Blending Diverse Physical Priors with Neural Networks

Abstract

Machine learning in context of physical systems merits a re-examination of the learning strategy. In addition to data, one can leverage a vast library of physical prior models (e.g. kinematics, fluid flow, etc) to perform more robust inference. The nascent sub-field of \emph{physics-based learning} (PBL) studies the blending of neural networks with physical priors. While previous PBL algorithms have been applied successfully to specific tasks, it is hard to generalize existing PBL methods to a wide range of physics-based problems. Such generalization would require an architecture that can adapt to variations in the correctness of the physics, or in the quality of training data. No such architecture exists. In this paper, we aim to generalize PBL, by making a first attempt to bring neural architecture search (NAS) to the realm of PBL. We introduce a new method known as physics-based neural architecture search (PhysicsNAS) that is a top-performer across a diverse range of quality in the physical model and the dataset.

References in corpus (15)

• Adam: A Method for Stochastic Optimization
• Neural Architecture Search with Reinforcement Learning
• Deep Convolutional Neural Network for Inverse Problems in Imaging
• First M87 Event Horizon Telescope Results. IV. Imaging the Central Supermassive Black Hole
• Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations
• Efficient Neural Architecture Search via Parameter Sharing
• Neural Lander: Stable Drone Landing Control using Learned Dynamics
• DARTS+: Improved Differentiable Architecture Search with Early Stopping
• Low Photon Count Phase Retrieval Using Deep Learning
• PDE-Net: Learning PDEs from Data
• Unrolled Optimization with Deep Priors
• Learning Visual Predictive Models of Physics for Playing Billiards
• Physics-Based Generative Adversarial Models for Image Restoration and Beyond
• Inverse Transport Networks
• Deep Shape from Polarization

Cited by in corpus (5)

• A Review of Machine Learning Methods Applied to Structural Dynamics and Vibroacoustic
• PID-GAN: A GAN Framework based on a Physics-informed Discriminator for Uncertainty Quantification with Physics
• Physics-Integrated Variational Autoencoders for Robust and Interpretable Generative Modeling
• Kinematically consistent recurrent neural networks for learning inverse problems in wave propagation
• Domain-specific loss design for unsupervised physical training: A new approach to modeling medical ML solutions