng, Z.; Savitha, R.; Chandrasekhar, V. R.; Hippalgaonkar, K. “ Graph convolutional neural networks for polymers property prediction, 2018, arXiv:1811.06231. arXiv.org e-Print archive https://arxiv.org/abs/1811.06231.
Google Scholar
There is no corresponding record for this reference.
111
Queen, O.; McCarver, G. A.; Thatigotla, S.; Abolins, B. P.; Brown, C. L.; Maroulas, V.; Vogiatzis, K. D. Polymer graph neural networks for multitask property learning. npj Computat. Mater. 2023, 9, 90, DOI: 10.1038/s41524-023-01034-3
Google Scholar
There is no corresponding record for this reference.
112
Bongini, P.; Bianchini, M.; Scarselli, F. Molecular generative graph neural networks for drug discovery. Neurocomputing 2021, 450, 242– 252, DOI: 10.1016/j.neucom.2021.04.039
Google Scholar
There is no corresponding record for this reference.
113
Scarselli, F.; Gori, M.; Tsoi, A. C.; Hagenbuchner, M.; Monfardini, G. The graph neural network model. IEEE Transact. Neural Networks 2009, 20 (1), 61– 80, DOI: 10.1109/TNN.2008.2005605
Google Scholar
There is no corresponding record for this reference.
114
Dwivedi, V. P.; Joshi, C. K.; Luu, A. T.; Laurent, T.; Bengio, Y.; Bresson, X. Benchmarking graph neural networks. J. Machine Learning Res. 2023, 24 (43), 1– 48
Google Scholar
There is no corresponding record for this reference.
115
Veličković, P. Everything is connected: Graph neural networks. Curr. Opin. Struct. Biol. 2023, 79, 102538 DOI: 10.1016/j.sbi.2023.102538
Google Scholar
There is no corresponding record for this reference.
116
Jegelka, S. Theory of graph neural networks: Representation and learning. In International Congress of Mathematicians , 2022; pp 1– 23.
Google Scholar
There is no corresponding record for this reference.
117
Liu, Z.; Zhou, J. Introduction to Graph Neural Networks; Springer Nature, 2022.
Google Scholar
There is no corresponding record for this reference.
118
Tang, M.; Li, B.; Chen, H. Application of message passing neural networks for molecular property prediction. Curr. Opin. Struct. Biol. 2023, 81, 102616 DOI: 10.1016/j.sbi.2023.102616
Google Scholar
There is no corresponding record for this reference.
119
Hasebe, T. Knowledge-embedded message-passing neural networks: improving molecular property prediction with human knowledge. ACS omega 2021, 6 (42), 27955– 27967, DOI: 10.1021/acsomega.1c03839
Google Scholar
There is no corresponding record for this reference.
120
Zhang, H.; Lin, Y.; Li, S.; Dai, M.; Zhang, Y.; Huang, L.; Pang, J.; Wu, P.; Peng, J.; Tang, Z.; Ding, P.; Xiao, W.; Song, N.; Dongbo, D. Substructure-enhanced mpnn for polymer discovery and knowledge: A study in predicting glass transition temperature. Macromolecules 2025, 58, 9515, DOI: 10.1021/acs.macromol.4c02859
Google Scholar
There is no corresponding record for this reference.
121
Aldeghi, M.; Coley, C. W. A graph representation of molecular ensembles for polymer property prediction. Chem. Sci. 2022, 13 (35), 10486– 10498, DOI: 10.1039/D2SC02839E
Google Scholar
There is no corresponding record for this reference.
122
Heid, E.; Greenman, K. P.; Chung, Y.; Li, S.-C.; Graff, D. E.; Vermeire, F. H.; Wu, H.; Green, W. H.; McGill, C. J. Chemprop: a machine learning package for chemical property prediction. J. Chem. Inf. Model. 2024, 64 (1), 9– 17, DOI: 10.1021/acs.jcim.3c01250
Google Scholar
There is no corresponding record for this reference.
123
Antoniuk, E. R.; Li, P.; Kailkhura, B.; Hiszpanski, A. M. Representing polymers as periodic graphs with learned descriptors for accurate polymer property predictions. J. Chem. Inf. Model. 2022, 62, 5435– 5445, DOI: 10.1021/acs.jcim.2c00875
Google Scholar
There is no corresponding record for this reference.
124
Sanchez Medina, E. I.; Kunchapu, S.; Sundmacher, K. Gibbs-helmholtz graph neural network for the prediction of activity coefficients of polymer solutions at infinite dilution. J. Phys. Chem. A 2023, 127 (46), 9863– 9873, DOI: 10.1021/acs.jpca.3c05892
Google Scholar
There is no corresponding record for this reference.
125
Battaglia, P. W.; Hamrick, J. B.; Bapst, V.; Sanchez-Gonzalez, A.; Zambaldi, V.; Malinowski, M.; Tacchetti, A.; Raposo, D.; Santoro, A.; Faulkner, R.; C, G.; F, S.; Ballard, A.; Justin, G.; George, D.; Ashish, V.; Kelsey, A.; Charles, N.; Victoria, L.; Chris, D.; Nicolas, H.; Dann, W.; Pushmeet, K.; Matt, B.; Oriol, V.; Yujia, L.; Razvan, P. Relational inductive biases, deep learning, and graph networks, 2018, arXiv:1806.01261. arXiv.org e-Print archive https://arxiv.org/abs/1806.01261.
Google Scholar
There is no corresponding record for this reference.
126
Gurnani, R.; Kuenneth, C.; Toland, A.; Ramprasad, R. Polymer informatics at scale with multitask graph neural networks. Chem. Mater. 2023, 35 (6), 1560– 1567, DOI: 10.1021/acs.chemmater.2c02991
Google Scholar
There is no corresponding record for this reference.
127
St John, P. C.; Phillips, C.; Kemper, T. W.; Wilson, A. N.; Guan, Y.; Crowley, M. F.; Nimlos, M. R.; Larsen, R. E. Message-passing neural networks for high-throughput polymer screening. J. Chem. Phys. 2019, 150, 234111 DOI: 10.1063/1.5099132
Google Scholar
There is no corresponding record for this reference.
128
Li, Y.; Zemel, R.; Brockschmidt, M.; Tarlow, D. Gated graph sequence neural networks. In Proceedings of ICLR16 , 2016.
Google Scholar
There is no corresponding record for this reference.
129
Dey, R.; Salem, F. M. Gate-variants of gated recurrent unit (gru) neural networks. In 2017 IEEE 60th international midwest symposium on circuits and systems (MWSCAS); IEEE, 2017; pp 1597– 1600.
Google Scholar
There is no corresponding record for this reference.
130
Sun, M.; Zhao, S.; Gilvary, C.; Elemento, O.; Zhou, J.; Wang, F. Graph convolutional networks for computational drug development and discovery. Briefings Bioinf. 2020, 21 (3), 919– 935, DOI: 10.1093/bib/bbz042
Google Scholar
There is no corresponding record for this reference.
131
Xu, X.; Zhao, X.; Wei, M.; Li, Z. A comprehensive review of graph convolutional networks: approaches and applications. Electronic Res. Archive 2023, 31 (7), 4185– 4215, DOI: 10.3934/era.2023213
Google Scholar
There is no corresponding record for this reference.
132
Zhou, J.; Cui, G.; Hu, S.; Zhang, Z.; Yang, C.; Liu, Z.; Wang, L.; Li, C.; Sun, M. Graph neural networks: A review of methods and applications. IEEE Transactions on Pattern Analysis and Machine Intelligence 2019, 42, 283– 300
Google Scholar
There is no corresponding record for this reference.
133
Hu, J.; Li, Z.; Lin, J.; Zhang, L. Prediction and interpretability of glass transition temperature of homopolymers by data-augmented graph convolutional neural networks. ACS Appl. Mater. Interfaces 2023, 15 (46), 54006– 54017, DOI: 10.1021/acsami.3c13698
Google Scholar
There is no corresponding record for this reference.
134
Hickey, K.; Feinstein, J.; Sivaraman, G.; MacDonell, M.; Yan, E.; Matherson, C.; Coia, S.; Xu, J.; Picel, K. Applying machine learning and quantum chemistry to predict the glass transition temperatures of polymers. Comput. Mater. Sci. 2024, 238, 112933 DOI: 10.1016/j.commatsci.2024.112933
Google Scholar
There is no corresponding record for this reference.
135
Kimmig, J.; Schuett, T.; Vollrath, A.; Zechel, S.; Schubert, U. S. Prediction of nanoparticle sizes for arbitrary methacrylates using artificial neuronal networks. Adv. Sci. 2021, 8 (23), 2102429 DOI: 10.1002/advs.202102429
Google Scholar
There is no corresponding record for this reference.
136
Tian, Y.; Zhang, Y.; Zhang, H. Recent advances in stochastic gradient descent in deep learning. Mathematics 2023, 11 (3), 682, DOI: 10.3390/math11030682
Google Scholar
There is no corresponding record for this reference.
137
Merity, S. “ Single headed attention rnn: Stop thinking with your head, 2019, arXiv:1911.11423. arXiv.org e-Print archive https://arxiv.org/abs/1911.11423.
Google Scholar
There is no corresponding record for this reference.
138
Inoue, H. Multi-sample dropout for accelerated training and better generalization, 2019, arXiv:1905.09788. arXiv.org e-Print archive https://arxiv.org/abs/1905.09788.
Google Scholar
There is no corresponding record for this reference.
139
Cho, K.; Van Merriënboer, B.; Bahdanau, D.; Bengio, Y. On the properties of neural machine translation: Encoder-decoder approaches, 2014, arXiv:1409.1259. arXiv.org e-Print archive https://arxiv.org/abs/1409.1259.
Google Scholar
There is no corresponding record for this reference.
140
Morris, C.; Ritzert, M.; Fey, M.; Hamilton, W. L.; Lenssen, J. E.; Rattan, G.; Grohe, M. Weisfeiler and leman go neural: Higher-order graph neural networks. Proceedings of the AAAI Conference on Artificial Intelligence 2019, 33, 4602– 4609, DOI: 10.1609/aaai.v33i01.33014602
Google Scholar
There is no corresponding record for this reference.
141
Ye, X.-b.; Guan, Q.; Luo, W.; Fang, L.; Lai, Z.-R.; Wang, J. Molecular substructure graph attention network for molecular property identification in drug discovery. Pattern Recognition 2022, 128, 108659 DOI: 10.1016/j.patcog.2022.108659
Google Scholar
There is no corresponding record for this reference.
142
Vrahatis, A. G.; Lazaros, K.; Kotsiantis, S. Graph attention networks: a comprehensive review of methods and applications. Future Internet 2024, 16 (9), 318, DOI: 10.3390/fi16090318
Google Scholar
There is no corresponding record for this reference.
143
Xiong, Z.; Wang, D.; Liu, X.; Zhong, F.; Wan, X.; Li, X.; Li, Z.; Luo, X.; Chen, K.; Jiang, H.; Z M Pushing the boundaries of molecular representation for drug discovery with the graph attention mechanism. J. Med. Chem. 2020, 63 (16), 8749– 8760, DOI: 10.1021/acs.jmedchem.9b00959
Google Scholar
There is no corresponding record for this reference.
144
Li, D.; Ru, Y.; Liu, J. Gatboost: Mining graph attention networks-based important substructures of polymers for a better property prediction. Mater. Today Commun. 2024, 38, 107577 DOI: 10.1016/j.mtcomm.2023.107577
Google Scholar
There is no corresponding record for this reference.
145
Chen, T.; Guestrin, C. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , 2016; pp 785– 794.
Google Scholar
There is no corresponding record for this reference.
146
Liu, T.; Jiang, A.; Zhou, J.; Li, M.; Kwan, H. K. Graphsage-based dynamic spatial-temporal graph convolutional network for traffic prediction. IEEE Transactions on Intelligent Transportation Systems 2023, 24 (10), 11210– 11224, DOI: 10.1109/TITS.2023.3279929
Google Scholar
There is no corresponding record for this reference.
147
Xie, T.; France-Lanord, A.; Wang, Y.; Lopez, J.; Stolberg, M. A.; Hill, M.; Leverick, G. M.; Gomez-Bombarelli, R.; Johnson, J. A.; Shao-Horn, Y.; Grossman, J. C. Accelerating amorphous polymer electrolyte screening by learning to reduce errors in molecular dynamics simulated properties. Nat. Commun. 2022, 13 (1), 3415 DOI: 10.1038/s41467-022-30994-1
Google Scholar
There is no corresponding record for this reference.
148
Qiu, H.; Wang, J.; Qiu, X.; Dai, X.; Sun, Z.-Y. Heat-resistant polymer discovery by utilizing interpretable graph neural network with small data. Macromolecules 2024, 57 (8), 3515– 3528, DOI: 10.1021/acs.macromol.4c00508
Google Scholar
There is no corresponding record for this reference.
149
Park, J.; Shim, Y.; Lee, F.; Rammohan, A.; Goyal, S.; Shim, M.; Jeong, C.; Kim, D. S. Prediction and interpretation of polymer properties using the graph convolutional network. ACS Polymers Au 2022, 2, 213, DOI: 10.1021/acspolymersau.1c00050
Google Scholar
There is no corresponding record for this reference.
150
Burkholz, R.; Quackenbush, J.; Bojar, D. Using graph convolutional neural networks to learn a representation for glycans. Cell Rep. 2021, 35 (11), 109251, DOI: 10.1016/j.celrep.2021.109251
Google Scholar
There is no corresponding record for this reference.
151
Volgin, I. V.; Batyr, P. A.; Matseevich, A. V.; Dobrovskiy, A. Y.; Andreeva, M. V.; Nazarychev, V. M.; Larin, S. V.; Goikhman, M. Y.; Vizilter, Y. V.; Askadskii, A. A.; Lyulin, S. V. Machine learning with enormous “synthetic” data sets: Predicting glass transition temperature of polyimides using graph convolutional neural networks. ACS Omega 2022, 7 (48), 43678– 43691, DOI: 10.1021/acsomega.2c04649
Google Scholar
There is no corresponding record for this reference.
152
Otsuka, S.; Kuwajima, I.; Hosoya, J.; Xu, Y.; Yamazaki, M. Polyinfo: Polymer database for polymeric materials design. In 2011 International Conference on Emerging Intelligent Data and Web Technologies; IEEE, 2011; pp 22– 29.
Google Scholar
There is no corresponding record for this reference.
153
Brandrup, J.; Immergut, E. H.; Grulke, E. A.; Abe, A.; Bloch, D. R. Polymer Handbook; Wiley: New York, 1999; Vol. 89.
Google Scholar
There is no corresponding record for this reference.
154
Mark, J. E. Physical Properties of Polymers Handbook; Springer, 2007; Vol. 1076.
Google Scholar
There is no corresponding record for this reference.
155
Ma, R.; Luo, T. Pi1m: a benchmark database for polymer informatics. J. Chem. Inf. Model. 2020, 60 (10), 4684– 4690, DOI: 10.1021/acs.jcim.0c00726
Google Scholar
There is no corresponding record for this reference.
156
Walsh, D. J.; Zou, W.; Schneider, L.; Mello, R.; Deagen, M. E.; Mysona, J.; Lin, T.-S.; de Pablo, J. J.; Jensen, K. F.; Audus, D. J. Community resource for innovation in polymer technology (cript): a scalable polymer material data structure, 2023.
Google Scholar
There is no corresponding record for this reference.
157
Yan, C.; Li, G. The rise of machine learning in polymer discovery. Advanced Intelligent Systems 2023, 5 (4), 2200243 DOI: 10.1002/aisy.202200243
Google Scholar
There is no corresponding record for this reference.
158
Levine, D. S.; Shuaibi, M.; Spotte-Smith, E. W. C.; Taylor, M. G.; Hasyim, M. R.; Michel, K.; Batatia, I.; Csányi, G.; Dzamba, M.; Eastman, P. The open molecules 2025 (omol25) dataset, evaluations, and models, 2025, arXiv:2505.08762. arXiv.org e-Print archive https://arxiv.org/abs/2505.08762.
Google Scholar
There is no corresponding record for this reference.
159
Mark, J. E. Polymer Data Handbook, Oxford University Press, New York, NY, 2009.
Google Scholar
There is no corresponding record for this reference.
160
Hayashi, Y.; Shiomi, J.; Morikawa, J.; Yoshida, R. Radonpy: automated physical property calculation using all-atom classical molecular dynamics simulations for polymer informatics. npj Computat. Mater. 2022, 8 (1), 222, DOI: 10.1038/s41524-022-00906-4
Google Scholar
There is no corresponding record for this reference.
161
Ellis, B.; Smith, R. Polymers: A Property Database. CRC Press, 2008.
Google Scholar
There is no corresponding record for this reference.
162
Huan, T. D.; Mannodi-Kanakkithodi, A.; Kim, C.; Sharma, V.; Pilania, G.; Ramprasad, R. A polymer dataset for accelerated property prediction and design. Scientific data 2016, 3 (1), 1– 10, DOI: 10.1038/sdata.2016.12
Google Scholar
There is no corresponding record for this reference.
163
Pence, H. E.; Williams, A. Chemspider: an online chemical information resource, 2010.
Google Scholar
There is no corresponding record for this reference.
164
Polymer property predictor and database. https://pppdb.uchicago.edu/, 2019–2025.
Google Scholar
There is no corresponding record for this reference.
165
Campus plastic s database: Computer aided material preselection by uniform standards. https://www.campusplastics.com/, 1988–2025.
Google Scholar
There is no corresponding record for this reference.
166
Miccio, L. A.; Schwartz, G. A. From chemical structure to quantitative polymer properties prediction through convolutional neural networks. Polymer 2020, 193, 122341 DOI: 10.1016/j.polymer.2020.122341
Google Scholar
There is no corresponding record for this reference.
167
Xu, P.; Lu, T.; Ju, L.; Tian, L.; Li, M.; Lu, W. Machine learning aided design of polymer with targeted band gap based on dft computation. J. Phys. Chem. B 2021, 125 (2), 601– 611, DOI: 10.1021/acs.jpcb.0c08674
Google Scholar
There is no corresponding record for this reference.
168
Xiong, J.; Xiong, Z.; Chen, K.; Jiang, H.; Zheng, M. Graph neural networks for automated de novo drug design. Drug Discovery Today 2021, 26 (6), 1382– 1393, DOI: 10.1016/j.drudis.2021.02.011
Google Scholar
There is no corresponding record for this reference.
169
Medina, H.; Drake, R.; Farmer, C. Accelerated prediction of molecular properties for per-and polyfluoroalkyl substances using graph neural networks with adjacency-free message passing. Environ. Pollut. 2025, 382, 126705 DOI: 10.1016/j.envpol.2025.126705
Google Scholar
There is no corresponding record for this reference.
170
Tozzini, V. Coarse-grained models for proteins. Curr. Opin. Struct. Biol. 2005, 15 (2), 144– 150, DOI: 10.1016/j.sbi.2005.02.005
Google Scholar
There is no corresponding record for this reference.
171
Joshi, S. Y.; Deshmukh, S. A. A review of advancements in coarse-grained molecular dynamics simulations. Mol. Simul. 2021, 47 (10–11), 786– 803, DOI: 10.1080/08927022.2020.1828583
Google Scholar
There is no corresponding record for this reference.
172
Noid, W. G.; Liu, P.; Wang, Y.; Chu, J.-W.; Ayton, G. S.; Izvekov, S.; Andersen, H. C.; Voth, G. A. The multiscale coarse-graining method. ii. numerical implementation for coarse-grained molecular models. J. Chem. Phys. 2008, 128, 244115, DOI: 10.1063/1.2938857
Google Scholar
There is no corresponding record for this reference.
173
Pak, A. J.; Voth, G. A. Advances in coarse-grained modeling of macromolecular complexes. Curr. Opin. Struct. Biol. 2018, 52, 119– 126, DOI: 10.1016/j.sbi.2018.11.005
Google Scholar
There is no corresponding record for this reference.
174
Noid, W. G. Perspective: Coarse-grained models for biomolecular systems. J. Chem. Phys. 2013, 139, 090901, DOI: 10.1063/1.4818908
Google Scholar
There is no corresponding record for this reference.
175
Kmiecik, S.; Gront, D.; Kolinski, M.; Wieteska, L.; Dawid, A. E.; Kolinski, A. Coarse-grained protein models and their applications. Chem. Rev. 2016, 116 (14), 7898– 7936, DOI: 10.1021/acs.chemrev.6b00163
Google Scholar
There is no corresponding record for this reference.
176
Ingólfsson, H. I.; Lopez, C. A.; Uusitalo, J. J.; de Jong, D. H.; Gopal, S. M.; Periole, X.; Marrink, S. J. The power of coarse graining in biomolecular simulations. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2014, 4 (3), 225– 248, DOI: 10.1002/wcms.1169
Google Scholar
There is no corresponding record for this reference.
177
Noid, W.; Szukalo, R. J.; Kidder, K. M.; Lesniewski, M. C. Rigorous progress in coarse-graining. Annu. Rev. Phys. Chem. 2024, 75 (1), 21– 45, DOI: 10.1146/annurev-physchem-062123-010821
Google Scholar
There is no corresponding record for this reference.
178
Guenza, M. G. Everything you want to know about coarse-graining and never dared to ask: Macromolecules as a key example. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2025, 15 (2), e70022 DOI: 10.1002/wcms.70022
Google Scholar
There is no corresponding record for this reference.
179
Harmandaris, V. A.; Kremer, K. Predicting polymer dynamics at multiple length and time scales. Soft Matter 2009, 5 (20), 3920– 3926, DOI: 10.1039/b905361a
Google Scholar
There is no corresponding record for this reference.
180
Ye, H.; Xian, W.; Li, Y. Machine learning of coarse-grained models for organic molecules and polymers: Progress, opportunities, and challenges. ACS Omega 2021, 6 (3), 1758– 1772, DOI: 10.1021/acsomega.0c05321
Google Scholar
There is no corresponding record for this reference.
181
Noid, W. G. Perspective: Advances, challenges, and insight for predictive coarse-grained models. J. Phys. Chem. B 2023, 127 (19), 4174– 4207, DOI: 10.1021/acs.jpcb.2c08731
Google Scholar
There is no corresponding record for this reference.
182
Vettorel, T.; Besold, G.; Kremer, K. Fluctuating soft-sphere approach to coarse-graining of polymer models. Soft Matter 2010, 6 (10), 2282– 2292, DOI: 10.1039/b921159d
Google Scholar
There is no corresponding record for this reference.
183
Español, P.; Serrano, M.; Pagonabarraga, I.; Zúñiga, I. Energy-conserving coarse-graining of complex molecules. Soft Matter 2016, 12 (21), 4821– 4837, DOI: 10.1039/C5SM03038B
Google Scholar
There is no corresponding record for this reference.
184
Husic, B. E.; Charron, N. E.; Lemm, D.; Wang, J.; Pérez, A.; Majewski, M.; Krämer, A.; Chen, Y.; Olsson, S.; De Fabritiis, G.; Noe, F.; Clementi, C. Coarse graining molecular dynamics with graph neural networks J. Chem. Phys. 2020; Vol. 153, DOI: 10.1063/5.0026133 .
Google Scholar
There is no corresponding record for this reference.
185
Wang, X.; Wu, Y.; Zhang, A.; He, X.; Chua, T.-S. Towards multi-grained explainability for graph neural networks. Adv. Neural Inf. Process. Syst. 2021, 34, 18446– 18458
Google Scholar
There is no corresponding record for this reference.
186
Peter, C.; Kremer, K. Multiscale simulation of soft matter systems-from the atomistic to the coarse-grained level and back. Soft Matter 2009, 5 (22), 4357– 4366, DOI: 10.1039/b912027k
Google Scholar
There is no corresponding record for this reference.
187
Zhao, Y.; Chen, G.; Liu, J. Polymer data challenges in the ai era: Bridging gaps for next-generation energy materials, 2025, arXiv:2505.13494. arXiv.org e-Print archive https://arxiv.org/abs/2505.13494.
Google Scholar
There is no corresponding record for this reference.
188
Orio, M.; Pantazis, D. A.; Neese, F. Density functional theory. Photosynth. Res. 2009, 102, 443– 453, DOI: 10.1007/s11120-009-9404-8
Google Scholar
There is no corresponding record for this reference.
189
Joshi, S. P.; Bucsek, A.; Pagan, D. C.; Daly, S.; Ravindran, S.; Marian, J.; Bessa, M. A.; Kalidindi, S. R.; Admal, N. C.; Reina, C.; Ghosh, S.; Warren, J. A.; Vinals, J.; Tadmor, E. B. “ Integrated experiment and simulation co-design: A key infrastructure for predictive mesoscale materials modeling, 2025, arXiv:2503.09793. arXiv.org e-Print archive https://arxiv.org/abs/2503.09793.
