| 本文已被：浏览 905次 下载 1959次 |
码上扫一扫！ |
|
|
| 基于层重组扩展卡尔曼滤波的神经网络力场训练 |
|
胡思宇1,2, 周远昌1,2, 赵曈1,2, 汪林望3, 贾伟乐1,2, 谭光明1,2
|
|
1.处理器全国重点实验室 (中国科学院 计算技术研究所), 北京 100190;2.中国科学院大学, 北京 100049;3.中国科学院 半导体研究所, 北京 100190
|
|
| 摘要: |
| 分子动力学模拟在材料模拟、生物制药等领域发挥着重要作用. 近年来, 科学智能(AI-for-Science)发展, 尤其是神经网络力场在预测能量、力等性质的问题上, 相比于传统势函数方法在准确性上有大幅提升. 针对当前的神经网络力场模型在使用一阶训练方法时出现的超参设置敏感和梯度爆炸问题, 给出层重组卡尔曼滤波器在避免超参数设置问题上的若干策略和防止梯度爆炸的理论证明. 基于层重组卡尔曼滤波器, 制定交替训练方法并分析该方法的精度收益和时间成本、提出分块阈值的性能模型并论述该模型的有效性、证明防止梯度爆炸的性质并验证该优化器关于激活函数和权重初始化的鲁棒性. 对4种典型的神经网络力场模型在11个有代表性的数据集进行测试, 实验表明, 当层重组卡尔曼滤波器和一阶优化器达到相当的预测精度时, 层重组卡尔曼滤波器相比于一阶方法快8–10倍. 可以相信, 所提出的层重组卡尔曼滤波训练方法能给其他的科学智能应用带来启发. |
| 关键词: 科学智能 神经网络 力场训练 层重组扩展卡尔曼滤波优化器 分子动力学模拟 |
| DOI：10.13328/j.cnki.jos.007258 |
| 分类号:TP18 |
| 基金项目:国家自然科学基金(T2125013, 92270206, 62372435, 62032023, 61972377, 61972380, T2293700, T2293702); 中国科学院战略性先导科技专项(XDB0500102); 中国科学院稳定支持青年科学家团队(YSBR-005) |
|
| Neural Network Force Field Training Based on Reorganized Layer-wised Extended Kalman Filtering |
|
HU Si-Yu1,2, ZHOU Yuan-Chang1,2, ZHAO Tong1,2, WANG Lin-Wang3, JIA Wei-Le1,2, TAN Guang-Ming1,2
|
|
1.State Key Lab of Processors (Institute of Computing Technology, Chinese Academy of Sciences), Beijing 100190, China;2.University of Chinese Academy of Sciences, Beijing 100049, China;3.Institute of Semiconductors, Chinese Academy of Sciences, Beijing 100190, China
|
| Abstract: |
| Molecular dynamics simulation plays an important role in material simulation, biopharmaceuticals, and other areas. In recent years, the development of AI-for-Science has greatly improved the accuracy of neural network force fields in predicting energy, force, and other properties, compared to traditional methods using potential functions. Neural network force field models may challenges such as hyperparameter settings and gradient explosion when trained by the first-order method. Based on an optimizer named reorganized layer extended Kalman filtering, this study provides several strategies to avoid hyperparameters and offers theoretical evidence for preventing gradient explosion. This study also proposes an alternate training method and analyzes its accuracy gains and time costs. A performance model of block thresholding is proposed, and its effectiveness is explored. Additionally, the property of preventing gradient explosion is proven, and the optimizer’s robustness with respect to activation functions and weight initialization is validated. Four typical neural network force field models are tested on 11 representative datasets. Results show that when the proposed optimizer and the first-order optimizer achieve comparable prediction accuracy, the proposed optimizer is 8 to 10 times faster than the first-order optimizer. It is believed that the proposed training method can inspire other AI-for-Science applications. |
| Key words: AI-for-Science neural network force field training reorganized layer-wised extended Kalman filtering optimizer molecular dynamics simulation |
