海归学者发起的公益学术平台
分享信息,整合资源
交流学术,偶尔风月
随着人工智能(AI)的迅猛发展,原子层面的建模、模拟与设计正经历深远的变革。基于机器学习的势能函数模型如今在精度上已可媲美从头算电子结构方法,并支持大规模、长时程模拟。然而,模型的生成与训练过程仍然是实现大规模应用的主要瓶颈。如何用同一个模型来解决不同材料体系、不同任务需求,一直是个备受关注的难题。多年来,人们先后尝试了多种方法,但仍缺少一种能够同时在化学和材料等多学科体系保持高精度和高效率、又能兼容不同计算任务的通用模型。
Fig. 1 | An overview of the proposed LAM workflow.
本工作由来自北京科学智能研究院、北京深势科技等29个机构的42位合作者(通讯作者为Linfeng Zhang和Han Wang)共同完成,提出了名为DPA-2的大原子模型新架构。与传统需要“单打独斗”的单一任务训练方式不同,DPA-2利用多任务预训练方法,一次性学习多种化学与材料体系(包括金属合金、电池材料、药物分子及铁电材料等共18个数据集、73种元素)的特征,这让它能够在面对“从未见过”的下游任务时给出更准确的预测。进一步而言,DPA-2大原子模型为后续的微调提供了极为便利的起点,从而在收集很少数据的情况下获得令人满意的结果,相比从头开始训练能将数据效率提升1-3个数量级,可大幅降低开发定制化模型的门槛,例如用于新型材料或新化合物的模拟与设计。
展望未来,随着更多多样化数据不断加入,DPA模型将为材料模拟、药物设计乃至化工过程的预测与优化提供更广阔的路径。该文近期发表于npj Computational Materials 10: 293 (2024),英文标题与摘要如下,点击左下角“阅读原文”可以自由获取论文PDF。
DPA-2: a large atomic model as a multi-task learner
Duo Zhang, Xinzijian Liu, Linfeng Zhang, Han Wang et al.
The rapid advancements in artificial intelligence (AI) are catalyzing transformative changes in atomic modeling, simulation, and design. AI-driven potential energy models have demonstrated the capability to conduct large-scale, long-duration simulations with the accuracy of ab initio electronic structure methods. However, the model generation process remains a bottleneck for large-scale applications. We propose a shift towards a model-centric ecosystem, wherein a large atomic model (LAM), pre-trained across multiple disciplines, can be efficiently fine-tuned and distilled for various downstream tasks, thereby establishing a new framework for molecular modeling. In this study, we introduce the DPA-2 architecture as a prototype for LAMs. Pre-trained on a diverse array of chemical and materials systems using a multi-task approach, DPA-2 demonstrates superior generalization capabilities across multiple downstream tasks compared to the traditional single-task pre-training and fine-tuning methodologies. Our approach sets the stage for the development and broad application of LAMs in molecular and materials simulation research.
扩展阅读