Open Polymer Challenge: Post-Competition Report

2025-12-09

AI Summary

Current polymer AI research is hindered by the scarcity of high-quality, open-source datasets. To address this, we introduce PolyBench—the first open-source polymer informatics benchmark tailored for sustainable polymeric materials—comprising 10,000 polymers annotated with five critical physicochemical properties, including thermal conductivity, and explicitly designed for few-shot, multi-task, and cross-source heterogeneous prediction scenarios. Methodologically, we propose polymer-specific feature enhancement, self-supervised pretraining, and targeted ensemble modeling, complemented by the ADEPT framework to generate molecular dynamics simulation data and mitigate distribution shift. We release a standardized test suite alongside reproducible, multi-model baselines. Our approach significantly improves property prediction accuracy, establishes methodological standards for polymer AI modeling, and accelerates virtual screening of energy-efficient polymeric materials.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Machine learning (ML) offers a powerful path toward discovering sustainable polymer materials, but progress has been limited by the lack of large, high-quality, and openly accessible polymer datasets. The Open Polymer Challenge (OPC) addresses this gap by releasing the first community-developed benchmark for polymer informatics, featuring a dataset with 10K polymers and 5 properties: thermal conductivity, radius of gyration, density, fractional free volume, and glass transition temperature. The challenge centers on multi-task polymer property prediction, a core step in virtual screening pipelines for materials discovery. Participants developed models under realistic constraints that include small data, label imbalance, and heterogeneous simulation sources, using techniques such as feature-based augmentation, transfer learning, self-supervised pretraining, and targeted ensemble strategies. The competition also revealed important lessons about data preparation, distribution shifts, and cross-group simulation consistency, informing best practices for future large-scale polymer datasets. The resulting models, analysis, and released data create a new foundation for molecular AI in polymer science and are expected to accelerate the development of sustainable and energy-efficient materials. Along with the competition, we release the test dataset at https://www.kaggle.com/datasets/alexliu99/neurips-open-polymer-prediction-2025-test-data. We also release the data generation pipeline at https://github.com/sobinalosious/ADEPT, which simulates more than 25 properties, including thermal conductivity, radius of gyration, and density.

Problem

• Addresses the lack of large, high-quality, open polymer datasets
• Focuses on multi-task polymer property prediction for virtual screening
• Develops models under realistic constraints like small data and imbalance

Innovation

• Multi-task polymer property prediction using machine learning models
• Feature-based augmentation and transfer learning for small datasets
• Self-supervised pretraining and targeted ensemble strategies applied