PolyMon: A Unified Framework for Polymer Property Prediction

Gaopeng Ren; Yijie Yang; Jiajun Zhou; Kim E. Jelfs

cond-mat.softMar 2, 2026

PolyMon: A Unified Framework for Polymer Property Prediction

Authors: Gaopeng Ren, Yijie Yang, Jiajun Zhou, Kim E. Jelfs

Organizations: Department of Chemistry, Imperial College London

Abstract

Accurate prediction of polymer properties is essential for materials design, but remains challenging due to data scarcity, diverse polymer representations, and the lack of systematic evaluation across modelling choices. Here, we present PolyMon, a unified and accessible framework that integrates multiple polymer representations, machine learning methods, and training strategies within a single, accessible platform. PolyMon supports various descriptors and graph construction strategies for polymer representations, and includes a wide range of models, from tabular models to graph neural networks, along with flexible training strategies including multi-fidelity learning, Δ-learning, active learning, and ensemble learning. Using five key polymer properties as benchmarks, we perform systematic evaluations to assess how representations and models affect predictive performance. These case studies further illustrate how different training strategies can be applied within a consistent workflow to leverage limited data and incorporate physical model derived information. Overall, PolyMon provides a comprehensive and extensible foundation for benchmarking and advancing machine learning-based polymer property prediction. The code is available at github.com/fate1997/polymon.

Figures & tables

Dataset	Description	Unit	# data points	Source
$T_{g}$	Glass transition temperature	∘ C	7,367	[ 38 ]
FFV	Fractional free volume	1	7,030	[ 41 ]
TC	Thermal conductivity	$\rm W/(m\cdot K)$	1,070	[ 26 ]
$\rho_{\rm md}$	MD-calculated density	$\rm g/cm^{3}$	1,077	[ 26 ]
$R_{g}$	Radius of gyration	Å	1,328	[ 26 , 41 ]
$\rho_{\rm exp}$	Experimentally measured density	$\rm g/cm^{3}$	117	[ 2 ]

Table 1: Dataset summary

Appendix figures & tables22 assets

Supplementary material from the paper’s appendix.

Appendix

Argument	Type	Description
--raw-csv	str	Path to the raw CSV file. The CSV file should contain the SMILES column
--sources	str (multiple)	Sources to use for training (only when the input CSV file contains the Source column.)
--tag	str	Tag to use for training.
--labels	str (multiple)	Labels to be predict. The available choice is from the CSV columns.
--feature-names	str (multiple)	Feature names to use for training. Available features can be found in Table 2 .
--n-trials	int	Number of trials for hyperparameter optimisation.

Appendix

Table 1: Arguments for the polymon train command.

Descriptor Name	Description
xenonpy_atom	XenonPy atom features
cgcnn	CGCNN atom features
source	Source labels as features
z	Atomic numbers as integers
edge	Default edge features ( bond )
bond	Chemical bond features

Appendix

Table 2: Available feature names in PolyMon

Model Type	Model Name	Description
GNNs	gatv2	Graph Attention Network v2
	attentivefp	AttentiveFP
	dimenetpp	DimeNet++
	gatv2vn	GATv2 with virtual node
	gin	Graph Isomorphism Network
	pna	Principal Neighbourhood Aggregation

Appendix

Table 3: Available models in PolyMon.

Hyperparameter	Description	Search Space
n_estimators	Number of estimators	1000
learning_rate	Boosting learning rate	$\log$ -uniform $(10^{-3},10^{-1})$
max_depth	Maximum tree depth	Integer $[1,10]$
subsample	Row subsampling ratio	Uniform $(0.05,1.0)$
colsample_bytree	Feature subsampling ratio	Uniform $(0.05,1.0)$
min_child_weight	Minimum child weight	Integer $[1,20]$

Appendix

Table 4: Hyperparameter search space for XGBoost.

Hyperparameter	Description	Search Space
n_estimators	Number of trees	Integer $[100,1000]$
max_depth	Maximum tree depth	Integer $[2,32]$
min_samples_split	Minimum samples required to split a node	Integer $[2,10]$
min_samples_leaf	Minimum samples required at a leaf	Integer $[1,10]$

Appendix

Table 5: Hyperparameter search space for Random Forest.

Hyperparameter	Description	Search Space
n_estimators	Number of estimators	Integer $[1000,10000]$
learning_rate	Boosting learning rate	$\{0.006,0.008,0.01,0.014,0.017,0.02\}$
max_depth	Maximum tree depth	10, 20, 100
num_leaves	Number of leaves	Integer $[1,1000]$
reg_alpha	L1 regularization strength	$\log$ -uniform $(10^{-3},10)$
reg_lambda	L2 regularization strength	$\log$ -uniform $(10^{-3},10)$

Appendix

Table 6: Hyperparameter search space for LightGBM.