Google Scholar
There is no corresponding record for this reference.
190
Zaporozhets, I.; Clementi, C. Multibody terms in protein coarse-grained models: A top-down perspective. J. Phys. Chem. B 2023, 127 (31), 6920– 6927, DOI: 10.1021/acs.jpcb.3c04493
Google Scholar
There is no corresponding record for this reference.
191
Jin, J.; Pak, A. J.; Durumeric, A. E.; Loose, T. D.; Voth, G. A. Bottom-up coarse-graining: Principles and perspectives. J. Chem. Theory Comput. 2022, 18 (10), 5759– 5791, DOI: 10.1021/acs.jctc.2c00643
Google Scholar
There is no corresponding record for this reference.
192
Praprotnik, M.; Delle Site, L.; Kremer, K. Multiscale simulation of soft matter: From scale bridging to adaptive resolution. Annu. Rev. Phys. Chem. 2008, 59, 545– 571, DOI: 10.1146/annurev.physchem.59.032607.093707
Google Scholar
There is no corresponding record for this reference.
193
Milano, G.; Kawakatsu, T. Hybrid particle-field molecular dynamics simulations for dense polymer systems J. Chem. Phys. 2009; Vol. 130, DOI: 10.1063/1.3142103 .
Google Scholar
There is no corresponding record for this reference.
194
Wang, J.; Han, Y.; Xu, Z.; Yang, X.; Ramakrishna, S.; Liu, Y. Dissipative particle dynamics simulation: A review on investigating mesoscale properties of polymer systems. Macromol. Mater. Eng. 2021, 306 (4), 2000724 DOI: 10.1002/mame.202000724
Google Scholar
There is no corresponding record for this reference.
195
Santo, K. P.; Neimark, A. V. Dissipative particle dynamics simulations in colloid and interface science: A review. Adv. Colloid Interface Sci. 2021, 298, 102545 DOI: 10.1016/j.cis.2021.102545
Google Scholar
There is no corresponding record for this reference.
196
Hoogerbrugge, P. J.; Koelman, J. Simulating microscopic hydrodynamic phenomena with dissipative particle dynamics. Europhys. Lett. 1992, 19 (3), 155, DOI: 10.1209/0295-5075/19/3/001
Google Scholar
There is no corresponding record for this reference.
197
Español, P.; Warren, P. Statistical mechanics of dissipative particle dynamics. Europhys. Lett. 1995, 30 (4), 191, DOI: 10.1209/0295-5075/30/4/001
Google Scholar
There is no corresponding record for this reference.
198
Marrink, S. J.; Monticelli, L.; Melo, M. N.; Alessandri, R.; Tieleman, D. P.; Souza, P. C. Two decades of martini: Better beads, broader scope. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2023, 13 (1), e1620 DOI: 10.1002/wcms.1620
Google Scholar
There is no corresponding record for this reference.
199
Alessandri, R.; Grünewald, F.; Marrink, S. J. The martini model in materials science. Adv. Mater. 2021, 33 (24), 2008635 DOI: 10.1002/adma.202008635
Google Scholar
There is no corresponding record for this reference.
200
Souza, P. C. T.; Alessandri, R.; Barnoud, J.; Thallmair, S.; Faustino, I.; Grünewald, F.; Patmanidis, I.; Abdizadeh, H.; Bruininks, B. M.; Wassenaar, T. A.; Kroon, P.; Melcr, K.; Nieto, V.; Corradi, V.; Khan, H. M.; Domanski, J.; Javanainen, M.; Martinez-Seara, H.; Reuter, N.; Best, R. B.; Vattulainen, I.; Monticelli, L.; Periole, X.; Peter, T. D.; Vries, dH.; Marrink, S. J. Martini 3: a general purpose force field for coarse-grained molecular dynamics. Nat. Methods 2021, 18 (4), 382– 388, DOI: 10.1038/s41592-021-01098-3
Google Scholar
There is no corresponding record for this reference.
201
Fredrickson, G. H. The Equilibrium Theory of Inhomogeneous Polymers. Oxford University Press, 2006.
Google Scholar
There is no corresponding record for this reference.
202
Schmid, F. Self-consistent-field theories for complex fluids. J. Phys.: Condens. Matter 1998, 10 (37), 8105, DOI: 10.1088/0953-8984/10/37/002
Google Scholar
There is no corresponding record for this reference.
203
Maerzke, K. A.; Siepmann, J. I. Transferable potentials for phase equilibria- coarse-grain description for linear alkanes. J. Phys. Chem. B 2011, 115 (13), 3452– 3465, DOI: 10.1021/jp1063935
Google Scholar
There is no corresponding record for this reference.
204
Liwo, A.; Ołdziej, S.; Pincus, M. R.; Wawak, R. J.; Rackovsky, S.; Scheraga, H. A. A united-residue force field for off-lattice protein-structure simulations. i. functional forms and parameters of long-range side-chain interaction potentials from protein crystal data. J. Comput. Chem. 1997, 18 (7), 849– 873, DOI: 10.1002/(SICI)1096-987X(199705)18:7<849::AID-JCC1>3.0.CO;2-R
Google Scholar
There is no corresponding record for this reference.
205
Maisuradze, G. G.; Senet, P.; Czaplewski, C.; Liwo, A.; Scheraga, H. A. Investigation of protein folding by coarse-grained molecular dynamics with the unres force field. J. Phys. Chem. A 2010, 114 (13), 4471– 4485, DOI: 10.1021/jp9117776
Google Scholar
There is no corresponding record for this reference.
206
Reith, D.; Pütz, M.; Müller-Plathe, F. Deriving effective mesoscale potentials from atomistic simulations. J. Comput. Chem. 2003, 24 (13), 1624– 1636, DOI: 10.1002/jcc.10307
Google Scholar
There is no corresponding record for this reference.
207
Agrawal, V.; Arya, G.; Oswald, J. Simultaneous iterative boltzmann inversion for coarse-graining of polyurea. Macromolecules 2014, 47 (10), 3378– 3389, DOI: 10.1021/ma500320n
Google Scholar
There is no corresponding record for this reference.
208
Tschöp, W.; Kremer, K.; Batoulis, J.; Bürger, T.; Hahn, O. Simulation of polymer melts. i. coarse-graining procedure for polycarbonates. Acta Polym. 1998, 49 (2–3), 61– 74, DOI: 10.1002/(SICI)1521-4044(199802)49:2/3<61::AID-APOL61>3.0.CO;2-V
Google Scholar
There is no corresponding record for this reference.
209
McCoy, J. D.; Curro, J. G. Mapping of explicit atom onto united atom potentials. Macromolecules 1998, 31 (26), 9362– 9368, DOI: 10.1021/ma981060g
Google Scholar
There is no corresponding record for this reference.
210
Baschnagel, J.; Binder, K.; Doruker, P.; Gusev, A. A.; Hahn, O.; Kremer, K.; Mattice, W. L.; Müller-Plathe, F.; Murat, M.; Paul, W.; Santos, S.; Suter, U. W.; Tries, V. Bridging the gap between atomistic and coarse-grained models of polymers: Status and perspectives. Viscoelasticity, Atomistic Models, Statistical Chemistry 2000, 152, 41– 156, DOI: 10.1007/3-540-46778-5_2
Google Scholar
There is no corresponding record for this reference.
211
Akkermans, R. L. C.; Briels, W. J. A structure-based coarse-grained model for polymer melts. J. Chem. Phys. 2001, 114 (2), 1020– 1031, DOI: 10.1063/1.1330744
Google Scholar
There is no corresponding record for this reference.
212
Uneyama, T.; Masubuchi, Y. Plateau moduli of several single-chain slip-link and slip-spring models. Macromolecules 2021, 54 (3), 1338– 1353, DOI: 10.1021/acs.macromol.0c01790
Google Scholar
There is no corresponding record for this reference.
213
Behbahani, A. F.; Schneider, L.; Rissanou, A.; Chazirakis, A.; Bacova, P.; Jana, P. K.; Li, W.; Doxastakis, M.; Polinska, P.; Burkhart, C.; Muller, M.; Harmandaris, V. A. Dynamics and rheology of polymer melts via hierarchical atomistic, coarse-grained, and slip-spring simulations. Macromolecules 2021, 54 (6), 2740– 2762, DOI: 10.1021/acs.macromol.0c02583
Google Scholar
There is no corresponding record for this reference.
214
Wu, Z.; Müller-Plathe, F. Slip-spring hybrid particle-field molecular dynamics for coarse-graining branched polymer melts: Polystyrene melts as an example. J. Chem. Theory Comput. 2022, 18 (6), 3814– 3828, DOI: 10.1021/acs.jctc.2c00107
Google Scholar
There is no corresponding record for this reference.
215
Underhill, P. T.; Doyle, P. S. On the coarse-graining of polymers into bead-spring chains. J. Non-Newtonian Fluid Mech. 2004, 122 (1–3), 3– 31, DOI: 10.1016/j.jnnfm.2003.10.006
Google Scholar
There is no corresponding record for this reference.
216
Kremer, K.; Grest, G. S. Dynamics of entangled linear polymer melts: A molecular-dynamics simulation. J. Chem. Phys. 1990, 92 (8), 5057– 5086, DOI: 10.1063/1.458541
Google Scholar
There is no corresponding record for this reference.
217
Everaers, R.; Karimi-Varzaneh, H. A.; Fleck, F.; Hojdis, N.; Svaneborg, C. Kremer-grest models for commodity polymer melts: Linking theory, experiment, and simulation at the kuhn scale. Macromolecules 2020, 53 (6), 1901– 1916, DOI: 10.1021/acs.macromol.9b02428
Google Scholar
There is no corresponding record for this reference.
218
Nitta, H.; Ozawa, T.; Yasuoka, K. Construction of full-atomistic polymer amorphous structures using reverse-mapping from kremer–grest models J. Chem. Phys. , 2023 159 no. 19.
Google Scholar
There is no corresponding record for this reference.
219
Miwatani, R.; Takahashi, K. Z.; Arai, N. Performance of coarse graining in estimating polymer properties: Comparison with the atomistic model. Polymers 2020, 12 (2), 382, DOI: 10.3390/polym12020382
Google Scholar
There is no corresponding record for this reference.
220
Praprotnik, M.; Delle Site, L.; Kremer, K. Adaptive resolution scheme for efficient hybrid atomistic-mesoscale molecular dynamics simulations of dense liquids. Phys. Rev. E 2006, 73 (6), 066701 DOI: 10.1103/PhysRevE.73.066701
Google Scholar
There is no corresponding record for this reference.
221
Darré, L.; Machado, M. R.; Brandner, A. F.; González, H. C.; Ferreira, S.; Pantano, S. Sirah: a structurally unbiased coarse-grained force field for proteins with aqueous solvation and long-range electrostatics. J. Chem. Theory Comput. 2015, 11 (2), 723– 739, DOI: 10.1021/ct5007746
Google Scholar
There is no corresponding record for this reference.
222
Das, R.; Baker, D. Macromolecular modeling with rosetta. Annu. Rev. Biochem. 2008, 77, 363– 382, DOI: 10.1146/annurev.biochem.77.062906.171838
Google Scholar
There is no corresponding record for this reference.
223
Leman, J. K.; Weitzner, B. D.; Lewis, S. M.; Adolf-Bryfogle, J.; Alam, N.; Alford, R. F.; Aprahamian, M.; Baker, D.; Barlow, K. A.; Barth, P.; Basanta, B.; Bender, B. J.; Blacklock, K.; Bonet, J.; Boyken, S. E.; Bradley, P.; Bystroff, C.; Conway, P.; Cooper, S.; Correia, B. E.; Coventry, B.; Das, R.; De Jong, R. M.; DiMaio, F.; Dsilva, L.; Dunbrack, R.; Ford, A. S.; Frenz, B.; Fu, D. Y.; Geniesse, C.; Goldschmidt, L.; Gowthaman, R.; Gray, J. J.; Gront, D.; Guffy, S.; Horowitz, S.; Huang, P.-S.; Huber, T.; Jacobs, T. M.; Jeliazkov, J. R.; Johnson, D. K.; Kappel, K.; Karanicolas, J.; Khakzad, H.; Khar, K. R.; Khare, S. D.; Khatib, F.; Khramushin, A.; King, I. C.; Kleffner, R.; Koepnick, B.; Kortemme, T.; Kuenze, G.; Kuhlman, B.; Kuroda, D.; Labonte, J. W.; Lai, J. K.; Lapidoth, G.; Leaver-Fay, A.; Lindert, S.; Linsky, T.; London, N.; Lubin, J. H.; Lyskov, S.; Maguire, J.; Malmström, L.; Marcos, E.; Marcu, O.; Marze, N. A.; Meiler, J.; Moretti, R.; Mulligan, V. K.; Nerli, S.; Norn, C.; Ó’Conchúir, S.; Ollikainen, N.; Ovchinnikov, S.; Pacella, M. S.; Pan, X.; Park, H.; Pavlovicz, R. E.; Pethe, M.; Pierce, B. G.; Pilla, K. B.; Raveh, B.; Renfrew, P. D.; Roy Burman, S. S.; Rubenstein, A.; Sauer, M. F.; Scheck, A.; Schief, W.; Schueler-Furman, O.; Sedan, Y.; Sevy, A. M.; Sgourakis, N. G.; Shi, L.; Siegel, J. B.; Silva, D.-A.; Smith, S.; Song, Y.; Stein, A.; Szegedy, M.; Teets, F. D.; Thyme, S. B.; Wang, R. Y.-R.; Watkins, A.; Zimmerman, L.; Bonneau, R. Macromolecular modeling and design in rosetta: recent methods and frameworks. Nat. Methods 2020, 17 (7), 665– 680, DOI: 10.1038/s41592-020-0848-2
Google Scholar
There is no corresponding record for this reference.
224
Baldassarre, F.; Azizpour, H. Explainability techniques for graph convolutional networks, 2019, arXiv:1905.13686 arXiv.org e-Print archive https://arxiv.org/abs/1905.13686.
Google Scholar
There is no corresponding record for this reference.
225
Bai, Y.; Wilbraham, L.; Slater, B. J.; Zwijnenburg, M. A.; Sprick, R. S.; Cooper, A. I. Accelerated discovery of organic polymer photocatalysts for hydrogen evolution from water through the integration of experiment and theory. J. Am. Chem. Soc. 2019, 141 (22), 9063– 9071, DOI: 10.1021/jacs.9b03591
Google Scholar
There is no corresponding record for this reference.
226
Vosko, S. H.; Wilk, L.; Nusair, M. Accurate spin-dependent electron liquid correlation energies for local spin density calculations: a critical analysis. Can. J. Phys. 1980, 58 (8), 1200– 1211, DOI: 10.1139/p80-159
Google Scholar
There is no corresponding record for this reference.
227
Becke, A. D. Density-functional thermochemistry. iii. the role of exact exchange. J. Chem. Phys. 1993, 98 (7), 5648– 5652, DOI: 10.1063/1.464913
Google Scholar
There is no corresponding record for this reference.
228
Lee, C.; Yang, W.; Parr, R. G. Development of the colle-salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B 1988, 37 (2), 785, DOI: 10.1103/PhysRevB.37.785
Google Scholar
There is no corresponding record for this reference.
229
Grimme, S.; Bannwarth, C.; Shushkov, P. A robust and accurate tight-binding quantum chemical method for structures, vibrational frequencies, and noncovalent interactions of large molecular systems parametrized for all spd-block elements (z= 1–86). J. Chem. Theory Comput. 2017, 13 (5), 1989– 2009, DOI: 10.1021/acs.jctc.7b00118
Google Scholar
There is no corresponding record for this reference.
230
Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. Ab initio calculation of vibrational absorption and circular dichroism spectra using density functional force fields. J. Phys. Chem. A 1994, 98 (45), 11623– 11627, DOI: 10.1021/j100096a001
Google Scholar
There is no corresponding record for this reference.
231
Yang, K.; Swanson, K.; Jin, W.; Coley, C.; Eiden, P.; Gao, H.; Guzman-Perez, A.; Hopper, T.; Kelley, B.; Mathea, M. Analyzing learned molecular representations for property prediction. J. Chem. Inf. Model. 2019, 59 (8), 3370– 3388, DOI: 10.1021/acs.jcim.9b00237
Google Scholar
There is no corresponding record for this reference.
232
Rogers, D.; Hahn, M. Extended-connectivity fingerprints. J. Chem. Inf. Model. 2010, 50 (5), 742– 754, DOI: 10.1021/ci100050t
Google Scholar
There is no corresponding record for this reference.
233
Landrum, G. Rdkit: open-source cheminformatics from machine learning to chemical registration. In Abstracts of Papers of the American Chemical Society; Vol. 258, AMER CHEMICAL SOC 1155 16TH ST, NW, WASHINGTON, DC 20036 USA, 2019.
Google Scholar
There is no corresponding record for this reference.
234
Mannodi-Kanakkithodi, A.; Chandrasekaran, A.; Kim, C.; Huan, T. D.; Pilania, G.; Botu, V.; Ramprasad, R. Scoping the polymer genome: A roadmap for rational polymer dielectrics design and beyond. Mater. Today 2018, 21, 785– 796, DOI: 10.1016/j.mattod.2017.11.021
Google Scholar
There is no corresponding record for this reference.
235
Kim, C.; Chandrasekaran, A.; Huan, T. D.; Das, D.; Ramprasad, R. Polymer genome: A data-powered polymer informatics platform for property predictions. J. Phys. Chem. C 2018, 122, 17575– 17585, DOI: 10.1021/acs.jpcc.8b02913
Google Scholar
There is no corresponding record for this reference.
236
Ward, L.; Dunn, A.; Faghaninia, A.; Zimmermann, N. E.; Bajaj, S.; Wang, Q.; Montoya, J.; Chen, J.; Bystrom, K.; Dylla, M.; Chard, K.; Asta, M.; Persson, K. A.; Snyder, G. J.; Foster, I.; Jain, A. Matminer: An open source toolkit for materials data mining. Comput. Mater. Sci. 2018, 152, 60– 69, DOI: 10.1016/j.commatsci.2018.05.018
Google Scholar
There is no corresponding record for this reference.
237
Lightstone, J. P.; Chen, L.; Kim, C.; Batra, R.; Ramprasad, R. Refractive index prediction models for polymers using machine learning J. Appl. Phys. 2020; Vol. 127, DOI: 10.1063/5.0008026 .
Google Scholar
There is no corresponding record for this reference.
238
Chen, L.; Kim, C.; Batra, R.; Lightstone, J. P.; Wu, C.; Li, Z.; Deshmukh, A. A.; Wang, Y.; Tran, H. D.; Vashishta, P.; Sotzing, G. A.; Cao, Y.; Ramprasad, R. Frequency-dependent dielectric constant prediction of polymers using machine learning. npj Computat. Mater. 2020, 6 (1), 61, DOI: 10.1038/s41524-020-0333-6
Google Scholar
There is no corresponding record for this reference.
239
Doan Tran, H.; Kim, C.; Chen, L.; Chandrasekaran, A.; Batra, R.; Venkatram, S.; Kamal, D.; Lightstone, J. P.; Gurnani, R.; Shetty, P. Machine-learning predictions of polymer properties with polymer genome. J. Appl. Phys. 2020, 128 (17), 171104 DOI: 10.1063/5.0023759
Google Scholar
There is no corresponding record for this reference.
240
Godwin, J.; Schaarschmidt, M.; Gaunt, A.; Sanchez-Gonzalez, A.; Rubanova, Y.; Veličković, P.; Kirkpatrick, J.; Battaglia, P. Simple gnn regularisation for 3d molecular property prediction & beyond, 2021, arXiv:2106.07971. arXiv.org e-Print archive https://arxiv.org/abs/2106.07971.
Google Scholar
There is no corresponding record for this reference.
241
Chen, D.; Lin, Y.; Li, W.; Li, P.; Zhou, J.; Sun, X. Measuring and relieving the over-smoothing problem for graph neural networks from the topological view. Proc. AAAI Conf. Artificial Intelligence 2020, 34, 3438– 3445, DOI: 10.1609/aaai.v34i04.5747
Google Scholar
There is no corresponding record for this reference.
242
Frisch, M. Gaussian 09, Revision D.01, Gaussian , 2009.
Google Scholar
There is no corresponding record for this reference.
243
Kuenneth, C.; Schertzer, W.; Ramprasad, R. Copolymer informatics with multitask deep neural networks. Macromolecules 2021, 54 (13), 5957– 5961, DOI: 10.1021/acs.macromol.1c00728
Google Scholar
There is no corresponding record for this reference.
244
You, J.; Ying, Z.; Leskovec, J. Design space for graph neural networks. In Advances in Neural Information Processing Systems (NeurIPS) 2020.
Google Scholar
There is no corresponding record for this reference.
245
Rupp, M.; Tkatchenko, A.; Müller, K.-R.; Von Lilienfeld, O. A. Fast and accurate modeling of molecular atomization energies with machine learning. Phys. Rev. Lett. 2012, 108 (5), 058301 DOI: 10.1103/PhysRevLett.108.058301
Google Scholar
There is no corresponding record for this reference.
246
Bartók, A. P.; Kondor, R.; Csányi, G. On representing chemical environments. Phys. Rev. B 2013, 87 (18), 184115 DOI: 10.1103/PhysRevB.87.184115
Google Scholar
There is no corresponding record for this reference.
247
Townsend, J.; Micucci, C. P.; Hymel, J. H.; Maroulas, V.; Vogiatzis, K. D. Representation of molecular structures with persistent homology for machine learning applications in chemistry. Nat. Commun. 2020, 11 (1), 3230 DOI: 10.1038/s41467-020-17035-5
Google Scholar
There is no corresponding record for this reference.
248
Huo, H.; Rupp, M. Unified representation of molecules and crystals for machine learning. Machine Learning: Sci. Technol. 2022, 3 (4), 045017 DOI: 10.1088/2632-2153/aca005
Google Scholar
There is no corresponding record for this reference.
249
Sanchez Medina, E. I.; Linke, S.; Stoll, M.; Sundmacher, K. Gibbs-helmholtz graph neural network: capturing the temperature dependency of activity coefficients at infinite dilution. Digital Discovery 2023, 2 (3), 781– 798, DOI: 10.1039/D2DD00142J
Google Scholar
There is no corresponding record for this reference.
250
Zhong, C.; Sato, Y.; Masuoka, H.; Chen, X. Improvement of predictive accuracy of the unifac model for vapor-liquid equilibria of polymer solutions. Fluid Phase Equilib. 1996, 123 (1–2), 97– 106, DOI: 10.1016/S0378-3812(96)90015-1
Google Scholar
There is no corresponding record for this reference.
251
Kontogeorgis, G. M.; Fredenslund, A.; Tassios, D. P. Simple activity coefficient model for the prediction of solvent activities in polymer solutions. Ind. Eng. Chem. Res. 1993, 32 (2), 362– 372, DOI: 10.1021/ie00014a013
Google Scholar
There is no corresponding record for this reference.
252
Pappa, G. D.; Voutsas, E. C.; Tassios, D. P. Prediction of activity coefficients in polymer and copolymer solutions using simple activity coefficient models. Ind. Eng. Chem. Res. 1999, 38 (12), 4975– 4984, DOI: 10.1021/ie990265q
Google Scholar
There is no corresponding record for this reference.
253
Cremer, D. Møller-plesset perturbation theory: from small molecule methods to methods for thousands of atoms. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2011, 1 (4), 509– 530, DOI: 10.1002/wcms.58
Google Scholar
There is no corresponding record for this reference.
254
Kearnes, S.; McCloskey, K.; Berndl, M.; Pande, V.; Riley, P. Molecular graph convolutions: moving beyond fingerprints. J. Comput.-Aided Mol. Des. 2016, 30, 595– 608, DOI: 10.1007/s10822-016-9938-8
Google Scholar
There is no corresponding record for this reference.
255
Wang, M. Y. Deep graph library: Towards efficient and scalable deep learning on graphs. In ICLR Workshop on Representation Learning on Graphs and Manifolds , 2019.
Google Scholar
There is no corresponding record for this reference.
256
Bojar, D.; Powers, R. K.; Camacho, D. M.; Collins, J. J. Deep-learning resources for studying glycan-mediated host-microbe interactions. Cell Host Microbe 2021, 29, 132– 144.e3, DOI: 10.1016/j.chom.2020.10.004
Google Scholar
There is no corresponding record for this reference.
257
Montavon, G.; Binder, A.; Lapuschkin, S.; Samek, W.; Müller, K.-R. Layer-wise relevance propagation: an overview. In Lecture Notes in Computer Science 2019; pp 193– 209.
Google Scholar
There is no corresponding record for this reference.
258
Binder, A.; Bach, S.; Montavon, G.; Müller, K.-R.; Samek, W. Layer-wise relevance propagation for deep neural network architectures. In Information Science and Applications (ICISA) 2016; Springer, 2016; pp 913– 922.
Google Scholar
There is no corresponding record for this reference.
259
Adebayo, J.; Gilmer, J.; Muelly, M.; Goodfellow, I.; Hardt, M.; Kim, B. Sanity checks for saliency maps Advances in Neural Information Processing Systems , 2018 31.
Google Scholar
There is no corresponding record for this reference.
260
Qi, Z.; Khorram, S.; Li, F. Visualizing deep networks by optimizing with integrated gradients. AAAI 2020, 34, 11890– 11898, DOI: 10.1609/aaai.v34i07.6863
Google Scholar
There is no corresponding record for this reference.

