Visual Reinforcement Learning

Visual reinforcement learning (VRL) aims to train agents to make decisions and control actions directly from visual inputs, a challenging problem due to the high dimensionality and complexity of visual data. Current research heavily focuses on improving sample efficiency and generalization capabilities, employing techniques like contrastive learning, knowledge distillation, and data augmentation within various model architectures including convolutional neural networks and transformers. These advancements are crucial for enabling robots and autonomous systems to learn complex tasks from visual experience, impacting fields like robotics, autonomous driving, and video game AI.

Papers

October 10, 2022

A Comprehensive Survey of Data Augmentation in Visual Reinforcement Learning
Guozheng Ma, Zhen Wang, Zhecheng Yuan, Xueqian Wang, Bo Yuan, Dacheng Tao
Data Augmentation Comprehensive Survey Augmentation Technique Visual Reinforcement Learning

September 16, 2022

Look where you look! Saliency-guided Q-networks for generalization in visual Reinforcement Learning
David Bertoin, Adil Zouitine, Mehdi Zouitine, Emmanuel Rachelson
Strong Generalization Visual Reinforcement Learning Q Network

June 6, 2022

Real2Sim or Sim2Real: Robotics Visual Insertion using Deep Reinforcement Learning and Real2Sim Policy Adaptation
Yiwen Chen, Xue Li, Sheng Guo, Xian Yao Ng, Marcelo Ang
Reinforcement Learning Deep Reinforcement Learning Visual Reinforcement Learning Insertion Task Sim2real Strategy

May 20, 2022

Learning Task-relevant Representations for Generalization via Characteristic Functions of Reward Sequence Distributions
Rui Yang, Jie Wang, Zijie Geng, Mingxuan Ye, Shuiwang Ji, Bin Li, Feng Wu
Strong Generalization Visual Reinforcement Learning DeepMind Control Suite Reward Prediction Characteristic Function Reward Distribution Task Induced Representation

May 12, 2022

Economical Precise Manipulation and Auto Eye-Hand Coordination with Binocular Visual Reinforcement Learning
Yiwen Chen, Sheng Guo, Zedong Zhang, Lei Zhou, Xian Yao Ng, Marcelo H. Ang
Manipulation Task Real Robot Visual Reinforcement Learning Precise Manipulation Eye in Hand Eye Hand Coordination

February 28, 2022

Weakly Supervised Disentangled Representation for Goal-conditioned Reinforcement Learning
Zhifeng Qian, Mingyu You, Hongjun Zhou, Bin He
Goal Conditioned Reinforcement Learning Visual Reinforcement Learning Goal Conditioned Policy Generalization

February 21, 2022

Don't Touch What Matters: Task-Aware Lipschitz Data Augmentation for Visual Reinforcement Learning
Zhecheng Yuan, Guozheng Ma, Yao Mu, Bo Xia, Bo Yuan, Xueqian Wang, Ping Luo, Huazhe Xu
Data Augmentation Proper Issue Labeling Visual Reinforcement Learning DeepMind Control Suite Control Benchmark

Visual Reinforcement Learning

Papers

A Comprehensive Survey of Data Augmentation in Visual Reinforcement Learning

Look where you look! Saliency-guided Q-networks for generalization in visual Reinforcement Learning

Real2Sim or Sim2Real: Robotics Visual Insertion using Deep Reinforcement Learning and Real2Sim Policy Adaptation

Learning Task-relevant Representations for Generalization via Characteristic Functions of Reward Sequence Distributions

Economical Precise Manipulation and Auto Eye-Hand Coordination with Binocular Visual Reinforcement Learning

Weakly Supervised Disentangled Representation for Goal-conditioned Reinforcement Learning

Don't Touch What Matters: Task-Aware Lipschitz Data Augmentation for Visual Reinforcement Learning