Hyperparameter	Description	Search Space
learning_rate	Learning rate	Discrete grid: $[0.001,0.020]$ with step $0.001$
depth	Tree depth	Integer $[9,15]$
l2_leaf_reg	L2 regularization strength	Discrete grid: $[1.0,5.5]$ with step $0.5$
min_child_samples	Minimum samples per leaf	$\{1,4,8,16,32\}$
grow_policy	Tree growing policy	Depthwise
iterations	Number of iterations for training	10000

Appendix

Table 7: Hyperparameter search space for CatBoost.

Hyperparameter	Description	Search Space
n_estimators	Number of ensemble estimators	Integer $[8,64]$
softmax_temperature	Softmax temperature	$\{0.75,0.8,0.85,0.9,0.95,1.0\}$
average_before_softmax	Logit averaging strategy	{True, False}

Appendix

Table 8: Hyperparameter search space for TabPFN.

Hyperparameter	Description	Search Space
hidden_dim	Hidden feature dimension	Discrete grid: $[32,512]$ with step $32$
num_layers	Number of hidden layers in the FastKAN network	Integer: $[2,5]$
grid_min	Minimum value of the spline grid range	Integer” $[-5.0,-2.0]$
grid_max	Maximum value of the spline grid range	Integer: $[2.0,5.0]$
num_grids	Number of spline interpolation grids	Discrete grid: $[6,10]$ with step $2$

Appendix

Table 9: Hyperparameter search space for FastKAN.

Hyperparameter	Description	Search Space
hidden_dim	Hidden feature dimension	Discrete grid: $[32,512]$ with step $32$
num_layers	Number of hidden layers in the FastKAN network	Integer: $[2,5]$

Appendix

Table 10: Hyperparameter search space for FourierKAN.

Hyperparameter	Description	Search Space
hidden_dim	Hidden feature dimension	Discrete grid: $[32,512]$ with step $32$
num_layers	Number of hidden layers in the FastKAN network	Integer: $[2,5]$

Appendix

Table 11: Hyperparameter search space for EfficientKAN.

Hyperparameter	Description	Search Space
hidden_dim	Hidden feature dimension	$\{16,32,48,64\}$
num_layers	Number of graph attention layers	Integer $[2,8]$
num_heads	Number of attention heads	$\{4,8\}$
pred_hidden_dim	Hidden dimension of the prediction head	Distance grid: $[16,256]$ with step $16$
pred_dropout	Dropout rate in the prediction head	$\{0.1,0.2,0.3,0.4,0.5\}$
pred_layers	Number of layers in the prediction head	$\{1,2,3\}$

Appendix

Table 12: Hyperparameter search space for GATv2.

Hyperparameter	Description	Search Space
hidden_dim	Hidden feature dimension	Discrete grid: $[16,256]$ with step $16$
num_layers	Number of GIN layers	$\{2,3,4\}$
pred_hidden_dim	Hidden dimension of the prediction (readout) head	Discrete grid: $[16,256]$ with step $16$
pred_dropout	Dropout rate applied to the prediction head	$\{0.0,0.1,0.2,0.3,0.4,0.5\}$
pred_layers	Number of layers in the prediction head	$\{1,2,3\}$
n_mlp_layers	Number of MLP layers per GIN block	$\{1,2,3\}$

Appendix

Table 13: Hyperparameter search space for the Graph Isomorphism Network (GIN).

Hyperparameter	Description	Search Space
in_channels	Dimension of input node feature vectors	Distance grid: $[16,256]$ with step $16$
hidden_dim	Hidden feature dimension	Discrete grid: $[16,256]$ with step $16$
num_layers	Number of GCN layers	$\{2,3,4\}$
dropout	Dropout rate applied to node features	$\{0.0,0.1,0.2,0.3,0.4,0.5\}$

Appendix

Table 14: Hyperparameter search space for the Graph Convolutional Network (GCN).

Hyperparameter	Description	Search Space
hidden_dim	Hidden feature dimension	Discrete grid: $[16,256]$ with step $16$
towers	Number of aggregation towers	$\{1,2,4,8\}$
num_layers	Number of PNA layers	Integer $[2,4]$
pred_hidden_dim	Hidden dimension of the prediction (readout) head	Discrete grid: $[16,256]$ with step $16$
pred_dropout	Dropout rate applied to the prediction head	$[0.0,0.1,0.2,0.3,0.4,0.5]$
pred_layers	Number of layers in the prediction head	$[1,2,3]$

Appendix

Table 15: Hyperparameter search space for PNA.