Cited By

Click to copy section linkSection link copied!

This article has not yet been cited by other publications.

Get e-Alerts

Journal of Chemical Information and Modeling

Cite this: J. Chem. Inf. Model. 2026, 66, 3, 1316–1336

Click to copy citationCitation copied!

https://doi.org/10.1021/acs.jcim.5c02421

Published January 28, 2026

CC-BY 4.0 .

Article Views

Altmetric

Citations

Learn about these metrics

Article Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.

Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.

The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated.

Abstract
High Resolution Image
Download MS PowerPoint Slide
Figure 1
Figure 1. General architecture of a GNN for molecular property prediction. The input molecular graph is processed through multiple message passing layers, where the node embeddings are iteratively updated by propagating information from neighboring atoms. These updated representations are then combined by a readout function (e.g., average or sum) to produce a graph-level embedding, which is fed into a downstream prediction layer.
High Resolution Image
Download MS PowerPoint Slide
Figure 2
Figure 2. Two graph augmentations in polymer GNN architectures: a virtual edges connecting distant nodes to encode periodic or long-range interactions (left), and a global node connected to all nodes to capture system-level or graph-wide information (right).
High Resolution Image
Download MS PowerPoint Slide
Figure 3
Figure 3. An illustration of GAT operations. The left portion of the diagram shows how attention coefficients α_ij are computed: the transformed node features Wh_i and Wh_j are concatenated and passed through a shared attention mechanism parametrized by the vector a, followed by a LeakyReLU nonlinearity and a softmax normalization over all neighbors j ∈ N_i. The resulting coefficient α_ij encodes the learned importance of neighbor j when updating node i. The right portion illustrates multihead attention for a specific node (node 1 in the drawing): each neighbor contributes messages through multiple attention heads (depicted by the green, blue and purple lines), producing head-specific weights α_1j^(k) and transformed messages W^(k)h_j. The index k refers to the attention head number, where each head has its own learnable weight matrix W^(k) and produces its own attention score. These per-head aggregations are combined via concatenation (for intermediate layers) or averaging (for final layers) to yield the updated node embedding h_i. The diagram highlights both key components of GATs: attention-based neighborhood weighting and multihead message aggregation.
High Resolution Image
Download MS PowerPoint Slide
Figure 4
Figure 4. Conceptual depiction of CG in polymer informatics. The top bar represents the (abstract) spectrum of CG strategies from macroscopic, geometry-based top-down strategies to atomistic, energetics-based bottom-up approaches. The molecular structure is overlaid with color-coded regions representing multiple levels of abstraction: atoms (gray), monomer/repeat units (red), substructures, functional groups or motifs (blue), small domains (purple), and large domains (yellow). This hierarchy also illustrates the conceptual transition from atomistic to increasingly coarse-grained graph representations, by grouping chemically or structurally important substructures so that scalable and interpretable ML pipelines are achievable.
High Resolution Image
Download MS PowerPoint Slide
References
This article references 260 other publications.
1. 1
  Liang, Y.; Yu, L. Development of semiconducting polymers for solar energy harvesting. Polym. Rev. 2010, 50 (4), 454– 473, DOI: 10.1080/15583724.2010.515765
  There is no corresponding record for this reference.
2. 2
  Baytekin, B.; Baytekin, H. T.; Grzybowski, B. A. Retrieving and converting energy from polymers: deployable technologies and emerging concepts. Energy Environ. Sci. 2013, 6 (12), 3467– 3482, DOI: 10.1039/c3ee41360h
  There is no corresponding record for this reference.
3. 3
  Abdelhamid, M. E.; O’Mullane, A. P.; Snook, G. A. Storing energy in plastics: a review on conducting polymers & their role in electrochemical energy storage. RSC Adv. 2015, 5 (15), 11611– 11626, DOI: 10.1039/C4RA15947K
  There is no corresponding record for this reference.
4. 4
  Fan, J.; Njuguna, J. An introduction to lightweight composite materials and their use in transport structures. In Lightweight Composite Structures in Transport; Elsevier, 2016; pp 3– 34.
  There is no corresponding record for this reference.
5. 5
  Singh, J.; Srivastawa, K.; Jana, S.; Dixit, C.; Ravichandran, S. Advancements in lightweight materials for aerospace structures: A comprehensive review. Acceleron Aerosp. J. 2024, 2 (3), 173– 183, DOI: 10.61359/11.2106-2409
  There is no corresponding record for this reference.
6. 6
  Yeo, S. J.; Oh, M. J.; Yoo, P. J. Structurally controlled cellular architectures for high-performance ultra-lightweight materials. Adv. Mater. 2019, 31 (34), 1803670 DOI: 10.1002/adma.201803670
  There is no corresponding record for this reference.
7. 7
  Cui, H.; Zhao, Q.; Zhang, L.; Du, X. Intelligent polymer-based bioinspired actuators: from monofunction to multifunction. Adv. Intell. Syst. 2020, 2 (11), 2000138 DOI: 10.1002/aisy.202000138
  There is no corresponding record for this reference.
8. 8
  Weng, M.; Ding, M.; Zhou, P.; Ye, Y.; Luo, Z.; Ye, X.; Guo, Q.; Chen, L. Multi-functional and integrated actuators made with bio-inspired cobweb carbon nanotube-polymer composites. Chem. Eng. J. 2023, 452, 139146 DOI: 10.1016/j.cej.2022.139146
  There is no corresponding record for this reference.
9. 9
  Lendlein, A.; Rehahn, M.; Buchmeiser, M. R.; Haag, R. Polymers in biomedicine and electronics. Macromol. Rapid Commun. 2010, 31 (17), 1487– 1491, DOI: 10.1002/marc.201000426
  There is no corresponding record for this reference.
10. 10
  Medina, H.; Child, N. A review of developments in carbon-based nanocomposite electrodes for noninvasive electroencephalography. Sensors 2025, 25 (7), 2274 DOI: 10.3390/s25072274
  There is no corresponding record for this reference.
11. 11
  Farmer, C.; Medina, H. Effects of electrostriction on the bifurcated electro-mechanical performance of conical dielectric elastomer actuators and sensors. Robotica 2023, 41 (1), 215– 235, DOI: 10.1017/S0263574722001254
  There is no corresponding record for this reference.
12. 12
  Medina, H.; Farmer, C.; Liu, I. Dielectric elastomer-based actuators: a modeling and control review for non-experts. In Actuators; MDPI, 2024; Vol. 13, p 151.
  There is no corresponding record for this reference.
13. 13
  Korn, D.; Farmer, C.; Medina, H. A detailed solution framework for the out-of-plane displacement of circular dielectric elastomer actuators. Eng. Rep. 2022, 4 (1), e12442 DOI: 10.1002/eng2.12442
  There is no corresponding record for this reference.
14. 14
  Peterson, G. I.; Choi, T.-L. Cascade polymerizations: recent developments in the formation of polymer repeat units by cascade reactions. Chem. Sci. 2020, 11 (19), 4843– 4854, DOI: 10.1039/D0SC01475C
  There is no corresponding record for this reference.
15. 15
  Cao, C.; Lin, Y. Correlation between the glass transition temperatures and repeating unit structure for high molecular weight polymers. J. Chem. Inf. Comput. Sci. 2003, 43 (2), 643– 650, DOI: 10.1021/ci0202990
  There is no corresponding record for this reference.
16. 16
  Tezuka, Y.; Oike, H. Topological polymer chemistry: systematic classification of nonlinear polymer topologies. J. Am. Chem. Soc. 2001, 123 (47), 11570– 11576, DOI: 10.1021/ja0114409
  There is no corresponding record for this reference.
17. 17
  Gu, Y.; Zhao, J.; Johnson, J. A. A (macro) molecular-level understanding of polymer network topology. Trends Chem. 2019, 1 (3), 318– 334, DOI: 10.1016/j.trechm.2019.02.017
  There is no corresponding record for this reference.
18. 18
  Li, Y.; Abberton, B. C.; Kröger, M.; Liu, W. K. Challenges in multiscale modeling of polymer dynamics. Polymers 2013, 5 (2), 751– 832, DOI: 10.3390/polym5020751
  There is no corresponding record for this reference.
19. 19
  Schmid, F. Understanding and modeling polymers: The challenge of multiple scales. ACS Polymers Au 2023, 3 (1), 28– 58, DOI: 10.1021/acspolymersau.2c00049
  There is no corresponding record for this reference.
20. 20
  Jones, R. Density functional theory: Its origins, rise to prominence, and future. Rev. Mod. Phys. 2015, 87 (3), 897, DOI: 10.1103/RevModPhys.87.897
  There is no corresponding record for this reference.
21. 21
  Parr, R. G. Density-Functional Theory of Atoms and Molecules; Oxford University Press, 1989.
  There is no corresponding record for this reference.
22. 22
  Karplus, M.; McCammon, J. A. Molecular dynamics simulations of biomolecules. Nat. Struct. Biol. 2002, 9 (9), 646– 652, DOI: 10.1038/nsb0902-646
  There is no corresponding record for this reference.
23. 23
  Hollingsworth, S. A.; Dror, R. O. Molecular dynamics simulation for all. Neuron 2018, 99 (6), 1129– 1143, DOI: 10.1016/j.neuron.2018.08.011
  There is no corresponding record for this reference.
24. 24
  Tuckerman, M. E.; Martyna, G. J. Understanding modern molecular dynamics: Techniques and applications. J. Phys. Chem. B 2000, 104, 159– 178, DOI: 10.1021/jp992433y
  There is no corresponding record for this reference.
25. 25
  Bhavikatti, S. Finite Element Analysis. New Age International, 2005.
  There is no corresponding record for this reference.
26. 26
  Kim, N.-H.; Sankar, B. V.; Kumar, A. V. Introduction to Finite Element Analysis and Design; John Wiley & Sons, 2018.
  There is no corresponding record for this reference.
27. 27
  Farmer, C.; Medina, H. Hyperelastics. jl: A julia package for hyperelastic material modelling with a large collection of models. J. Open Source Softw. 2024, 9 (96), 6314 DOI: 10.21105/joss.06314
  There is no corresponding record for this reference.
28. 28
  Farmer, C.; Medina, H. Impact-4ccs: Integrated modeling and prediction using ab initio and trained potentials for collision cross sections. J. Comput. Chem. 2025, 46 (11), e70106 DOI: 10.1002/jcc.70106
  There is no corresponding record for this reference.
29. 29
  Medina, H.; Farmer, C. Current challenges in monitoring low contaminant levels of per-and polyfluoroalkyl substances in water matrices in the field. Toxics 2024, 12 (8), 610 DOI: 10.3390/toxics12080610
  There is no corresponding record for this reference.
30. 30
  Farmer, C.; Medina, H. Universal interatomic potentials with dft for understanding orbital localization in polydimethylsiloxane-amorphous silica nanocomposites. ACS Omega 2025, 10 (49), 60812– 60818, DOI: 10.1021/acsomega.5c09273
  There is no corresponding record for this reference.
31. 31
  McCormick, T. M.; Bridges, C. R.; Carrera, E. I.; DiCarmine, P. M.; Gibson, G. L.; Hollinger, J.; Kozycz, L. M.; Seferos, D. S. Conjugated polymers: Evaluating dft methods for more accurate orbital energy modeling. Macromolecules 2013, 46 (10), 3879– 3886, DOI: 10.1021/ma4005023
  There is no corresponding record for this reference.
32. 32
  Adekoya, O. C.; Adekoya, G. J.; Sadiku, E. R.; Hamam, Y.; Ray, S. S. Application of dft calculations in designing polymer-based drug delivery systems: An overview. Pharmaceutics 2022, 14 (9), 1972, DOI: 10.3390/pharmaceutics14091972
  There is no corresponding record for this reference.
33. 33
  Cohen, A. J.; Mori-Sánchez, P.; Yang, W. Challenges for density functional theory. Chem. Rev. 2012, 112 (1), 289– 320, DOI: 10.1021/cr200107z
  There is no corresponding record for this reference.
34. 34
  Kohn, W.; Becke, A. D.; Parr, R. G. Density functional theory of electronic structure. J. Phys. Chem. A 1996, 100 (31), 12974– 12980, DOI: 10.1021/jp960669l
  There is no corresponding record for this reference.
35. 35
  Dawson, W.; Degomme, A.; Stella, M.; Nakajima, T.; Ratcliff, L. E.; Genovese, L. Density functional theory calculations of large systems: Interplay between fragments, observables, and computational complexity. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2022, 12 (3), e1574 DOI: 10.1002/wcms.1574
  There is no corresponding record for this reference.
36. 36
  Bishop, M.; Kalos, M.; Frisch, H. Molecular dynamics of polymeric systems. J. Chem. Phys. 1979, 70 (3), 1299– 1304, DOI: 10.1063/1.437567
  There is no corresponding record for this reference.
37. 37
  Han, J.; Gee, R. H.; Boyd, R. H. Glass transition temperatures of polymers from molecular dynamics simulations. Macromolecules 1994, 27 (26), 7781– 7784, DOI: 10.1021/ma00104a036
  There is no corresponding record for this reference.
38. 38
  Lau, D.; Jian, W.; Yu, Z.; Hui, D. Nano-engineering of construction materials using molecular dynamics simulations: Prospects and challenges. Compos. Part B 2018, 143, 282– 291, DOI: 10.1016/j.compositesb.2018.01.014
  There is no corresponding record for this reference.
39. 39
  Mackerle, J. Finite element analysis and simulation of polymersan addendum: a bibliography (1996–2002). Modell. Simul. Mater. Sci. Eng. 2003, 11 (2), 195 DOI: 10.1088/0965-0393/11/2/307
  There is no corresponding record for this reference.
40. 40
  Zeng, X.; Brown, L. P.; Endruweit, A.; Matveev, M.; Long, A. C. Geometrical modelling of 3d woven reinforcements for polymer composites: Prediction of fabric permeability and composite mechanical properties. Compos. Part A: Appl. Sci. Manuf. 2014, 56, 150– 160, DOI: 10.1016/j.compositesa.2013.10.004
  There is no corresponding record for this reference.
41. 41
  Alshahrani, H. Characterization and finite element modeling of coupled properties during polymer composites forming processes. Mech. Mater. 2020, 144, 103370 DOI: 10.1016/j.mechmat.2020.103370
  There is no corresponding record for this reference.
42. 42
  Thompson, A. P.; Aktulga, H. M.; Berger, R.; Bolintineanu, D. S.; Brown, W. M.; Crozier, P. S.; In’t Veld, P. J.; Kohlmeyer, A.; Moore, S. G.; Nguyen, T. D.; Shan, R.; Stevens, M. J.; Tranchida, J.; Trott, C.; Plimpton, S. J. Lammps-a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales. Comput. Phys. Commun. 2022, 271, 108171 DOI: 10.1016/j.cpc.2021.108171
  There is no corresponding record for this reference.
43. 43
  Ratcliff, L. E.; Mohr, S.; Huhs, G.; Deutsch, T.; Masella, M.; Genovese, L. Challenges in large scale quantum mechanical calculations. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2017, 7 (1), e1290 DOI: 10.1002/wcms.1290
  There is no corresponding record for this reference.
44. 44
  Fredrickson, G. H. Computational field theory of polymers: opportunities and challenges. Soft Matter 2007, 3 (11), 1329– 1334, DOI: 10.1039/b710604a
  There is no corresponding record for this reference.
45. 45
  Sindu, B.; Hamaekers, J. Feature-based prediction of properties of cross-linked epoxy polymers by molecular dynamics and machine learning techniques. Modell. Simul. Mater. Sci. Eng. 2025, 33 (6), 065010 DOI: 10.1088/1361-651X/adf56c
