Machine Learning Force Fields with Data Cost Aware Training
Machine Learning Force Fields with Data Cost Aware Training
复制标题
DOI：
10.48550/arxiv.2306.03109
复制
发表时间：
2023-06
期刊：
影响因子：
--
通讯作者：
Alexander W. Bukharin;Tianyi Liu;Sheng Wang;Simiao Zuo;Weihao Gao;Wen Yan;Tuo Zhao
中科院分区：
文献类型：
--
作者：
Alexander W. Bukharin;Tianyi Liu;Sheng Wang;Simiao Zuo;Weihao Gao;Wen Yan;Tuo Zhao
已经提出了机器学习力场（MLFF）来加速分子动力学（MD）模拟，该模拟在化学和生物医学研究中发现了广泛的应用。即使对于最有效的MLFF，达到化学精度也可能需要数百帧由昂贵的量子机械算法产生的力和能量标签，该算法可能会扩展为$ O（n^3）$到$ O（n^7）$，$ n $与$ n $相比与基本功能的数量成比例。为了解决这个问题，我们提出了一个多阶段计算框架 - 小行星，该框架通过利用廉价的不准确数据和昂贵的准确数据来降低MLFF的数据成本。小行星背后的动机是，尽管存在大偏见，但不准确的数据可以帮助捕获基础力场的复杂结构。因此，我们首先采用大量不准确的培训数据训练MLFF模型，采用偏见感知损失函数，以防止该模型过度适合该数据的潜在偏差。然后，我们使用少量准确的培训数据对获得的模型进行了微调，从而保留了从不准确的培训数据中学到的知识，同时显着提高了模型的准确性。此外，我们根据不准确训练数据未标记的设置提出了一种基于分数匹配的小行星变体。在MD数据集和下游任务上进行的广泛实验验证了小行星的功效。我们的代码和数据可从https://github.com/abukharin3/asteroid获得。
Machine learning force fields (MLFF) have been proposed to accelerate molecular dynamics (MD) simulation, which finds widespread applications in chemistry and biomedical research. Even for the most data-efficient MLFFs, reaching chemical accuracy can require hundreds of frames of force and energy labels generated by expensive quantum mechanical algorithms, which may scale as $O(n^3)$ to $O(n^7)$, with $n$ proportional to the number of basis functions. To address this issue, we propose a multi-stage computational framework -- ASTEROID, which lowers the data cost of MLFFs by leveraging a combination of cheap inaccurate data and expensive accurate data. The motivation behind ASTEROID is that inaccurate data, though incurring large bias, can help capture the sophisticated structures of the underlying force field. Therefore, we first train a MLFF model on a large amount of inaccurate training data, employing a bias-aware loss function to prevent the model from overfitting tahe potential bias of this data. We then fine-tune the obtained model using a small amount of accurate training data, which preserves the knowledge learned from the inaccurate training data while significantly improving the model's accuracy. Moreover, we propose a variant of ASTEROID based on score matching for the setting where the inaccurate training data are unlabeled. Extensive experiments on MD datasets and downstream tasks validate the efficacy of ASTEROID. Our code and data are available at https://github.com/abukharin3/asteroid.