Multi-View Polymer Representations for the Open Polymer Prediction

2025-11-14

AI Summary

This work addresses polymer property prediction by proposing a multi-view molecular representation ensemble framework. Methodologically, it integrates four complementary modalities—RDKit/Morgan fingerprints, graph neural networks (GNNs), 3D conformation-aware representations, and pretrained SMILES language models—and employs uniform ensemble weighting alongside SMILES test-time augmentation to enhance generalization without complex hyperparameter tuning. Its key contribution is the first lightweight, highly robust multi-view ensemble paradigm specifically designed for polymer performance prediction, effectively leveraging multimodal structural information. Evaluated on the NeurIPS 2025 Open Polymer Prediction Challenge, the method ranked 9th among 2,241 teams, achieving MAEs of 0.057 and 0.082 on the public and private test sets, respectively—demonstrating state-of-the-art performance and practical applicability.

Technology Category

Machine Learning: Multi-instance/Multi-view LearningComputer Vision: Multi-modal VisionIntelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphsSearch and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAGSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

We address polymer property prediction with a multi-view design that exploits complementary representations. Our system integrates four families: (i) tabular RDKit/Morgan descriptors, (ii) graph neural networks, (iii) 3D-informed representations, and (iv) pretrained SMILES language models, and averages per-property predictions via a uniform ensemble. Models are trained with 10-fold splits and evaluated with SMILES test-time augmentation. The approach ranks 9th of 2241 teams in the Open Polymer Prediction Challenge at NeurIPS 2025. The submitted ensemble achieves a public MAE of 0.057 and a private MAE of 0.082.

Problem

• Predicting polymer properties using multi-view complementary representations
• Integrating four representation families through uniform ensemble averaging
• Evaluating model performance with cross-validation and test-time augmentation

Innovation

• Multi-view design integrates complementary polymer representations
• Uniform ensemble averages per-property predictions
• SMILES test-time augmentation enhances model evaluation