  There is no corresponding record for this reference.
46. 46
  Chew, A. K.; Afzal, M. A. F.; Chandrasekaran, A.; Kamps, J. H.; Ramakrishnan, V. Designing the next generation of polymers with machine learning and physics-based models. Mach. Learn. Sci. Technol. 2024, 5 (4), 045031 DOI: 10.1088/2632-2153/ad88d7
  There is no corresponding record for this reference.
47. 47
  He, J.; Rong, F. Prediction of molecularly imprinted polymer binding affinity based on graph neural networks. J. Phys.: Conference Series 2025, 3030, 012009
  There is no corresponding record for this reference.
48. 48
  Nagasawa, S.; Al-Naamani, E.; Saeki, A. Computer-aided screening of conjugated polymers for organic solar cell: classification by random forest. J. Phys. Chem. Lett. 2018, 9 (10), 2639– 2646, DOI: 10.1021/acs.jpclett.8b00635
  There is no corresponding record for this reference.
49. 49
  Kishino, M.; Matsumoto, K.; Kobayashi, Y.; Taguchi, R.; Akamatsu, N.; Shishido, A. Fatigue life prediction of bending polymer films using random forest. Int. J. Fatigue 2023, 166, 107230 DOI: 10.1016/j.ijfatigue.2022.107230
  There is no corresponding record for this reference.
50. 50
  Arora, A.; Lin, T.-S.; Rebello, N. J.; Av-Ron, S. H.; Mochigase, H.; Olsen, B. D. Random forest predictor for diblock copolymer phase behavior. ACS Macro Lett. 2021, 10 (11), 1339– 1345, DOI: 10.1021/acsmacrolett.1c00521
  There is no corresponding record for this reference.
51. 51
  Malashin, I.; Tynchenko, V.; Gantimurov, A.; Nelyub, V.; Borodulin, A. Support vector machines in polymer science: a review. Polymers 2025, 17 (4), 491, DOI: 10.3390/polym17040491
  There is no corresponding record for this reference.
52. 52
  Ziaee, H.; Hosseini, S. M.; Sharafpoor, A.; Fazavi, M.; Ghiasi, M. M.; Bahadori, A. Prediction of solubility of carbon dioxide in different polymers using support vector machine algorithm. J. Taiwan Inst. Chem. Eng. 2015, 46, 205– 213, DOI: 10.1016/j.jtice.2014.09.015
  There is no corresponding record for this reference.
53. 53
  Chen, F.-C. Virtual screening of conjugated polymers for organic photovoltaic devices using support vector machines and ensemble learning. Int. J. Polym. Sci. 2019, 2019 (1), 4538514 DOI: 10.1155/2019/4538514
  There is no corresponding record for this reference.
54. 54
  Brereton, R. G.; Lloyd, G. R. Support vector machines for classification and regression. Analyst 2010, 135 (2), 230– 267, DOI: 10.1039/B918972F
  There is no corresponding record for this reference.
55. 55
  Youshia, J.; Ali, M. E.; Lamprecht, A. Artificial neural network based particle size prediction of polymeric nanoparticles. Eur. J. Pharm. Biopharm. 2017, 119, 333– 342, DOI: 10.1016/j.ejpb.2017.06.030
  There is no corresponding record for this reference.
56. 56
  Zhang, Z.; Friedrich, K. Artificial neural networks applied to polymer composites: a review. Compos. Sci. Technol. 2003, 63 (14), 2029– 2044, DOI: 10.1016/S0266-3538(03)00106-4
  There is no corresponding record for this reference.
57. 57
  El Kadi, H. Modeling the mechanical behavior of fiber-reinforced polymeric composite materials using artificial neural networksa review. Compos. Struct. 2006, 73 (1), 1– 23, DOI: 10.1016/j.compstruct.2005.01.020
  There is no corresponding record for this reference.
58. 58
  Liu, W.; Cao, C. Artificial neural network prediction of glass transition temperature of polymers. Colloid Polym. Sci. 2009, 287 (7), 811– 818, DOI: 10.1007/s00396-009-2035-y
  There is no corresponding record for this reference.
59. 59
  Chen, L.; Pilania, G.; Batra, R.; Huan, T. D.; Kim, C.; Kuenneth, C.; Ramprasad, R. Polymer informatics: Current status and critical next steps. Mater. Sci. Eng. R 2021, 144, 100595 DOI: 10.1016/j.mser.2020.100595
  There is no corresponding record for this reference.
60. 60
  Butler, K. T.; Davies, D. W.; Cartwright, H.; Isayev, O.; Walsh, A. Machine learning for molecular and materials science. Nature 2018, 559, 547– 555, DOI: 10.1038/s41586-018-0337-2
  There is no corresponding record for this reference.
61. 61
  Schleder, G. R.; Padilha, A. C. M.; Acosta, C. M.; Costa, M.; Fazzio, A. From dft to machine learning: Recent approaches to materials science-a review. J. Chem. Inf. Model. 2019, 2 (4), 1347– 1356, DOI: 10.1088/2515-7639/ab084b
  There is no corresponding record for this reference.
62. 62
  Sha, W.; Li, Y.; Tang, S.; Tian, J.; Zhao, Y.; Guo, Y.; Zhang, W.; Zhang, X.; Lu, S.; Cao, Y.-C.; Cheng, S. Machine learning in polymer informatics. InfoMat 2021, 3 (4), 353– 361, DOI: 10.1002/inf2.12167
  There is no corresponding record for this reference.
63. 63
  Martin, T. B.; Audus, D. J. Emerging trends in machine learning: a polymer perspective. ACS Polymers Au 2023, 3 (3), 239– 258, DOI: 10.1021/acspolymersau.2c00053
  There is no corresponding record for this reference.
64. 64
  Ramprasad, R.; Batra, R.; Pilania, G.; Mannodi-Kanakkithodi, A.; Kim, C. Machine learning in materials informatics: recent applications and prospects. npj Comput. Mater. 2017, 3 (1), 54, DOI: 10.1038/s41524-017-0056-5
  There is no corresponding record for this reference.
65. 65
  Chen, G.; Shen, Z.; Iyer, A.; Ghumman, U. F.; Tang, S.; Bi, J.; Chen, W.; Li, Y. Machine-learning-assisted de novo design of organic molecules and polymers: opportunities and challenges. Polymers 2020, 12 (1), 163, DOI: 10.3390/polym12010163
  There is no corresponding record for this reference.
66. 66
  Wu, S.; Yamada, H.; Hayashi, Y.; Zamengo, M.; Yoshida, R. Potentials and challenges of polymer informatics: exploiting machine learning for polymer design, 2020, arXiv:2010.07683. arXiv.org e-Print archive https://arxiv.org/abs/2010.07683.
  There is no corresponding record for this reference.
67. 67
  Kumar, J. N.; Li, Q.; Jun, Y. Challenges and opportunities of polymer design with machine learning and high throughput experimentation. MRS Commun. 2019, 9 (2), 537– 544, DOI: 10.1557/mrc.2019.54
  There is no corresponding record for this reference.
68. 68
  Wieder, O.; Kohlbacher, S.; Kuenemann, M.; Garon, A.; Ducrot, P.; Seidel, T.; Langer, T. A compact review of molecular property prediction with graph neural networks. Drug Discovery Today: Technol. 2020, 37, 1– 12, DOI: 10.1016/j.ddtec.2020.11.009
  There is no corresponding record for this reference.
69. 69
  Wu, Z.; Pan, S.; Chen, F.; Long, G.; Zhang, C.; Philip, S. Y. A comprehensive survey on graph neural networks. IEEE Trans. Neural Netw. Learn. Syst. 2021, 32 (1), 4– 24, DOI: 10.1109/TNNLS.2020.2978386
  There is no corresponding record for this reference.
70. 70
  Gao, Q.; Dukker, T.; Schweidtmann, A. M.; Weber, J. M. Self-supervised graph neural networks for polymer property prediction. Mol. Syst. Des. Eng. 2024, 9 (11), 1130– 1143, DOI: 10.1039/D4ME00088A
  There is no corresponding record for this reference.
71. 71
  Inae, E.; Liu, Y.; Zhu, Y.; Xu, J.; Liu, G.; Zhang, R.; Luo, T.; Jiang, M. Modeling Polymers with Neural Networks; American Chemical Society, 2025.
  There is no corresponding record for this reference.
72. 72
  Gurney, K. An Introduction to Neural Networks; CRC Press, 2018.
  There is no corresponding record for this reference.
73. 73
  Haykin, S.; Network, N. A comprehensive foundation. Neural Netw. 2004, 2, 41
  There is no corresponding record for this reference.
74. 74
  LeCun, Y.; Bottou, L.; Bengio, Y.; Haffner, P. Gradient-based learning applied to document recognition. Proc. IEEE 1998, 86 (11), 2278– 2324, DOI: 10.1109/5.726791
  There is no corresponding record for this reference.
75. 75
  Li, Z.; Liu, F.; Yang, W.; Peng, S.; Zhou, J. A survey of convolutional neural networks: analysis, applications, and prospects. IEEE Trans. Neural Netw. Learn. Syst. 2022, 33 (12), 6999– 7019, DOI: 10.1109/TNNLS.2021.3084827
  There is no corresponding record for this reference.
76. 76
  Gu, J.; Wang, Z.; Kuen, J.; Ma, L.; Shahroudy, A.; Shuai, B.; Liu, T.; Wang, X.; Wang, G.; Cai, J.; Chen, T. Recent advances in convolutional neural networks. Pattern Recognit. 2018, 77, 354– 377, DOI: 10.1016/j.patcog.2017.10.013
  There is no corresponding record for this reference.
77. 77
  Ge, W.; De Silva, R.; Fan, Y.; Sisson, S. A.; Stenzel, M. H. Machine learning in polymer research. Adv. Mater. 2025, 37 (11), 2413695 DOI: 10.1002/adma.202413695
  There is no corresponding record for this reference.
78. 78
  Platonov, O.; Kuznedelev, D.; Diskin, M.; Babenko, A.; Prokhorenkova, L. A critical look at the evaluation of gnns under heterophily: Are we really making progress? 2023, arXiv:2302.11640. arXiv.org e-Print archive https://arxiv.org/abs/2302.11640.
  There is no corresponding record for this reference.
79. 79
  Wang, Y.; Li, Z.; Barati Farimani, A. Graph neural networks for molecules. In Machine Learning in Molecular Sciences; Springer, 2023; pp 21– 66.
  There is no corresponding record for this reference.
80. 80
  Hamilton, W.; Ying, Z.; Leskovec, J. Inductive representation learning on large graphs. In NIPS’17: Proceedings of the 31st International Conference on Neural Information Processing Systems; Department of Computer Science: Stanford, CA, 2017; pp 1025– 1035.
  There is no corresponding record for this reference.
81. 81
  Xu, K.; Hu, W.; Leskovec, J.; Jegelka, S. How powerful are graph neural networks? 2018, arXiv:1810.00826. arXiv.org e-Print archive https://arxiv.org/abs/1810.00826.
  There is no corresponding record for this reference.
82. 82
  Sharma, K.; Lee, Y.-C.; Nambi, S.; Salian, A.; Shah, S.; Kim, S.-W.; Kumar, S. A survey of graph neural networks for social recommender systems. ACM Computing Surveys 2024, 56 (10), 1– 34, DOI: 10.1145/3661821
  There is no corresponding record for this reference.
83. 83
  Kumar, S.; Mallik, A.; Khetarpal, A.; Panda, B. S. Influence maximization in social networks using graph embedding and graph neural network. Informat. Sci. 2022, 607, 1617– 1636, DOI: 10.1016/j.ins.2022.06.075
  There is no corresponding record for this reference.
84. 84
  Zhang, X.-M.; Liang, L.; Liu, L.; Tang, M.-J. Graph neural networks and their current applications in bioinformatics. Front. Genetics 2021, 12, 690049 DOI: 10.3389/fgene.2021.690049
  There is no corresponding record for this reference.
85. 85
  Malla, A. M.; Banka, A. A. A systematic review of deep graph neural networks: Challenges, classification, architectures, applications & potential utility in bioinformatics, 2023, arXiv:2311.02127. arXiv.org e-Print archive https://arxiv.org/abs/2311.02127.
  There is no corresponding record for this reference.
86. 86
  Zheng, P.; Wang, S.; Wang, X.; Zeng, X. Artificial intelligence in bioinformatics and drug repurposing: methods and applications. In Frontiers Research Topics , 2022.
  There is no corresponding record for this reference.
87. 87
  Reiser, P.; Neubert, M.; Eberhard, A.; Torresi, L.; Zhou, C.; Shao, C.; Metni, H.; van Hoesel, C.; Schopmans, H.; Sommer, T.; Friederich, P. Graph neural networks for materials science and chemistry. Commun. Mater. 2022, 3 (1), 93, DOI: 10.1038/s43246-022-00315-6
  There is no corresponding record for this reference.
88. 88
  Fung, V.; Zhang, J.; Juarez, E.; Sumpter, B. G. Benchmarking graph neural networks for materials chemistry. npj Computat. Mater. 2021, 7 (1), 84, DOI: 10.1038/s41524-021-00554-0
  There is no corresponding record for this reference.
89. 89
  Maurizi, M.; Gao, C.; Berto, F. Predicting stress, strain and deformation fields in materials and structures with graph neural networks. Sci. Rep. 2022, 12 (1), 21834 DOI: 10.1038/s41598-022-26424-3
  There is no corresponding record for this reference.
90. 90
  Kipf, T. N.; Welling, M. Semi-supervised classification with graph convolutional networks, 2016, arXiv:1609.02907. arXiv.org e-Print archive https://arxiv.org/abs/1609.02907.
  There is no corresponding record for this reference.
91. 91
  Gilmer, J.; Schoenholz, S. S.; Riley, P. F.; Vinyals, O.; Dahl, G. E. Neural message passing for quantum chemistry. In International Conference on Machine Learning; PMLR, 2017; pp 1263– 1272.
  There is no corresponding record for this reference.
92. 92
  Veličković, P.; Cucurull, G.; Casanova, A.; Romero, A.; Lio, P.; Bengio, Y. “ Graph attention networks, 2017, arXiv:1710.10903. arXiv.org e-Print archive https://arxiv.org/abs/1710.10903.
  There is no corresponding record for this reference.
93. 93
  Liu, G.; Inae, E.; Jiang, M. Deep learning for polymer property prediction. In Deep Learning for Polymer Discovery: Foundation and Advances; Springer, 2025; pp 17– 35.
  There is no corresponding record for this reference.
94. 94
  Dong, C.; Li, D.; Liu, J. Glass transition temperature prediction of polymers via graph reinforcement learning. Langmuir 2024, 40 (35), 18568– 18580, DOI: 10.1021/acs.langmuir.4c01906
  There is no corresponding record for this reference.
95. 95
  Wang, X.; Liao, L.; Huang, C.; Wu, X. Prediction of mechanical properties of cross-linked polymer interface by graph convolution network. Acta Mechanica Sinica 2026, 42 (3), 424627 DOI: 10.1007/s10409-024-24627-x
  There is no corresponding record for this reference.
96. 96
  Safari, H.; Bavarian, M. Enhancing polymer reaction engineering through the power of machine learning. In Systems and Control Transactions , 2024; p 157792.
  There is no corresponding record for this reference.
97. 97
  Liao, H.-C.; Lin, Y.-H.; Peng, C.-H.; Li, Y.-P. Directed message passing neural networks for accurate prediction of polymer-solvent interaction parameters. ACS Engineering Au 2025, 5, 530, DOI: 10.1021/acsengineeringau.5c00027
  There is no corresponding record for this reference.
98. 98
  Kim, S.; Chen, J.; Cheng, T.; Gindulyte, A.; He, J.; He, S.; Li, Q.; Shoemaker, B. A.; Thiessen, P. A.; Yu, B. Pubchem 2023 update. Nucleic Acids Res. 2023, 51 (D1), D1373– D1380, DOI: 10.1093/nar/gkac956
  There is no corresponding record for this reference.
99. 99
  Gao, Z.-Y.; Liu, G.-Q. Recent progress of web-enable material database and a case study of nims and matweb. J. Mater. Eng. 2013, 3 (11), 89– 96, DOI: 10.3969/j.issn.1001-4381.2013.06.001
  There is no corresponding record for this reference.
100. 100
  Wypych, G. Handbook of Polymers; Elsevier, 2022.
  There is no corresponding record for this reference.
101. 101
  Patra, A.; Batra, R.; Chandrasekaran, A.; Kim, C.; Huan, T. D.; Ramprasad, R. A multi-fidelity information-fusion approach to machine learn and predict polymer bandgap. Comput. Mater. Sci. 2020, 172, 109286 DOI: 10.1016/j.commatsci.2019.109286
  There is no corresponding record for this reference.
102. 102
  Dhamankar, S.; Webb, M. A. Chemically specific coarse-graining of polymers: methods and prospects. J. Polym. Sci. 2021, 59 (22), 2613– 2643, DOI: 10.1002/pol.20210555
  There is no corresponding record for this reference.
103. 103
