Building informative materials datasets beyond targeted objectives

TL;DR

This paper introduces a diversity-aware framework for constructing materials datasets that maximizes informativeness for specific properties while maintaining broad coverage for future learning, improving prediction performance and reducing bias.

Contribution

The authors propose a novel diversity-aware selection method for dataset construction that enhances informativeness across multiple properties and mitigates cold-start issues in materials discovery.

Findings

Prediction performance on untargeted properties improved by up to 10% with the framework.

Performance on targeted properties can be increased by up to 25% using the proposed method.

Diversity-aware dataset construction prevents degradation of predictive accuracy compared to random sampling.

Abstract

Materials science data collection can be expensive, making the reuse and long-term utility of datasets critical important for future discovery campaigns. In practice, researchers prioritize a subset of properties due to research interests. However, ignoring a subset of outcomes in data collection campaigns potentially generate datasets poorly suited for future learning tasks. Here, we present a framework for dataset construction that maximizes informativeness for target properties of interest while preserving performance on untargeted ones. Our approach uses diversity-aware selection to ensure broad coverage of the materials space. In noisy experimental dataset construction, we find that without our diversity-aware framework, prediction performance on untargeted properties can degrade by up to 40% relative to random sampling, whereas applying our framework yields improvements of up to 10% . For targeted properties, performance can degrade with respect to random sampling by up to 12.5% without diversity, while our framework achieves gains of up to 25%. Incorporating diversity into dataset construction not only preserves informativeness for the targeted properties, but also improves materials coverage for potential future objectives. As a result, the constructed datasets remain broadly informative across considered and unconsidered outcomes, ensuring unbiased quality entries and mitigating cold-start limitations in subsequent modeling and discovery campaigns.