Hyperparameter	Description	Search Space
hidden_dim	Hidden feature dimension	Discrete grid: $[16,256]$ with step $16$
num_layers	Number of interaction layers	Integer $[2,4]$
num_spherical	Number of spherical basis functions	Integer $[5,10]$
num_radial	Number of radial basis functions	$\{4,6,8,10,12\}$
cutoff	Distance cutoff for message passing	$\{2.0,2.5,3.0,3.5,4.0,4.5,5.0\}$
int_emb_size	Dimension of interaction embedding layers	Discrete grid: $[8,64]$ with step $8$

Appendix

Table 16: Hyperparameter search space for DimeNet++.

Hyperparameter	Description	Search Space
in_channels	Dimension of input node feature vectors	Discrete grid: $[16,256]$ with step $16$
hidden_dim	Hidden feature dimension	Discrete grid: $[16,256]$ with step $16$
num_layers	Number of message-passing layers	$\{1,2,3\}$
num_timesteps	Number of iterative attention refinement steps	$\{1,2,3\}$

Appendix

Table 17: Hyperparameter search space for AttentiveFP.

Hyperparameter	Description	Search Space
hidden_dim	Hidden feature dimension	Distance grid: $[16,512]$ with step $16$
num_layers	Number of graph attention layers	Integer $[2,6]$
num_heads	Number of attention heads	$\{4,8\}$
pred_hidden_dim	Hidden dimension of the prediction head	Distance grid: $[128,1024]$ with step $128$
pred_dropout	Dropout rate in the prediction head	$\{0.0,0.1,0.2,0.3,0.4,0.5\}$
pred_layers	Number of layers in the prediction head	$\{1,2,3\}$

Appendix

Table 18: Hyperparameter search space for Graph Transformers(GT).

Hyperparameter	Description	Search Space
hidden_dim	Hidden feature dimension	Discrete grid: $[16,256]$ with step $16$
num_layers	Number of stacked GPS layers	Discrete grid: $[2,10]$ with step $1$
pe_dim	Dimension of positional encodings	$\{4,5,6,7,8\}$
heads	Number of attention heads in the transformer component	$\{4,8\}$
attn_type	Type of attention mechanism used in GPS	{ performer , multihead }
attention_dropout	Dropout probability used inside the attention module	$\{0.0,0.1,0.2,0.3,0.4,0.5\}$

Appendix

Table 19: Hyperparameter search space for the Graph Positioning System (GPS) model.

Model	Descriptor	$\rho_{\rm{md}}$	FFV	$R_{g}$	TC	$T_{g}$
CatBoost	ECFP4	$0.0358_{\pm 0.0033}$	$0.0072_{\pm 0.0004}$	$2.134_{\pm 0.191}$	$0.0214_{\pm 0.0012}$	$23.86_{\pm 0.30}$
	MACCS	$0.0322_{\pm 0.0047}$	$0.0082_{\pm 0.0003}$	$2.502_{\pm 0.184}$	$0.0243_{\pm 0.0011}$	$28.09_{\pm 0.70}$
	Mordred	$0.0231_{\pm 0.0037}$	$0.0059_{\pm 0.0005}$	$2.293_{\pm 0.198}$	$0.0199_{\pm 0.0007}$	$23.85_{\pm 0.47}$
	Oligomer-ECFP4	$0.0372_{\pm 0.0039}$	$0.0075_{\pm 0.0004}$	$2.234_{\pm 0.253}$	$0.0209_{\pm 0.0011}$	$24.22_{\pm 0.44}$
	Oligomer-Mordred	$0.0212_{\pm 0.0037}$	$0.0056_{\pm 0.0005}$	$2.334_{\pm 0.232}$	$0.0194_{\pm 0.0008}$	$23.31_{\pm 0.46}$
	Oligomer-RDKit2D	$0.0226_{\pm 0.0030}$	$0.0061_{\pm 0.0005}$	$2.355_{\pm 0.221}$	$0.0213_{\pm 0.0009}$	$25.19_{\pm 0.42}$

Appendix

Table 20: Prediction performance (mean

{}_{\pm\,\mathrm{std}}

) of different models and molecular descriptors.

Model	$\rho_{\rm md}$	FFV	$R_{g}$	TC	$T_{g}$
GATv2	$0.0136_{\pm 0.0005}$	$0.0046_{\pm 0.0001}$	$1.7065_{\pm 0.0711}$	$0.0172_{\pm 0.0007}$	$22.4999_{\pm 0.7899}$
gatv2_chain_readout	$0.0166_{\pm 0.0010}$		$1.6924_{\pm 0.0852}$	$0.0186_{\pm 0.0018}$
gatv2_pe	$0.0143_{\pm 0.0007}$	$0.0048_{\pm 0.0002}$	$1.8040_{\pm 0.0953}$	$0.0181_{\pm 0.0026}$	$25.9572_{\pm 0.9894}$
DMPNN	$0.0114_{\pm 0.0006}$		$1.7144_{\pm 0.0926}$	$0.0157_{\pm 0.0019}$

Appendix

Table 21: Performance of additional GNN models on polymer property prediction (* indicates that training was not completed within 72 hours).

