A Generalizable Physics-informed Learning Framework for Risk Probability Estimation

被引:0
|
作者
Wang, Zhuoyuan [1 ]
Nakahira, Yorie [1 ]
机构
[1] Carnegie Mellon Univ, Elect & Comp Engn Dept, Pittsburgh, PA 15213 USA
来源
LEARNING FOR DYNAMICS AND CONTROL CONFERENCE, VOL 211 | 2023年 / 211卷
关键词
Stochastic safe control; physics-informed learning; risk probability estimation; SAFE EXPLORATION; REACHABILITY;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Accurate estimates of long-term risk probabilities and their gradients are critical for many stochastic safe control methods. However, computing such risk probabilities in real-time and in unseen or changing environments is challenging. Monte Carlo (MC) methods cannot accurately evaluate the probabilities and their gradients as an infinitesimal devisor can amplify the sampling noise. In this paper, we develop an efficient method to evaluate the probabilities of long-term risk and their gradients. The proposed method exploits the fact that long-term risk probability satisfies certain partial differential equations (PDEs), which characterize the neighboring relations between the probabilities, to integrate MC methods and physics-informed neural networks. We provide theoretical guarantees of the estimation error given certain choices of training configurations. Numerical results show the proposed method has better sample efficiency, generalizes well to unseen regions, and can adapt to systems with changing parameters. The proposed method can also accurately estimate the gradients of risk probabilities, which enables first- and second-order techniques on risk probabilities to be used for learning and control.
引用
收藏
页数:13
相关论文
共 39 条
  • [21] Extended Physics-Informed Neural Networks (XPINNs): A Generalized Space-Time Domain Decomposition Based Deep Learning Framework for Nonlinear Partial Differential Equations
    Jagtap, Ameya D.
    Karniadakis, George Em
    COMMUNICATIONS IN COMPUTATIONAL PHYSICS, 2020, 28 (05) : 2002 - 2041
  • [22] GENERATIVE ENSEMBLE REGRESSION: LEARNING PARTICLE DYNAMICS FROM OBSERVATIONS OF ENSEMBLES WITH PHYSICS-INFORMED DEEP GENERATIVE MODELS
    Yang, Liu
    Daskalakis, Constantinos
    Karniadakis, George E.
    SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2022, 44 (01): : B80 - B99
  • [23] Physics-informed Graph Neural Network for Dynamic Reconfiguration of power systems
    Authier, Jules
    Haider, Rabab
    Annaswamy, Anuradha
    Dorfler, Florian
    ELECTRIC POWER SYSTEMS RESEARCH, 2024, 235
  • [24] PhySR: Physics-informed deep super-resolution for spatiotemporal data
    Ren, Pu
    Rao, Chengping
    Liu, Yang
    Ma, Zihan
    Wang, Qi
    Wang, Jian-Xun
    Sun, Hao
    JOURNAL OF COMPUTATIONAL PHYSICS, 2023, 492
  • [25] Physics-informed variational inference for uncertainty quantification of stochastic differential equations
    Shin, Hyomin
    Choi, Minseok
    JOURNAL OF COMPUTATIONAL PHYSICS, 2023, 487
  • [26] Design of a Novel Explainable Adversarial Autoencoder Model for the Electromagnetic Analysis of Functional Materials Based on Physics-Informed Learning
    Narang, Naina
    Lingam, Greeshma
    IEEE ACCESS, 2024, 12 : 166044 - 166057
  • [27] Physics-Informed Neural Networks for Tissue Elasticity Reconstruction in Magnetic Resonance Elastography
    Ragoza, Matthew
    Batmanghelich, Kayhan
    MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION, MICCAI 2023, PT X, 2023, 14229 : 333 - 343
  • [28] hp-VPINNs: Variational physics-informed neural networks with domain decomposition
    Kharazmi, Ehsan
    Zhang, Zhongqiang
    Karniadakis, George E. M.
    COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING, 2021, 374
  • [29] Deep fuzzy physics-informed neural networks for forward and inverse PDE problems
    Wu, Wenyuan
    Duan, Siyuan
    Sun, Yuan
    Yu, Yang
    Liu, Dong
    Peng, Dezhong
    NEURAL NETWORKS, 2025, 181
  • [30] Physics-informed learning for thermophysical field reconstruction and parameter measurement in a nano-porous insulator's heat transfer problem
    Pang, Hao-Qiang
    Shao, Xia
    Zhang, Zi-Tong
    Xie, Xin
    Dai, Ming-Yang
    Guo, Jiang-Feng
    Zhang, Yan-Bo
    Liu, Tian-Yuan
    Gao, Yan-Feng
    INTERNATIONAL COMMUNICATIONS IN HEAT AND MASS TRANSFER, 2023, 148