Variational Loss

Variational loss functions are central to many machine learning methods, particularly those solving partial differential equations (PDEs) or performing dimensionality reduction. Current research focuses on improving the efficiency and accuracy of these methods, including advancements in architectures like Physics-Informed Neural Networks (PINNs) and the development of adaptive sampling techniques to optimize training data usage. These improvements are crucial for tackling complex scientific and engineering problems, enabling faster and more accurate solutions for applications ranging from inverse problems to image processing and network analysis.

Papers

April 18, 2024

FastVPINNs: Tensor-Driven Acceleration of VPINNs for Complex Geometries
Thivin Anandh, Divij Ghose, Himanshu Jain, Sashikumaar Ganesan
Variational Method Tensor Data Reg PINNs Complex Geometry Variational Loss Sparse Tensor Accelerator

November 9, 2023

Dual Pipeline Style Transfer with Input Distribution Differentiation
ShiQi Jiang, JunJie Kang, YuJian Li
Input Distribution Texture Representation Color Transfer Stage Pipeline Variational Loss

October 26, 2023

Adaptive importance sampling for Deep Ritz
Xiaoliang Wan, Tao Zhou, Yuancheng Zhou
Adaptive Importance Adaptive Sampling Adaptive Method Deep Ritz Method Variational Loss

October 5, 2023

Deep Variational Multivariate Information Bottleneck -- A Framework for Variational Losses
Eslam Abdelaleem, Ilya Nemenman, K. Michael Martini
New Framework Variational Auto Variational Information Bottleneck Variational Representation Variational Loss

August 29, 2023

Class Prior-Free Positive-Unlabeled Learning with Taylor Variational Loss for Hyperspectral Remote Sensing Imagery
Hengwei Zhao, Xinyu Wang, Jingtao Li, Yanfei Zhong
LeArning Abstract Natural Gradient Hyperspectral Image Classification Pu Learning Class Prior Hyperspectral Remote Sensing Variational Loss

January 12, 2023

Variational Inference: Posterior Threshold Improves Network Clustering Accuracy in Sparse Regimes
Xuezhen Li, Can M. Le
Variational Inference Dense Network Variational Loss Sparse Target Network Clustering

November 23, 2022

Replacing Automatic Differentiation by Sobolev Cubatures fastens Physics Informed Neural Nets and strengthens their Approximation Power
Juan Esteban Suarez Cardona, Michael Hecht
Physic Informed Neural Network Automatic Differentiation Approximation Power Inverse PDE Problem Loss Estimation Variational Loss

October 28, 2022

Evaluating the Impact of Loss Function Variation in Deep Learning for Classification
Simon Dräger, Jannik Dunkelau
Deep Learning Deep Neural Network Classification Code Loss Function Global Impact Variational Loss Novel Stealth Loss

December 10, 2021

How to Avoid Trivial Solutions in Physics-Informed Neural Networks
Raphael Leiteritz, Dirk Pflüger
Physic Informed Neural Network Scientific Machine Learning Global Solution Variational Loss Data Driven Modelling

Variational Loss

Papers

FastVPINNs: Tensor-Driven Acceleration of VPINNs for Complex Geometries

Dual Pipeline Style Transfer with Input Distribution Differentiation

Adaptive importance sampling for Deep Ritz

Deep Variational Multivariate Information Bottleneck -- A Framework for Variational Losses

Class Prior-Free Positive-Unlabeled Learning with Taylor Variational Loss for Hyperspectral Remote Sensing Imagery

Variational Inference: Posterior Threshold Improves Network Clustering Accuracy in Sparse Regimes

Replacing Automatic Differentiation by Sobolev Cubatures fastens Physics Informed Neural Nets and strengthens their Approximation Power

Evaluating the Impact of Loss Function Variation in Deep Learning for Classification

How to Avoid Trivial Solutions in Physics-Informed Neural Networks