Quadruple Attention in Many-body Systems for Accurate Molecular Property Predictions

Jiahua Rao, Dahao Xu, Wentao Wei, Yicong Chen, Mingjun Yang, Yuedong Yang

Abstract: While Graph Neural Networks and Transformers have shown promise in predicting molecular properties, they struggle with directly modeling complex many-body interactions. Current methods often approximate interactions like three- and four-body terms in message passing, while attention-based models, despite enabling direct atom communication, are typically limited to triplets, making higher-order interactions computationally demanding. To address the limitations, we introduce MABNet, a geometric attention framework designed to model four-body interactions by facilitating direct communication among atomic quartets. This approach bypasses the computational bottlenecks associated with traditional triplet-based attention mechanisms, allowing for the efficient handling of higher-order interactions. MABNet achieves state-of-the-art performance on benchmarks like MD22 and SPICE. These improvements underscore its capability to accurately capture intricate many-body interactions in large molecules. By unifying rigorous many-body physics with computational efficiency, MABNet advances molecular simulations for applications in drug design and materials discovery, while its extensible framework paves the way for modeling higher-order quantum effects.

Lay Summary: Predicting molecular properties is crucial for advancements in drug discovery and materials science. However, current computational methods face challenges in accurately capturing complex interactions between multiple atoms, especially when these interactions involve more than three atoms. To overcome this, we developed MABNet, a novel framework that introduces a new way for atoms to communicate directly in groups of four. This innovation allows our model to better understand and simulate intricate molecular behaviors without the usual computational hurdles. By achieving state-of-the-art results on key benchmarks, MABNet demonstrates its potential to enhance molecular simulations and accelerate discoveries in chemistry and biology.

Primary Area: Applications->Chemistry, Physics, and Earth Sciences

Keywords: Molecular Property Prediction, Many-body System, Quadruple Attention, Molecular Force Fields

Submission Number: 1358