  Müller-Plathe, F. Coarse-graining in polymer simulation: from the atomistic to the mesoscopic scale and back. ChemPhysChem 2002, 3 (9), 754– 769, DOI: 10.1002/1439-7641(20020916)3:9<754::AID-CPHC754>3.0.CO;2-U
  There is no corresponding record for this reference.
104. 104
  Harmandaris, V. A.; Reith, D.; Van der Vegt, N. F.; Kremer, K. Comparison between coarse-graining models for polymer systems: Two mapping schemes for polystyrene. Macromol. Chem. Phys. 2007, 208 (19–20), 2109– 2120, DOI: 10.1002/macp.200700245
  There is no corresponding record for this reference.
105. 105
  Padding, J. T.; Briels, W. J. Systematic coarse-graining of the dynamics of entangled polymer melts: the road from chemistry to rheology. J. Phys.: Condens. Matter 2011, 23 (23), 233101 DOI: 10.1088/0953-8984/23/23/233101
  There is no corresponding record for this reference.
106. 106
  Zhang, X.; Sheng, Y.; Liu, X.; Yang, J.; Goddard, W. A., III; Ye, C.; Zhang, W. Polymer-unit graph: Advancing interpretability in graph neural network machine learning for organic polymer semiconductor materials. J. Chem. Theory Comput. 2024, 20 (7), 2908– 2920, DOI: 10.1021/acs.jctc.3c01385
  There is no corresponding record for this reference.
107. 107
  Mohapatra, S.; An, J.; Gómez-Bombarelli, R. Chemistry-informed macromolecule graph representation for similarity computation, unsupervised and supervised learning. Machine Learning: Sci. Technol. 2022, 3 (1), 015028 DOI: 10.1088/2632-2153/ac545e
  There is no corresponding record for this reference.
108. 108
  Wang, X.; Xu, Y.; Zheng, H.; Yu, K. A scalable graph neural network method for developing an accurate force field of large flexible organic molecules. J. Phys. Chem. Lett. 2021, 12 (33), 7982– 7987, DOI: 10.1021/acs.jpclett.1c02214
  There is no corresponding record for this reference.
109. 109
  Qiu, H.; Qiu, X.; Dai, X.; Sun, Z.-Y. Design of polyimides with targeted glass transition temperature using a graph neural network. J. Mater. Chem. C 2023, 11 (8), 2930– 2940, DOI: 10.1039/D2TC05174E
  There is no corresponding record for this reference.
110. 110
  Zeng, M.; Kumar, J. N.; Zeng, Z.; Savitha, R.; Chandrasekhar, V. R.; Hippalgaonkar, K. “ Graph convolutional neural networks for polymers property prediction, 2018, arXiv:1811.06231. arXiv.org e-Print archive https://arxiv.org/abs/1811.06231.
  There is no corresponding record for this reference.
111. 111
  Queen, O.; McCarver, G. A.; Thatigotla, S.; Abolins, B. P.; Brown, C. L.; Maroulas, V.; Vogiatzis, K. D. Polymer graph neural networks for multitask property learning. npj Computat. Mater. 2023, 9, 90, DOI: 10.1038/s41524-023-01034-3
  There is no corresponding record for this reference.
112. 112
  Bongini, P.; Bianchini, M.; Scarselli, F. Molecular generative graph neural networks for drug discovery. Neurocomputing 2021, 450, 242– 252, DOI: 10.1016/j.neucom.2021.04.039
  There is no corresponding record for this reference.
113. 113
  Scarselli, F.; Gori, M.; Tsoi, A. C.; Hagenbuchner, M.; Monfardini, G. The graph neural network model. IEEE Transact. Neural Networks 2009, 20 (1), 61– 80, DOI: 10.1109/TNN.2008.2005605
  There is no corresponding record for this reference.
114. 114
  Dwivedi, V. P.; Joshi, C. K.; Luu, A. T.; Laurent, T.; Bengio, Y.; Bresson, X. Benchmarking graph neural networks. J. Machine Learning Res. 2023, 24 (43), 1– 48
  There is no corresponding record for this reference.
115. 115
  Veličković, P. Everything is connected: Graph neural networks. Curr. Opin. Struct. Biol. 2023, 79, 102538 DOI: 10.1016/j.sbi.2023.102538
  There is no corresponding record for this reference.
116. 116
  Jegelka, S. Theory of graph neural networks: Representation and learning. In International Congress of Mathematicians , 2022; pp 1– 23.
  There is no corresponding record for this reference.
117. 117
  Liu, Z.; Zhou, J. Introduction to Graph Neural Networks; Springer Nature, 2022.
  There is no corresponding record for this reference.
118. 118
  Tang, M.; Li, B.; Chen, H. Application of message passing neural networks for molecular property prediction. Curr. Opin. Struct. Biol. 2023, 81, 102616 DOI: 10.1016/j.sbi.2023.102616
  There is no corresponding record for this reference.
119. 119
  Hasebe, T. Knowledge-embedded message-passing neural networks: improving molecular property prediction with human knowledge. ACS omega 2021, 6 (42), 27955– 27967, DOI: 10.1021/acsomega.1c03839
  There is no corresponding record for this reference.
120. 120
  Zhang, H.; Lin, Y.; Li, S.; Dai, M.; Zhang, Y.; Huang, L.; Pang, J.; Wu, P.; Peng, J.; Tang, Z.; Ding, P.; Xiao, W.; Song, N.; Dongbo, D. Substructure-enhanced mpnn for polymer discovery and knowledge: A study in predicting glass transition temperature. Macromolecules 2025, 58, 9515, DOI: 10.1021/acs.macromol.4c02859
  There is no corresponding record for this reference.
121. 121
  Aldeghi, M.; Coley, C. W. A graph representation of molecular ensembles for polymer property prediction. Chem. Sci. 2022, 13 (35), 10486– 10498, DOI: 10.1039/D2SC02839E
  There is no corresponding record for this reference.
122. 122
  Heid, E.; Greenman, K. P.; Chung, Y.; Li, S.-C.; Graff, D. E.; Vermeire, F. H.; Wu, H.; Green, W. H.; McGill, C. J. Chemprop: a machine learning package for chemical property prediction. J. Chem. Inf. Model. 2024, 64 (1), 9– 17, DOI: 10.1021/acs.jcim.3c01250
  There is no corresponding record for this reference.
123. 123
  Antoniuk, E. R.; Li, P.; Kailkhura, B.; Hiszpanski, A. M. Representing polymers as periodic graphs with learned descriptors for accurate polymer property predictions. J. Chem. Inf. Model. 2022, 62, 5435– 5445, DOI: 10.1021/acs.jcim.2c00875
  There is no corresponding record for this reference.
124. 124
  Sanchez Medina, E. I.; Kunchapu, S.; Sundmacher, K. Gibbs-helmholtz graph neural network for the prediction of activity coefficients of polymer solutions at infinite dilution. J. Phys. Chem. A 2023, 127 (46), 9863– 9873, DOI: 10.1021/acs.jpca.3c05892
  There is no corresponding record for this reference.
125. 125
  Battaglia, P. W.; Hamrick, J. B.; Bapst, V.; Sanchez-Gonzalez, A.; Zambaldi, V.; Malinowski, M.; Tacchetti, A.; Raposo, D.; Santoro, A.; Faulkner, R.; C, G.; F, S.; Ballard, A.; Justin, G.; George, D.; Ashish, V.; Kelsey, A.; Charles, N.; Victoria, L.; Chris, D.; Nicolas, H.; Dann, W.; Pushmeet, K.; Matt, B.; Oriol, V.; Yujia, L.; Razvan, P. Relational inductive biases, deep learning, and graph networks, 2018, arXiv:1806.01261. arXiv.org e-Print archive https://arxiv.org/abs/1806.01261.
  There is no corresponding record for this reference.
126. 126
  Gurnani, R.; Kuenneth, C.; Toland, A.; Ramprasad, R. Polymer informatics at scale with multitask graph neural networks. Chem. Mater. 2023, 35 (6), 1560– 1567, DOI: 10.1021/acs.chemmater.2c02991
  There is no corresponding record for this reference.
127. 127
  St John, P. C.; Phillips, C.; Kemper, T. W.; Wilson, A. N.; Guan, Y.; Crowley, M. F.; Nimlos, M. R.; Larsen, R. E. Message-passing neural networks for high-throughput polymer screening. J. Chem. Phys. 2019, 150, 234111 DOI: 10.1063/1.5099132
  There is no corresponding record for this reference.
128. 128
  Li, Y.; Zemel, R.; Brockschmidt, M.; Tarlow, D. Gated graph sequence neural networks. In Proceedings of ICLR16 , 2016.
  There is no corresponding record for this reference.
129. 129
  Dey, R.; Salem, F. M. Gate-variants of gated recurrent unit (gru) neural networks. In 2017 IEEE 60th international midwest symposium on circuits and systems (MWSCAS); IEEE, 2017; pp 1597– 1600.
  There is no corresponding record for this reference.
130. 130
  Sun, M.; Zhao, S.; Gilvary, C.; Elemento, O.; Zhou, J.; Wang, F. Graph convolutional networks for computational drug development and discovery. Briefings Bioinf. 2020, 21 (3), 919– 935, DOI: 10.1093/bib/bbz042
  There is no corresponding record for this reference.
131. 131
  Xu, X.; Zhao, X.; Wei, M.; Li, Z. A comprehensive review of graph convolutional networks: approaches and applications. Electronic Res. Archive 2023, 31 (7), 4185– 4215, DOI: 10.3934/era.2023213
  There is no corresponding record for this reference.
132. 132
  Zhou, J.; Cui, G.; Hu, S.; Zhang, Z.; Yang, C.; Liu, Z.; Wang, L.; Li, C.; Sun, M. Graph neural networks: A review of methods and applications. IEEE Transactions on Pattern Analysis and Machine Intelligence 2019, 42, 283– 300
  There is no corresponding record for this reference.
133. 133
  Hu, J.; Li, Z.; Lin, J.; Zhang, L. Prediction and interpretability of glass transition temperature of homopolymers by data-augmented graph convolutional neural networks. ACS Appl. Mater. Interfaces 2023, 15 (46), 54006– 54017, DOI: 10.1021/acsami.3c13698
  There is no corresponding record for this reference.
134. 134
  Hickey, K.; Feinstein, J.; Sivaraman, G.; MacDonell, M.; Yan, E.; Matherson, C.; Coia, S.; Xu, J.; Picel, K. Applying machine learning and quantum chemistry to predict the glass transition temperatures of polymers. Comput. Mater. Sci. 2024, 238, 112933 DOI: 10.1016/j.commatsci.2024.112933
  There is no corresponding record for this reference.
135. 135
  Kimmig, J.; Schuett, T.; Vollrath, A.; Zechel, S.; Schubert, U. S. Prediction of nanoparticle sizes for arbitrary methacrylates using artificial neuronal networks. Adv. Sci. 2021, 8 (23), 2102429 DOI: 10.1002/advs.202102429
  There is no corresponding record for this reference.
136. 136
  Tian, Y.; Zhang, Y.; Zhang, H. Recent advances in stochastic gradient descent in deep learning. Mathematics 2023, 11 (3), 682, DOI: 10.3390/math11030682
  There is no corresponding record for this reference.
137. 137
  Merity, S. “ Single headed attention rnn: Stop thinking with your head, 2019, arXiv:1911.11423. arXiv.org e-Print archive https://arxiv.org/abs/1911.11423.
  There is no corresponding record for this reference.
138. 138
  Inoue, H. Multi-sample dropout for accelerated training and better generalization, 2019, arXiv:1905.09788. arXiv.org e-Print archive https://arxiv.org/abs/1905.09788.
  There is no corresponding record for this reference.
139. 139
  Cho, K.; Van Merriënboer, B.; Bahdanau, D.; Bengio, Y. On the properties of neural machine translation: Encoder-decoder approaches, 2014, arXiv:1409.1259. arXiv.org e-Print archive https://arxiv.org/abs/1409.1259.
  There is no corresponding record for this reference.
140. 140
  Morris, C.; Ritzert, M.; Fey, M.; Hamilton, W. L.; Lenssen, J. E.; Rattan, G.; Grohe, M. Weisfeiler and leman go neural: Higher-order graph neural networks. Proceedings of the AAAI Conference on Artificial Intelligence 2019, 33, 4602– 4609, DOI: 10.1609/aaai.v33i01.33014602
  There is no corresponding record for this reference.
141. 141
  Ye, X.-b.; Guan, Q.; Luo, W.; Fang, L.; Lai, Z.-R.; Wang, J. Molecular substructure graph attention network for molecular property identification in drug discovery. Pattern Recognition 2022, 128, 108659 DOI: 10.1016/j.patcog.2022.108659
  There is no corresponding record for this reference.
142. 142
  Vrahatis, A. G.; Lazaros, K.; Kotsiantis, S. Graph attention networks: a comprehensive review of methods and applications. Future Internet 2024, 16 (9), 318, DOI: 10.3390/fi16090318
  There is no corresponding record for this reference.
143. 143
  Xiong, Z.; Wang, D.; Liu, X.; Zhong, F.; Wan, X.; Li, X.; Li, Z.; Luo, X.; Chen, K.; Jiang, H.; Z M Pushing the boundaries of molecular representation for drug discovery with the graph attention mechanism. J. Med. Chem. 2020, 63 (16), 8749– 8760, DOI: 10.1021/acs.jmedchem.9b00959
  There is no corresponding record for this reference.
144. 144
  Li, D.; Ru, Y.; Liu, J. Gatboost: Mining graph attention networks-based important substructures of polymers for a better property prediction. Mater. Today Commun. 2024, 38, 107577 DOI: 10.1016/j.mtcomm.2023.107577
  There is no corresponding record for this reference.
145. 145
  Chen, T.; Guestrin, C. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , 2016; pp 785– 794.
  There is no corresponding record for this reference.
146. 146
  Liu, T.; Jiang, A.; Zhou, J.; Li, M.; Kwan, H. K. Graphsage-based dynamic spatial-temporal graph convolutional network for traffic prediction. IEEE Transactions on Intelligent Transportation Systems 2023, 24 (10), 11210– 11224, DOI: 10.1109/TITS.2023.3279929
  There is no corresponding record for this reference.
147. 147
  Xie, T.; France-Lanord, A.; Wang, Y.; Lopez, J.; Stolberg, M. A.; Hill, M.; Leverick, G. M.; Gomez-Bombarelli, R.; Johnson, J. A.; Shao-Horn, Y.; Grossman, J. C. Accelerating amorphous polymer electrolyte screening by learning to reduce errors in molecular dynamics simulated properties. Nat. Commun. 2022, 13 (1), 3415 DOI: 10.1038/s41467-022-30994-1
  There is no corresponding record for this reference.
148. 148
  Qiu, H.; Wang, J.; Qiu, X.; Dai, X.; Sun, Z.-Y. Heat-resistant polymer discovery by utilizing interpretable graph neural network with small data. Macromolecules 2024, 57 (8), 3515– 3528, DOI: 10.1021/acs.macromol.4c00508
  There is no corresponding record for this reference.
149. 149
  Park, J.; Shim, Y.; Lee, F.; Rammohan, A.; Goyal, S.; Shim, M.; Jeong, C.; Kim, D. S. Prediction and interpretation of polymer properties using the graph convolutional network. ACS Polymers Au 2022, 2, 213, DOI: 10.1021/acspolymersau.1c00050
  There is no corresponding record for this reference.
150. 150
  Burkholz, R.; Quackenbush, J.; Bojar, D. Using graph convolutional neural networks to learn a representation for glycans. Cell Rep. 2021, 35 (11), 109251, DOI: 10.1016/j.celrep.2021.109251
  There is no corresponding record for this reference.
151. 151
  Volgin, I. V.; Batyr, P. A.; Matseevich, A. V.; Dobrovskiy, A. Y.; Andreeva, M. V.; Nazarychev, V. M.; Larin, S. V.; Goikhman, M. Y.; Vizilter, Y. V.; Askadskii, A. A.; Lyulin, S. V. Machine learning with enormous “synthetic” data sets: Predicting glass transition temperature of polyimides using graph convolutional neural networks. ACS Omega 2022, 7 (48), 43678– 43691, DOI: 10.1021/acsomega.2c04649
  There is no corresponding record for this reference.
152. 152
  Otsuka, S.; Kuwajima, I.; Hosoya, J.; Xu, Y.; Yamazaki, M. Polyinfo: Polymer database for polymeric materials design. In 2011 International Conference on Emerging Intelligent Data and Web Technologies; IEEE, 2011; pp 22– 29.
  There is no corresponding record for this reference.
153. 153
  Brandrup, J.; Immergut, E. H.; Grulke, E. A.; Abe, A.; Bloch, D. R. Polymer Handbook; Wiley: New York, 1999; Vol. 89.
  There is no corresponding record for this reference.
154. 154
  Mark, J. E. Physical Properties of Polymers Handbook; Springer, 2007; Vol. 1076.
  There is no corresponding record for this reference.
155. 155
  Ma, R.; Luo, T. Pi1m: a benchmark database for polymer informatics. J. Chem. Inf. Model. 2020, 60 (10), 4684– 4690, DOI: 10.1021/acs.jcim.0c00726
  There is no corresponding record for this reference.
156. 156
