Physics-Integrated Variational Autoencoders for Robust and Interpretable Generative Modeling

Abstract

Integrating physics models within machine learning models holds considerable promise toward learning robust models with improved interpretability and abilities to extrapolate. In this work, we focus on the integration of incomplete physics models into deep generative models. In particular, we introduce an architecture of variational autoencoders (VAEs) in which a part of the latent space is grounded by physics. A key technical challenge is to strike a balance between the incomplete physics and trainable components such as neural networks for ensuring that the physics part is used in a meaningful manner. To this end, we propose a regularized learning method that controls the effect of the trainable components and preserves the semantics of the physics-based latent variables as intended. We not only demonstrate generative performance improvements over a set of synthetic and real-world datasets, but we also show that we learn robust models that can consistently extrapolate beyond the training distribution in a meaningful manner. Moreover, we show that we can control the generative process in an interpretable manner.

References in corpus (26)

• B-PINNs: Bayesian Physics-Informed Neural Networks for Forward and Inverse PDE Problems with Noisy Data
• Deep Learning for Physical Processes: Incorporating Prior Scientific Knowledge
• Kohn-Sham equations as regularizer: building prior knowledge into machine-learned physics
• Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers
• Hybrid FEM-NN models: Combining artificial neural networks with the finite element method
• Augmenting Physical Models with Deep Networks for Complex Dynamics Forecasting
• Solving inverse-PDE problems with physics-aware neural networks
• Lagrangian Neural Networks
• Embedding Hard Physical Constraints in Neural Network Coarse-Graining of 3D Turbulence
• Grey-box models for wave loading prediction
• Blending Diverse Physical Priors with Neural Networks
• Explainable Machine Learning with Prior Knowledge: An Overview
• SimGANs: Simulator-Based Generative Adversarial Networks for ECG Synthesis to Improve Deep ECG Classification
• A probabilistic generative model for semi-supervised training of coarse-grained surrogates and enforcing physical constraints through virtual observables
• Integrating Expert ODEs into Neural ODEs: Pharmacology and Disease Progression
• SimGAN: Hybrid Simulator Identification for Domain Adaptation via Adversarial Reinforcement Learning
• AdjointNet: Constraining machine learning models with physics-based codes
• Neural Dynamical Systems: Balancing Structure and Flexibility in Physical Prediction
• Hybrid Physical-Deep Learning Model for Astronomical Inverse Problems
• Thermodynamic Consistent Neural Networks for Learning Material Interfacial Mechanics
• Living in the Physics and Machine Learning Interplay for Earth Observation
• Variational Autoencoding of PDE Inverse Problems
• Interpretable machine learning models: a physics-based view
• Physics-aware, probabilistic model order reduction with guaranteed stability
• Amortized Synthesis of Constrained Configurations Using a Differentiable Surrogate
• Embedded-model flows: Combining the inductive biases of model-free deep learning and explicit probabilistic modeling