Figure 1: Performance of different empirical equations and atomic contribution methods.

Polymer Property Prediction

See all 13 papers

Sep 22, 2026

An open benchmark for machine learning-based polymer property prediction

Polymer property prediction lacks open, standardized benchmarks that enable rigorous comparison of machine-learning methods, with existing resources covering only a narrow fraction of polymer architectures, such as homopolymers. We introduce Polymer Benchmark 2026 (PolyBench26), an open dataset comprising nearly 250,000 polymer-property datapoints across eight physical properties, including data from experimental measurements, density functional theory, and molecular dynamics. The benchmark supports four evaluation tasks across homopolymers and alternating, random, and block copolymers: in-distribution property prediction, dataset-size scaling, repeat-unit complexity, and transfer to held-out polymer architectures. We compare language model, graph-based, and descriptor-based approaches and find graph-based models provide the lowest errors in property prediction, retain their advantage across the evaluated training-set sizes, and remain robust to increasing repeat-unit complexity. PolyBench26 provides a reproducible foundation for developing models for the increasingly complex polymer design space. The PolyBench26 benchmark is available open-source at https://github.com/rlearsch/PolymerBenchmark2026.

Robert W. Learsch, Nicholas Liesen, Daniel S. Levine +2

Lawrence Livermore National Laboratory · FAIR at Meta

Materials Property Prediction Polymer Property Prediction

Ask PDF

May 7, 2026

Can LLMs Predict Polymer Physics Just by Reading Synthesis and Processing Prose?

Can large language models predict physical and mechanical polymer properties simply by reading unstructured scientific prose? Polymer performance is rarely determined by chemical structure alone; identical nominal polymers can exhibit drastically different behaviors depending on their synthesis route, processing history, morphology, and testing conditions. Yet, state-of-the-art polymer property models typically rely on structure-only representations -- such as SMILES or molecular graphs -- which strip away this vital experimental context. In this work, we introduce PolyLM, a natural-language-only, process- and condition-aware framework that predicts materials performance directly from full-text literature. By circumventing structural inputs entirely, PolyLM preserves the nuanced, unstructured descriptions of synthesis and processing reported by domain scientists. To train this framework, we curated an unprecedented, literature-scale dataset encompassing 185,000 scientific papers and over 276,400 unique polymer samples across 22 physical, mechanical, and thermal properties. We fine-tuned a massive 9-billion-parameter language model (Qwen3.5-9B) using Low-Rank Adaptation (LoRA) and task-level uncertainty weighting. Evaluated on 68,283 held-out observations, the model achieves remarkably high predictive accuracy, establishing new state-of-the-art benchmarks for complex properties. Across the 22 diverse targets, the model achieves a median

R^2

of 0.74, with predictions for key thermal, mechanical, and physicochemical properties frequently surpassing an

R^2

of 0.80. These results unequivocally demonstrate that natural language is a powerful, highly scalable interface for realistic materials performance prediction.

Yuchu Liu, Rui Zhu, Jingwei Xiong +1

Yale University · UC Davis · Indiana University Bloomington

Materials Property Prediction Polymer Property Prediction

Ask PDF

Aug 2, 2026

PolymerGPT: Multi-property Optimization with a Decoder-Based GPT Model for Generative Polymer Design

Polymer property prediction and inverse generative design targeting desired properties are two crucial tasks in machine learning-assisted polymer design. While the former has received considerable attention, there have been limited methods developed for the latter. Existing methods focus on single-property optimization in the generative process, whereas accurate prediction of macroscopic material behavior requires simultaneous control of multiple physical properties. In this paper, we provide a transformative framework for direct optimization of a large collection of polymer properties. We propose PolymerGPT, a decoder-based GPT model that incorporates up to 37 commonly used polymer properties into the generative process via learned conditioning prefixes. It also supports a scaffold condition that specifies a desired scaffold for predicted structures. Our experimental results demonstrate that PolymerGPT achieves exceptional performance for unconditional and conditional generation while maintaining high validity, uniqueness, and novelty. Conditioning on five key properties yields generated structures whose predicted values closely match all target properties simultaneously.

Charlie Pyle, Adarsh Gadari, C. Adrian Figg +3

Texas A&M University · University of Pittsburgh · Department of Chemistry, Virginia Tech +2

Molecular Optimization Molecular Generation Conditional Generative Modeling

Ask PDF

Abstract

Polymer Property Prediction

Explore similar work

An open benchmark for machine learning-based polymer property prediction

Can LLMs Predict Polymer Physics Just by Reading Synthesis and Processing Prose?

PolymerGPT: Multi-property Optimization with a Decoder-Based GPT Model for Generative Polymer Design