  Walsh, D. J.; Zou, W.; Schneider, L.; Mello, R.; Deagen, M. E.; Mysona, J.; Lin, T.-S.; de Pablo, J. J.; Jensen, K. F.; Audus, D. J. Community resource for innovation in polymer technology (cript): a scalable polymer material data structure, 2023.
  There is no corresponding record for this reference.
157. 157
  Yan, C.; Li, G. The rise of machine learning in polymer discovery. Advanced Intelligent Systems 2023, 5 (4), 2200243 DOI: 10.1002/aisy.202200243
  There is no corresponding record for this reference.
158. 158
  Levine, D. S.; Shuaibi, M.; Spotte-Smith, E. W. C.; Taylor, M. G.; Hasyim, M. R.; Michel, K.; Batatia, I.; Csányi, G.; Dzamba, M.; Eastman, P. The open molecules 2025 (omol25) dataset, evaluations, and models, 2025, arXiv:2505.08762. arXiv.org e-Print archive https://arxiv.org/abs/2505.08762.
  There is no corresponding record for this reference.
159. 159
  Mark, J. E. Polymer Data Handbook, Oxford University Press, New York, NY, 2009.
  There is no corresponding record for this reference.
160. 160
  Hayashi, Y.; Shiomi, J.; Morikawa, J.; Yoshida, R. Radonpy: automated physical property calculation using all-atom classical molecular dynamics simulations for polymer informatics. npj Computat. Mater. 2022, 8 (1), 222, DOI: 10.1038/s41524-022-00906-4
  There is no corresponding record for this reference.
161. 161
  Ellis, B.; Smith, R. Polymers: A Property Database. CRC Press, 2008.
  There is no corresponding record for this reference.
162. 162
  Huan, T. D.; Mannodi-Kanakkithodi, A.; Kim, C.; Sharma, V.; Pilania, G.; Ramprasad, R. A polymer dataset for accelerated property prediction and design. Scientific data 2016, 3 (1), 1– 10, DOI: 10.1038/sdata.2016.12
  There is no corresponding record for this reference.
163. 163
  Pence, H. E.; Williams, A. Chemspider: an online chemical information resource, 2010.
  There is no corresponding record for this reference.
164. 164
  Polymer property predictor and database. https://pppdb.uchicago.edu/, 2019–2025.
  There is no corresponding record for this reference.
165. 165
  Campus plastic s database: Computer aided material preselection by uniform standards. https://www.campusplastics.com/, 1988–2025.
  There is no corresponding record for this reference.
166. 166
  Miccio, L. A.; Schwartz, G. A. From chemical structure to quantitative polymer properties prediction through convolutional neural networks. Polymer 2020, 193, 122341 DOI: 10.1016/j.polymer.2020.122341
  There is no corresponding record for this reference.
167. 167
  Xu, P.; Lu, T.; Ju, L.; Tian, L.; Li, M.; Lu, W. Machine learning aided design of polymer with targeted band gap based on dft computation. J. Phys. Chem. B 2021, 125 (2), 601– 611, DOI: 10.1021/acs.jpcb.0c08674
  There is no corresponding record for this reference.
168. 168
  Xiong, J.; Xiong, Z.; Chen, K.; Jiang, H.; Zheng, M. Graph neural networks for automated de novo drug design. Drug Discovery Today 2021, 26 (6), 1382– 1393, DOI: 10.1016/j.drudis.2021.02.011
  There is no corresponding record for this reference.
169. 169
  Medina, H.; Drake, R.; Farmer, C. Accelerated prediction of molecular properties for per-and polyfluoroalkyl substances using graph neural networks with adjacency-free message passing. Environ. Pollut. 2025, 382, 126705 DOI: 10.1016/j.envpol.2025.126705
  There is no corresponding record for this reference.
170. 170
  Tozzini, V. Coarse-grained models for proteins. Curr. Opin. Struct. Biol. 2005, 15 (2), 144– 150, DOI: 10.1016/j.sbi.2005.02.005
  There is no corresponding record for this reference.
171. 171
  Joshi, S. Y.; Deshmukh, S. A. A review of advancements in coarse-grained molecular dynamics simulations. Mol. Simul. 2021, 47 (10–11), 786– 803, DOI: 10.1080/08927022.2020.1828583
  There is no corresponding record for this reference.
172. 172
  Noid, W. G.; Liu, P.; Wang, Y.; Chu, J.-W.; Ayton, G. S.; Izvekov, S.; Andersen, H. C.; Voth, G. A. The multiscale coarse-graining method. ii. numerical implementation for coarse-grained molecular models. J. Chem. Phys. 2008, 128, 244115, DOI: 10.1063/1.2938857
  There is no corresponding record for this reference.
173. 173
  Pak, A. J.; Voth, G. A. Advances in coarse-grained modeling of macromolecular complexes. Curr. Opin. Struct. Biol. 2018, 52, 119– 126, DOI: 10.1016/j.sbi.2018.11.005
  There is no corresponding record for this reference.
174. 174
  Noid, W. G. Perspective: Coarse-grained models for biomolecular systems. J. Chem. Phys. 2013, 139, 090901, DOI: 10.1063/1.4818908
  There is no corresponding record for this reference.
175. 175
  Kmiecik, S.; Gront, D.; Kolinski, M.; Wieteska, L.; Dawid, A. E.; Kolinski, A. Coarse-grained protein models and their applications. Chem. Rev. 2016, 116 (14), 7898– 7936, DOI: 10.1021/acs.chemrev.6b00163
  There is no corresponding record for this reference.
176. 176
  Ingólfsson, H. I.; Lopez, C. A.; Uusitalo, J. J.; de Jong, D. H.; Gopal, S. M.; Periole, X.; Marrink, S. J. The power of coarse graining in biomolecular simulations. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2014, 4 (3), 225– 248, DOI: 10.1002/wcms.1169
  There is no corresponding record for this reference.
177. 177
  Noid, W.; Szukalo, R. J.; Kidder, K. M.; Lesniewski, M. C. Rigorous progress in coarse-graining. Annu. Rev. Phys. Chem. 2024, 75 (1), 21– 45, DOI: 10.1146/annurev-physchem-062123-010821
  There is no corresponding record for this reference.
178. 178
  Guenza, M. G. Everything you want to know about coarse-graining and never dared to ask: Macromolecules as a key example. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2025, 15 (2), e70022 DOI: 10.1002/wcms.70022
  There is no corresponding record for this reference.
179. 179
  Harmandaris, V. A.; Kremer, K. Predicting polymer dynamics at multiple length and time scales. Soft Matter 2009, 5 (20), 3920– 3926, DOI: 10.1039/b905361a
  There is no corresponding record for this reference.
180. 180
  Ye, H.; Xian, W.; Li, Y. Machine learning of coarse-grained models for organic molecules and polymers: Progress, opportunities, and challenges. ACS Omega 2021, 6 (3), 1758– 1772, DOI: 10.1021/acsomega.0c05321
  There is no corresponding record for this reference.
181. 181
  Noid, W. G. Perspective: Advances, challenges, and insight for predictive coarse-grained models. J. Phys. Chem. B 2023, 127 (19), 4174– 4207, DOI: 10.1021/acs.jpcb.2c08731
  There is no corresponding record for this reference.
182. 182
  Vettorel, T.; Besold, G.; Kremer, K. Fluctuating soft-sphere approach to coarse-graining of polymer models. Soft Matter 2010, 6 (10), 2282– 2292, DOI: 10.1039/b921159d
  There is no corresponding record for this reference.
183. 183
  Español, P.; Serrano, M.; Pagonabarraga, I.; Zúñiga, I. Energy-conserving coarse-graining of complex molecules. Soft Matter 2016, 12 (21), 4821– 4837, DOI: 10.1039/C5SM03038B
  There is no corresponding record for this reference.
184. 184
  Husic, B. E.; Charron, N. E.; Lemm, D.; Wang, J.; Pérez, A.; Majewski, M.; Krämer, A.; Chen, Y.; Olsson, S.; De Fabritiis, G.; Noe, F.; Clementi, C. Coarse graining molecular dynamics with graph neural networks J. Chem. Phys. 2020; Vol. 153, DOI: 10.1063/5.0026133 .
  There is no corresponding record for this reference.
185. 185
  Wang, X.; Wu, Y.; Zhang, A.; He, X.; Chua, T.-S. Towards multi-grained explainability for graph neural networks. Adv. Neural Inf. Process. Syst. 2021, 34, 18446– 18458
  There is no corresponding record for this reference.
186. 186
  Peter, C.; Kremer, K. Multiscale simulation of soft matter systems-from the atomistic to the coarse-grained level and back. Soft Matter 2009, 5 (22), 4357– 4366, DOI: 10.1039/b912027k
  There is no corresponding record for this reference.
187. 187
  Zhao, Y.; Chen, G.; Liu, J. Polymer data challenges in the ai era: Bridging gaps for next-generation energy materials, 2025, arXiv:2505.13494. arXiv.org e-Print archive https://arxiv.org/abs/2505.13494.
  There is no corresponding record for this reference.
188. 188
  Orio, M.; Pantazis, D. A.; Neese, F. Density functional theory. Photosynth. Res. 2009, 102, 443– 453, DOI: 10.1007/s11120-009-9404-8
  There is no corresponding record for this reference.
189. 189
  Joshi, S. P.; Bucsek, A.; Pagan, D. C.; Daly, S.; Ravindran, S.; Marian, J.; Bessa, M. A.; Kalidindi, S. R.; Admal, N. C.; Reina, C.; Ghosh, S.; Warren, J. A.; Vinals, J.; Tadmor, E. B. “ Integrated experiment and simulation co-design: A key infrastructure for predictive mesoscale materials modeling, 2025, arXiv:2503.09793. arXiv.org e-Print archive https://arxiv.org/abs/2503.09793.
  There is no corresponding record for this reference.
190. 190
  Zaporozhets, I.; Clementi, C. Multibody terms in protein coarse-grained models: A top-down perspective. J. Phys. Chem. B 2023, 127 (31), 6920– 6927, DOI: 10.1021/acs.jpcb.3c04493
  There is no corresponding record for this reference.
191. 191
  Jin, J.; Pak, A. J.; Durumeric, A. E.; Loose, T. D.; Voth, G. A. Bottom-up coarse-graining: Principles and perspectives. J. Chem. Theory Comput. 2022, 18 (10), 5759– 5791, DOI: 10.1021/acs.jctc.2c00643
  There is no corresponding record for this reference.
192. 192
  Praprotnik, M.; Delle Site, L.; Kremer, K. Multiscale simulation of soft matter: From scale bridging to adaptive resolution. Annu. Rev. Phys. Chem. 2008, 59, 545– 571, DOI: 10.1146/annurev.physchem.59.032607.093707
  There is no corresponding record for this reference.
193. 193
  Milano, G.; Kawakatsu, T. Hybrid particle-field molecular dynamics simulations for dense polymer systems J. Chem. Phys. 2009; Vol. 130, DOI: 10.1063/1.3142103 .
  There is no corresponding record for this reference.
194. 194
  Wang, J.; Han, Y.; Xu, Z.; Yang, X.; Ramakrishna, S.; Liu, Y. Dissipative particle dynamics simulation: A review on investigating mesoscale properties of polymer systems. Macromol. Mater. Eng. 2021, 306 (4), 2000724 DOI: 10.1002/mame.202000724
  There is no corresponding record for this reference.
195. 195
  Santo, K. P.; Neimark, A. V. Dissipative particle dynamics simulations in colloid and interface science: A review. Adv. Colloid Interface Sci. 2021, 298, 102545 DOI: 10.1016/j.cis.2021.102545
  There is no corresponding record for this reference.
196. 196
  Hoogerbrugge, P. J.; Koelman, J. Simulating microscopic hydrodynamic phenomena with dissipative particle dynamics. Europhys. Lett. 1992, 19 (3), 155, DOI: 10.1209/0295-5075/19/3/001
  There is no corresponding record for this reference.
197. 197
  Español, P.; Warren, P. Statistical mechanics of dissipative particle dynamics. Europhys. Lett. 1995, 30 (4), 191, DOI: 10.1209/0295-5075/30/4/001
  There is no corresponding record for this reference.
198. 198
  Marrink, S. J.; Monticelli, L.; Melo, M. N.; Alessandri, R.; Tieleman, D. P.; Souza, P. C. Two decades of martini: Better beads, broader scope. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2023, 13 (1), e1620 DOI: 10.1002/wcms.1620
  There is no corresponding record for this reference.
199. 199
  Alessandri, R.; Grünewald, F.; Marrink, S. J. The martini model in materials science. Adv. Mater. 2021, 33 (24), 2008635 DOI: 10.1002/adma.202008635
  There is no corresponding record for this reference.
200. 200
  Souza, P. C. T.; Alessandri, R.; Barnoud, J.; Thallmair, S.; Faustino, I.; Grünewald, F.; Patmanidis, I.; Abdizadeh, H.; Bruininks, B. M.; Wassenaar, T. A.; Kroon, P.; Melcr, K.; Nieto, V.; Corradi, V.; Khan, H. M.; Domanski, J.; Javanainen, M.; Martinez-Seara, H.; Reuter, N.; Best, R. B.; Vattulainen, I.; Monticelli, L.; Periole, X.; Peter, T. D.; Vries, dH.; Marrink, S. J. Martini 3: a general purpose force field for coarse-grained molecular dynamics. Nat. Methods 2021, 18 (4), 382– 388, DOI: 10.1038/s41592-021-01098-3
  There is no corresponding record for this reference.
201. 201
  Fredrickson, G. H. The Equilibrium Theory of Inhomogeneous Polymers. Oxford University Press, 2006.
  There is no corresponding record for this reference.
202. 202
  Schmid, F. Self-consistent-field theories for complex fluids. J. Phys.: Condens. Matter 1998, 10 (37), 8105, DOI: 10.1088/0953-8984/10/37/002
  There is no corresponding record for this reference.
203. 203
  Maerzke, K. A.; Siepmann, J. I. Transferable potentials for phase equilibria- coarse-grain description for linear alkanes. J. Phys. Chem. B 2011, 115 (13), 3452– 3465, DOI: 10.1021/jp1063935
  There is no corresponding record for this reference.
204. 204
  Liwo, A.; Ołdziej, S.; Pincus, M. R.; Wawak, R. J.; Rackovsky, S.; Scheraga, H. A. A united-residue force field for off-lattice protein-structure simulations. i. functional forms and parameters of long-range side-chain interaction potentials from protein crystal data. J. Comput. Chem. 1997, 18 (7), 849– 873, DOI: 10.1002/(SICI)1096-987X(199705)18:7<849::AID-JCC1>3.0.CO;2-R
  There is no corresponding record for this reference.
205. 205
  Maisuradze, G. G.; Senet, P.; Czaplewski, C.; Liwo, A.; Scheraga, H. A. Investigation of protein folding by coarse-grained molecular dynamics with the unres force field. J. Phys. Chem. A 2010, 114 (13), 4471– 4485, DOI: 10.1021/jp9117776
  There is no corresponding record for this reference.
206. 206
  Reith, D.; Pütz, M.; Müller-Plathe, F. Deriving effective mesoscale potentials from atomistic simulations. J. Comput. Chem. 2003, 24 (13), 1624– 1636, DOI: 10.1002/jcc.10307
  There is no corresponding record for this reference.
207. 207
  Agrawal, V.; Arya, G.; Oswald, J. Simultaneous iterative boltzmann inversion for coarse-graining of polyurea. Macromolecules 2014, 47 (10), 3378– 3389, DOI: 10.1021/ma500320n
  There is no corresponding record for this reference.
208. 208
  Tschöp, W.; Kremer, K.; Batoulis, J.; Bürger, T.; Hahn, O. Simulation of polymer melts. i. coarse-graining procedure for polycarbonates. Acta Polym. 1998, 49 (2–3), 61– 74, DOI: 10.1002/(SICI)1521-4044(199802)49:2/3<61::AID-APOL61>3.0.CO;2-V
  There is no corresponding record for this reference.
209. 209
  McCoy, J. D.; Curro, J. G. Mapping of explicit atom onto united atom potentials. Macromolecules 1998, 31 (26), 9362– 9368, DOI: 10.1021/ma981060g
  There is no corresponding record for this reference.
210. 210
  Baschnagel, J.; Binder, K.; Doruker, P.; Gusev, A. A.; Hahn, O.; Kremer, K.; Mattice, W. L.; Müller-Plathe, F.; Murat, M.; Paul, W.; Santos, S.; Suter, U. W.; Tries, V. Bridging the gap between atomistic and coarse-grained models of polymers: Status and perspectives. Viscoelasticity, Atomistic Models, Statistical Chemistry 2000, 152, 41– 156, DOI: 10.1007/3-540-46778-5_2
  There is no corresponding record for this reference.
211. 211
  Akkermans, R. L. C.; Briels, W. J. A structure-based coarse-grained model for polymer melts. J. Chem. Phys. 2001, 114 (2), 1020– 1031, DOI: 10.1063/1.1330744
  There is no corresponding record for this reference.
212. 212
  Uneyama, T.; Masubuchi, Y. Plateau moduli of several single-chain slip-link and slip-spring models. Macromolecules 2021, 54 (3), 1338– 1353, DOI: 10.1021/acs.macromol.0c01790
