Gradient Penalty

Gradient penalty methods are used to improve the stability and performance of various machine learning models, particularly generative adversarial networks (GANs) and reinforcement learning algorithms. Current research focuses on applying gradient penalties to enhance model robustness, address issues like mode collapse and hallucination in image generation and language models, and improve the efficiency of learning in challenging scenarios such as offline reinforcement learning and inverse problems. These techniques are proving valuable across diverse applications, including image denoising, survival analysis, and robotics, by promoting smoother, more generalizable, and less prone to overfitting models.

Papers

September 30, 2024

HELPD: Mitigating Hallucination of LVLMs by Hierarchical Feedback Learning with Vision-enhanced Penalty Decoding
Fan Yuan, Chi Qin, Xiaogang Xu, Piji Li
Large Vision Language Model Content Hallucination Hierarchical Learning Constrained Decoding Multimodal Hallucination Hallucination Benchmark Hierarchical Policy Learning Gradient Penalty

July 16, 2024

Novel Hybrid Integrated Pix2Pix and WGAN Model with Gradient Penalty for Binary Images Denoising
Luca Tirel, Ali Mohamed Ali, Hashim A. Hashim
Generative Adversarial Network Image Denoising GAN Training Wasserstein Gan Gradient Penalty

July 3, 2024

Efficient Imitation Without Demonstrations via Value-Penalized Auxiliary Control from Examples
Trevor Ablett, Bryan Chan, Jayce Haoran Wang, Jonathan Kelly
Reinforcement Learning LeArning Abstract Reward Function Noisy Demonstration Reward Report Fewer Example Auxiliary Loss Auxiliary Feature Gradient Penalty

June 8, 2023

Solution of physics-based inverse problems using conditional generative adversarial networks with full gradient penalty
Deep Ray, Javier Murgoitio-Esandi, Agnimitra Dasgupta, Assad A. Oberai
Generative Adversarial Network Inverse Problem Solution Path Conditional Generative Adversarial Network Gradient Penalty Physic Based Inverse Problem

October 19, 2022

Robust Offline Reinforcement Learning with Gradient Penalty and Constraint Relaxation
Chengqian Gao, Ke Xu, Liu Liu, Deheng Ye, Peilin Zhao, Zhiqiang Xu
Offline Reinforcement Learning Policy Constraint Robust Offline Reinforcement Learning Constraint Relaxation Gradient Penalty

October 1, 2022

Learning Globally Smooth Functions on Manifolds
Juan Cervino, Luiz F. O. Chamon, Benjamin D. Haeffele, Rene Vidal, Alejandro Ribeiro
Gradient Based Class Manifold Constrained Learning Manifold Regularization Gradient Penalty

May 27, 2022

Hazard Gradient Penalty for Survival Analysis
Seungjae Jung, Kyung-Min Kim
Survival Model Survival Analysis Gradient Penalty

Gradient Penalty

Papers

HELPD: Mitigating Hallucination of LVLMs by Hierarchical Feedback Learning with Vision-enhanced Penalty Decoding

Novel Hybrid Integrated Pix2Pix and WGAN Model with Gradient Penalty for Binary Images Denoising

Efficient Imitation Without Demonstrations via Value-Penalized Auxiliary Control from Examples

Solution of physics-based inverse problems using conditional generative adversarial networks with full gradient penalty

Robust Offline Reinforcement Learning with Gradient Penalty and Constraint Relaxation

Learning Globally Smooth Functions on Manifolds

Hazard Gradient Penalty for Survival Analysis