Deep learning methods for material property prediction have been widely explored to advance materials discovery. However, the prevailing pre-train then fine-tune paradigm often fails to address the inherent diversity and disparity of material tasks. To overcome these challenges, we introduce MoMa, a Modular framework for Materials that first trains specialized modules across a wide range of tasks and then adaptively composes synergistic modules tailored to each downstream scenario. Evaluation across 17 datasets demonstrates the superiority of MoMa, with a substantial 14% average improvement over the strongest baseline. Few-shot and continual learning experiments further highlight MoMa's potential for real-world applications. Pioneering a new paradigm of modular material learning, MoMa will be open-sourced to foster broader community collaboration.
@article{arxiv.2502.15483,
title = {MoMa: A Modular Deep Learning Framework for Material Property Prediction},
author = {Botian Wang and Yawen Ouyang and Yaohui Li and Mianzhi Pan and Yuanhang Tang and Yiqun Wang and Haorui Cui and Jianbing Zhang and Xiaonan Wang and Wei-Ying Ma and Hao Zhou},
journal= {arXiv preprint arXiv:2502.15483},
year = {2026}
}