  There is no corresponding record for this reference.
213. 213
  Behbahani, A. F.; Schneider, L.; Rissanou, A.; Chazirakis, A.; Bacova, P.; Jana, P. K.; Li, W.; Doxastakis, M.; Polinska, P.; Burkhart, C.; Muller, M.; Harmandaris, V. A. Dynamics and rheology of polymer melts via hierarchical atomistic, coarse-grained, and slip-spring simulations. Macromolecules 2021, 54 (6), 2740– 2762, DOI: 10.1021/acs.macromol.0c02583
  There is no corresponding record for this reference.
214. 214
  Wu, Z.; Müller-Plathe, F. Slip-spring hybrid particle-field molecular dynamics for coarse-graining branched polymer melts: Polystyrene melts as an example. J. Chem. Theory Comput. 2022, 18 (6), 3814– 3828, DOI: 10.1021/acs.jctc.2c00107
  There is no corresponding record for this reference.
215. 215
  Underhill, P. T.; Doyle, P. S. On the coarse-graining of polymers into bead-spring chains. J. Non-Newtonian Fluid Mech. 2004, 122 (1–3), 3– 31, DOI: 10.1016/j.jnnfm.2003.10.006
  There is no corresponding record for this reference.
216. 216
  Kremer, K.; Grest, G. S. Dynamics of entangled linear polymer melts: A molecular-dynamics simulation. J. Chem. Phys. 1990, 92 (8), 5057– 5086, DOI: 10.1063/1.458541
  There is no corresponding record for this reference.
217. 217
  Everaers, R.; Karimi-Varzaneh, H. A.; Fleck, F.; Hojdis, N.; Svaneborg, C. Kremer-grest models for commodity polymer melts: Linking theory, experiment, and simulation at the kuhn scale. Macromolecules 2020, 53 (6), 1901– 1916, DOI: 10.1021/acs.macromol.9b02428
  There is no corresponding record for this reference.
218. 218
  Nitta, H.; Ozawa, T.; Yasuoka, K. Construction of full-atomistic polymer amorphous structures using reverse-mapping from kremer–grest models J. Chem. Phys. , 2023 159 no. 19.
  There is no corresponding record for this reference.
219. 219
  Miwatani, R.; Takahashi, K. Z.; Arai, N. Performance of coarse graining in estimating polymer properties: Comparison with the atomistic model. Polymers 2020, 12 (2), 382, DOI: 10.3390/polym12020382
  There is no corresponding record for this reference.
220. 220
  Praprotnik, M.; Delle Site, L.; Kremer, K. Adaptive resolution scheme for efficient hybrid atomistic-mesoscale molecular dynamics simulations of dense liquids. Phys. Rev. E 2006, 73 (6), 066701 DOI: 10.1103/PhysRevE.73.066701
  There is no corresponding record for this reference.
221. 221
  Darré, L.; Machado, M. R.; Brandner, A. F.; González, H. C.; Ferreira, S.; Pantano, S. Sirah: a structurally unbiased coarse-grained force field for proteins with aqueous solvation and long-range electrostatics. J. Chem. Theory Comput. 2015, 11 (2), 723– 739, DOI: 10.1021/ct5007746
  There is no corresponding record for this reference.
222. 222
  Das, R.; Baker, D. Macromolecular modeling with rosetta. Annu. Rev. Biochem. 2008, 77, 363– 382, DOI: 10.1146/annurev.biochem.77.062906.171838
  There is no corresponding record for this reference.
223. 223
  Leman, J. K.; Weitzner, B. D.; Lewis, S. M.; Adolf-Bryfogle, J.; Alam, N.; Alford, R. F.; Aprahamian, M.; Baker, D.; Barlow, K. A.; Barth, P.; Basanta, B.; Bender, B. J.; Blacklock, K.; Bonet, J.; Boyken, S. E.; Bradley, P.; Bystroff, C.; Conway, P.; Cooper, S.; Correia, B. E.; Coventry, B.; Das, R.; De Jong, R. M.; DiMaio, F.; Dsilva, L.; Dunbrack, R.; Ford, A. S.; Frenz, B.; Fu, D. Y.; Geniesse, C.; Goldschmidt, L.; Gowthaman, R.; Gray, J. J.; Gront, D.; Guffy, S.; Horowitz, S.; Huang, P.-S.; Huber, T.; Jacobs, T. M.; Jeliazkov, J. R.; Johnson, D. K.; Kappel, K.; Karanicolas, J.; Khakzad, H.; Khar, K. R.; Khare, S. D.; Khatib, F.; Khramushin, A.; King, I. C.; Kleffner, R.; Koepnick, B.; Kortemme, T.; Kuenze, G.; Kuhlman, B.; Kuroda, D.; Labonte, J. W.; Lai, J. K.; Lapidoth, G.; Leaver-Fay, A.; Lindert, S.; Linsky, T.; London, N.; Lubin, J. H.; Lyskov, S.; Maguire, J.; Malmström, L.; Marcos, E.; Marcu, O.; Marze, N. A.; Meiler, J.; Moretti, R.; Mulligan, V. K.; Nerli, S.; Norn, C.; Ó’Conchúir, S.; Ollikainen, N.; Ovchinnikov, S.; Pacella, M. S.; Pan, X.; Park, H.; Pavlovicz, R. E.; Pethe, M.; Pierce, B. G.; Pilla, K. B.; Raveh, B.; Renfrew, P. D.; Roy Burman, S. S.; Rubenstein, A.; Sauer, M. F.; Scheck, A.; Schief, W.; Schueler-Furman, O.; Sedan, Y.; Sevy, A. M.; Sgourakis, N. G.; Shi, L.; Siegel, J. B.; Silva, D.-A.; Smith, S.; Song, Y.; Stein, A.; Szegedy, M.; Teets, F. D.; Thyme, S. B.; Wang, R. Y.-R.; Watkins, A.; Zimmerman, L.; Bonneau, R. Macromolecular modeling and design in rosetta: recent methods and frameworks. Nat. Methods 2020, 17 (7), 665– 680, DOI: 10.1038/s41592-020-0848-2
  There is no corresponding record for this reference.
224. 224
  Baldassarre, F.; Azizpour, H. Explainability techniques for graph convolutional networks, 2019, arXiv:1905.13686 arXiv.org e-Print archive https://arxiv.org/abs/1905.13686.
  There is no corresponding record for this reference.
225. 225
  Bai, Y.; Wilbraham, L.; Slater, B. J.; Zwijnenburg, M. A.; Sprick, R. S.; Cooper, A. I. Accelerated discovery of organic polymer photocatalysts for hydrogen evolution from water through the integration of experiment and theory. J. Am. Chem. Soc. 2019, 141 (22), 9063– 9071, DOI: 10.1021/jacs.9b03591
  There is no corresponding record for this reference.
226. 226
  Vosko, S. H.; Wilk, L.; Nusair, M. Accurate spin-dependent electron liquid correlation energies for local spin density calculations: a critical analysis. Can. J. Phys. 1980, 58 (8), 1200– 1211, DOI: 10.1139/p80-159
  There is no corresponding record for this reference.
227. 227
  Becke, A. D. Density-functional thermochemistry. iii. the role of exact exchange. J. Chem. Phys. 1993, 98 (7), 5648– 5652, DOI: 10.1063/1.464913
  There is no corresponding record for this reference.
228. 228
  Lee, C.; Yang, W.; Parr, R. G. Development of the colle-salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B 1988, 37 (2), 785, DOI: 10.1103/PhysRevB.37.785
  There is no corresponding record for this reference.
229. 229
  Grimme, S.; Bannwarth, C.; Shushkov, P. A robust and accurate tight-binding quantum chemical method for structures, vibrational frequencies, and noncovalent interactions of large molecular systems parametrized for all spd-block elements (z= 1–86). J. Chem. Theory Comput. 2017, 13 (5), 1989– 2009, DOI: 10.1021/acs.jctc.7b00118
  There is no corresponding record for this reference.
230. 230
  Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. Ab initio calculation of vibrational absorption and circular dichroism spectra using density functional force fields. J. Phys. Chem. A 1994, 98 (45), 11623– 11627, DOI: 10.1021/j100096a001
  There is no corresponding record for this reference.
231. 231
  Yang, K.; Swanson, K.; Jin, W.; Coley, C.; Eiden, P.; Gao, H.; Guzman-Perez, A.; Hopper, T.; Kelley, B.; Mathea, M. Analyzing learned molecular representations for property prediction. J. Chem. Inf. Model. 2019, 59 (8), 3370– 3388, DOI: 10.1021/acs.jcim.9b00237
  There is no corresponding record for this reference.
232. 232
  Rogers, D.; Hahn, M. Extended-connectivity fingerprints. J. Chem. Inf. Model. 2010, 50 (5), 742– 754, DOI: 10.1021/ci100050t
  There is no corresponding record for this reference.
233. 233
  Landrum, G. Rdkit: open-source cheminformatics from machine learning to chemical registration. In Abstracts of Papers of the American Chemical Society; Vol. 258, AMER CHEMICAL SOC 1155 16TH ST, NW, WASHINGTON, DC 20036 USA, 2019.
  There is no corresponding record for this reference.
234. 234
  Mannodi-Kanakkithodi, A.; Chandrasekaran, A.; Kim, C.; Huan, T. D.; Pilania, G.; Botu, V.; Ramprasad, R. Scoping the polymer genome: A roadmap for rational polymer dielectrics design and beyond. Mater. Today 2018, 21, 785– 796, DOI: 10.1016/j.mattod.2017.11.021
  There is no corresponding record for this reference.
235. 235
  Kim, C.; Chandrasekaran, A.; Huan, T. D.; Das, D.; Ramprasad, R. Polymer genome: A data-powered polymer informatics platform for property predictions. J. Phys. Chem. C 2018, 122, 17575– 17585, DOI: 10.1021/acs.jpcc.8b02913
  There is no corresponding record for this reference.
236. 236
  Ward, L.; Dunn, A.; Faghaninia, A.; Zimmermann, N. E.; Bajaj, S.; Wang, Q.; Montoya, J.; Chen, J.; Bystrom, K.; Dylla, M.; Chard, K.; Asta, M.; Persson, K. A.; Snyder, G. J.; Foster, I.; Jain, A. Matminer: An open source toolkit for materials data mining. Comput. Mater. Sci. 2018, 152, 60– 69, DOI: 10.1016/j.commatsci.2018.05.018
  There is no corresponding record for this reference.
237. 237
  Lightstone, J. P.; Chen, L.; Kim, C.; Batra, R.; Ramprasad, R. Refractive index prediction models for polymers using machine learning J. Appl. Phys. 2020; Vol. 127, DOI: 10.1063/5.0008026 .
  There is no corresponding record for this reference.
238. 238
  Chen, L.; Kim, C.; Batra, R.; Lightstone, J. P.; Wu, C.; Li, Z.; Deshmukh, A. A.; Wang, Y.; Tran, H. D.; Vashishta, P.; Sotzing, G. A.; Cao, Y.; Ramprasad, R. Frequency-dependent dielectric constant prediction of polymers using machine learning. npj Computat. Mater. 2020, 6 (1), 61, DOI: 10.1038/s41524-020-0333-6
  There is no corresponding record for this reference.
239. 239
  Doan Tran, H.; Kim, C.; Chen, L.; Chandrasekaran, A.; Batra, R.; Venkatram, S.; Kamal, D.; Lightstone, J. P.; Gurnani, R.; Shetty, P. Machine-learning predictions of polymer properties with polymer genome. J. Appl. Phys. 2020, 128 (17), 171104 DOI: 10.1063/5.0023759
  There is no corresponding record for this reference.
240. 240
  Godwin, J.; Schaarschmidt, M.; Gaunt, A.; Sanchez-Gonzalez, A.; Rubanova, Y.; Veličković, P.; Kirkpatrick, J.; Battaglia, P. Simple gnn regularisation for 3d molecular property prediction & beyond, 2021, arXiv:2106.07971. arXiv.org e-Print archive https://arxiv.org/abs/2106.07971.
  There is no corresponding record for this reference.
241. 241
  Chen, D.; Lin, Y.; Li, W.; Li, P.; Zhou, J.; Sun, X. Measuring and relieving the over-smoothing problem for graph neural networks from the topological view. Proc. AAAI Conf. Artificial Intelligence 2020, 34, 3438– 3445, DOI: 10.1609/aaai.v34i04.5747
  There is no corresponding record for this reference.
242. 242
  Frisch, M. Gaussian 09, Revision D.01, Gaussian , 2009.
  There is no corresponding record for this reference.
243. 243
  Kuenneth, C.; Schertzer, W.; Ramprasad, R. Copolymer informatics with multitask deep neural networks. Macromolecules 2021, 54 (13), 5957– 5961, DOI: 10.1021/acs.macromol.1c00728
  There is no corresponding record for this reference.
244. 244
  You, J.; Ying, Z.; Leskovec, J. Design space for graph neural networks. In Advances in Neural Information Processing Systems (NeurIPS) 2020.
  There is no corresponding record for this reference.
245. 245
  Rupp, M.; Tkatchenko, A.; Müller, K.-R.; Von Lilienfeld, O. A. Fast and accurate modeling of molecular atomization energies with machine learning. Phys. Rev. Lett. 2012, 108 (5), 058301 DOI: 10.1103/PhysRevLett.108.058301
  There is no corresponding record for this reference.
246. 246
  Bartók, A. P.; Kondor, R.; Csányi, G. On representing chemical environments. Phys. Rev. B 2013, 87 (18), 184115 DOI: 10.1103/PhysRevB.87.184115
  There is no corresponding record for this reference.
247. 247
  Townsend, J.; Micucci, C. P.; Hymel, J. H.; Maroulas, V.; Vogiatzis, K. D. Representation of molecular structures with persistent homology for machine learning applications in chemistry. Nat. Commun. 2020, 11 (1), 3230 DOI: 10.1038/s41467-020-17035-5
  There is no corresponding record for this reference.
248. 248
  Huo, H.; Rupp, M. Unified representation of molecules and crystals for machine learning. Machine Learning: Sci. Technol. 2022, 3 (4), 045017 DOI: 10.1088/2632-2153/aca005
  There is no corresponding record for this reference.
249. 249
  Sanchez Medina, E. I.; Linke, S.; Stoll, M.; Sundmacher, K. Gibbs-helmholtz graph neural network: capturing the temperature dependency of activity coefficients at infinite dilution. Digital Discovery 2023, 2 (3), 781– 798, DOI: 10.1039/D2DD00142J
  There is no corresponding record for this reference.
250. 250
  Zhong, C.; Sato, Y.; Masuoka, H.; Chen, X. Improvement of predictive accuracy of the unifac model for vapor-liquid equilibria of polymer solutions. Fluid Phase Equilib. 1996, 123 (1–2), 97– 106, DOI: 10.1016/S0378-3812(96)90015-1
  There is no corresponding record for this reference.
251. 251
  Kontogeorgis, G. M.; Fredenslund, A.; Tassios, D. P. Simple activity coefficient model for the prediction of solvent activities in polymer solutions. Ind. Eng. Chem. Res. 1993, 32 (2), 362– 372, DOI: 10.1021/ie00014a013
  There is no corresponding record for this reference.
252. 252
  Pappa, G. D.; Voutsas, E. C.; Tassios, D. P. Prediction of activity coefficients in polymer and copolymer solutions using simple activity coefficient models. Ind. Eng. Chem. Res. 1999, 38 (12), 4975– 4984, DOI: 10.1021/ie990265q
  There is no corresponding record for this reference.
253. 253
  Cremer, D. Møller-plesset perturbation theory: from small molecule methods to methods for thousands of atoms. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2011, 1 (4), 509– 530, DOI: 10.1002/wcms.58
  There is no corresponding record for this reference.
254. 254
  Kearnes, S.; McCloskey, K.; Berndl, M.; Pande, V.; Riley, P. Molecular graph convolutions: moving beyond fingerprints. J. Comput.-Aided Mol. Des. 2016, 30, 595– 608, DOI: 10.1007/s10822-016-9938-8
  There is no corresponding record for this reference.
255. 255
  Wang, M. Y. Deep graph library: Towards efficient and scalable deep learning on graphs. In ICLR Workshop on Representation Learning on Graphs and Manifolds , 2019.
  There is no corresponding record for this reference.
256. 256
  Bojar, D.; Powers, R. K.; Camacho, D. M.; Collins, J. J. Deep-learning resources for studying glycan-mediated host-microbe interactions. Cell Host Microbe 2021, 29, 132– 144.e3, DOI: 10.1016/j.chom.2020.10.004
  There is no corresponding record for this reference.
257. 257
  Montavon, G.; Binder, A.; Lapuschkin, S.; Samek, W.; Müller, K.-R. Layer-wise relevance propagation: an overview. In Lecture Notes in Computer Science 2019; pp 193– 209.
  There is no corresponding record for this reference.
258. 258
  Binder, A.; Bach, S.; Montavon, G.; Müller, K.-R.; Samek, W. Layer-wise relevance propagation for deep neural network architectures. In Information Science and Applications (ICISA) 2016; Springer, 2016; pp 913– 922.
  There is no corresponding record for this reference.
259. 259
  Adebayo, J.; Gilmer, J.; Muelly, M.; Goodfellow, I.; Hardt, M.; Kim, B. Sanity checks for saliency maps Advances in Neural Information Processing Systems , 2018 31.
  There is no corresponding record for this reference.
260. 260
  Qi, Z.; Khorram, S.; Li, F. Visualizing deep networks by optimizing with integrated gradients. AAAI 2020, 34, 11890– 11898, DOI: 10.1609/aaai.v34i07.6863
  There is no corresponding record for this reference.

