• Corpus ID: 203610380

Blending Diverse Physical Priors with Neural Networks

@article{Ba2019BlendingDP,
title={Blending Diverse Physical Priors with Neural Networks},
author={Yunhao Ba and Guangyuan Zhao and Achuta Kadambi},
journal={ArXiv},
year={2019},
volume={abs/1910.00201}
}
• Published 25 September 2019
• Computer Science, Physics
• ArXiv
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…
18 Citations

Figures and Tables from this paper

• Computer Science
NeurIPS
• 2021
This work introduces an architecture of variational autoencoders (VAEs) in which a part of the latent space is grounded by physics, and proposes a regularized learning method that controls the effect of the trainable components and preserves the semantics of the physics-based latent variables as intended.
• Computer Science
KDD
• 2021
This work proposes a novel physics-informed GAN architecture, termed PID-GAN, where the knowledge of physics is used to inform the learning of both the generator and discriminator models, making ample use of unlabeled data instances.
• Computer Science
ArXiv
• 2021
This work attempts to perform physically interpretable learning of inverse problems in wave propagation without suffering overfitting restrictions using long short-term memory networks endowed with a physical, hyperparameter-driven regularizer, analogous to an artificial bulk modulus.
• Computer Science
IEEE Access
• 2020
A meticulous and systematic attempt at organizing and standardizing the methods of combining ML and MB models as hybrid learning methods and shedding some light on the challenges of hybrid models.
• Computer Science
ArXiv
• 2020
An overview of techniques to integrate machine learning with physics-based modeling and classes of methodologies used to construct physics-guided machine learning models and hybrid physics-machine learning frameworks from a machine learning standpoint is provided.
• Computer Science
Communications of the American Mathematical Society
• 2022
A unifying framework for blending mechanistic and machine-learning approaches to identify dynamical systems from noisily and partially observed data is presented, and it is demonstrated numerically how data assimilation can be leveraged to learn hidden dynamics from noisy, partially-observed data.
• Computer Science
ACM Comput. Surv.
• 2023
A taxonomy of existing techniques that are able to integrate traditional physics-based modeling approaches with state-of-the-art machine learning (ML) techniques is provided, which uncovers knowledge gaps and potential crossovers of methods between disciplines.
• Computer Science
2022 IEEE International Conference on Computational Photography (ICCP)
• 2022
This paper presents a framework to understand the building blocks of this nascent field of end-to-end design of camera hardware and algorithms, and shows how methods that exploit both physics and data have become prevalent in imaging and computer vision.
The proposed approach, termed as Physics-informed Discriminator for cGAN (cGANPID), is more aligned to the adversarial learning idea of cGAN as opposed to existing methods on incorporating physical knowledge in GANs by adding physics based loss functions as additional terms in the optimization objective of GAN.
This paper describes a class of hierarchical models parameterized by neural networks: neural hierarchical models, providing a bridge between deep learning and ecological modeling that combines the function representation power of neural networks with the inferential capacity of hierarchical Models.

References

SHOWING 1-10 OF 53 REFERENCES

• Physics
ArXiv
• 2019
This work study the blending of physics and deep learning in the context of Shape from Polarization (SfP) in the framework of a two-stage encoder, finding that there is a subtlety to combining physics and neural networks.
• Computer Science
NeurIPS
• 2018
This paper demonstrates how to perform backpropagation analytically through a physical simulator defined via a linear complementarity problem, and highlights the system's ability to learn physical parameters from data, efficiently match and simulate observed visual behavior, and readily enable control via gradient-based planning methods.
• M. Raissi
• Computer Science
J. Mach. Learn. Res.
• 2018
This work puts forth a deep learning approach for discovering nonlinear partial differential equations from scattered and potentially noisy observations in space and time by approximate the unknown solution as well as the nonlinear dynamics by two deep neural networks.
• Computer Science, Mathematics
ArXiv
• 2017
This paper presents unrolled optimization with deep priors, a principled framework for infusing knowledge of the image formation into deep networks that solve inverse problems in imaging, inspired by classical iterative methods.
• Computer Science
ICML
• 2018
Numerical experiments show that the PDE-Net has the potential to uncover the hidden PDE of the observed dynamics, and predict the dynamical behavior for a relatively long time, even in a noisy environment.
• Computer Science
ArXiv
• 2017
A novel framework, termed as physics-guided neural network (PGNN), leverages the output of physics-based model simulations along with observational features to generate predictions using a neural network architecture to ensure better generalizability as well as physical consistency of results.
• Computer Science
ICLR
• 2017
This paper uses a recurrent network to generate the model descriptions of neural networks and trains this RNN with reinforcement learning to maximize the expected accuracy of the generated architectures on a validation set.
• Computer Science
ArXiv
• 2017
We introduce physics informed neural networks -- neural networks that are trained to solve supervised learning tasks while respecting any given law of physics described by general nonlinear partial
• Computer Science
ICLR
• 2017
MetaQNN is introduced, a meta-modeling algorithm based on reinforcement learning to automatically generate high-performing CNN architectures for a given learning task that beat existing networks designed with the same layer types and are competitive against the state-of-the-art methods that use more complex layer types.
• Computer Science
ICLR
• 2016
This paper explores how an agent can be equipped with an internal model of the dynamics of the external world, and how it can use this model to plan novel actions by running multiple internal simulations ("visual imagination").