Graph Neural Networks for Polymer Characterization and Property Prediction: Opportunities and ChallengesClick to copy article linkArticle link copied!

Journal of Chemical Information and Modeling

License Summary*

Abstract

License Summary*

License Summary*

License Summary*

1. Introduction

2. GNN Architectures

Figure 1

2.1. Message Passing Neural Networks

Figure 2

2.2. Graph Convolutional Networks (GCN)

2.3. GAT and Attention-Based Architectures

Figure 3

3. Data Sets

Figure 4

4. Coarse-Graining in Polymer Informatics

5. Applications and Implementations

5.1. Electronic and Optical Property Prediction

5.2. Thermal and Mechanical Properties

5.3. Transport and Dynamic Properties

5.4. Biological Properties

6. Discussion and Outlook

6.1. Understanding GNNs: Tools and Insights

6.2. Coarse-Graining in Graph Design

6.3. Outlook: from Prediction to Design

7. Conclusions

Author Information

References

Cited By

Journal of Chemical Information and Modeling

License Summary*

Article Views

Altmetric

Citations

Recommended Articles

Abstract

Figure 1

Figure 2

Figure 3

Figure 4

References

Graph Neural Networks for Polymer Characterization and Property Prediction: Opportunities and Challenges
Click to copy article linkArticle link copied!