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

November 15, 2024

The Surprising Ineffectiveness of Pre-Trained Visual Representations for Model-Based Reinforcement Learning
Moritz Schneider, Robert Krug, Narunas Vaskevicius, Luigi Palmieri, Joschka Boedecker
Reinforcement Learning Model Free Reinforcement Learning Pre Trained Representation Visual Reinforcement Learning Surprising Effectiveness

October 19, 2024

MENTOR: Mixture-of-Experts Network with Task-Oriented Perturbation for Visual Reinforcement Learning
Suning Huang, Zheyu Zhang, Tianhai Liang, Yihan Xu, Zhehao Kou, Chenhao Lu, Guowei Xu, Zhengrong Xue, Huazhe Xu
Real World Mixture of Expert Model Free Reinforcement Learning Visual Reinforcement Learning DeepMind Control Suite

October 17, 2024

Sliding Puzzles Gym: A Scalable Benchmark for State Representation in Visual Reinforcement Learning
Bryan L. M. de Oliveira, Murilo L. da Luz, Bruno Brandão, Luana G. B. Martins, Telma W. de L. Soares, Luckeciano C. Melo
Reinforcement Learning Representation Learning Large Scale Visual Representation State Representation Visual Reinforcement Learning Puzzle Game

September 28, 2024

Generalizing Consistency Policy to Visual RL with Prioritized Proximal Experience Regularization
Haoran Li, Zhennan Jiang, Yuhui Chen, Dongbin Zhao
Visual Reinforcement Learning Dimensional State Space Consistency Learning Actor Critic Framework

September 26, 2024

Learning Occlusion-aware Decision-making from Agent Interaction via Active Perception
Jie Jia, Yiming Shu, Zhongxue Gan, Wenchao Ding
Reinforcement Learning LeArning Abstract Visual Reinforcement Learning Active Perception Agent Interaction Occlusion Sensitivity Analysis

September 2, 2024

MOOSS: Mask-Enhanced Temporal Contrastive Learning for Smooth State Evolution in Visual Reinforcement Learning
Jiarui Sun, M. Ugur Akcal, Wei Zhang, Girish Chowdhary
Contrastive Learning Policy Learning Visual Reinforcement Learning

August 13, 2024

Integrating Saliency Ranking and Reinforcement Learning for Enhanced Object Detection
Matthias Bartolo, Dylan Seychell, Josef Bajada
Reinforcement Learning Object Detection Human Saliency Visual Reinforcement Learning Saliency Ranking

August 11, 2024

CURLing the Dream: Contrastive Representations for World Modeling in Reinforcement Learning
Victor Augusto Kich, Jair Augusto Bottega, Raul Steinmetz, Ricardo Bedin Grando, Ayano Yorozu, Akihisa Ohya
Reinforcement Learning Contrastive Learning Visual Reinforcement Learning Novel Reinforcement Learning Contrastive Representation DeepMind Control Suite Dream Report World Modeling

July 24, 2024

Pretrained Visual Representations in Reinforcement Learning
Emlyn Williams, Athanasios Polydoros
Reinforcement Learning Pre Trained Representation Visual Reinforcement Learning

July 22, 2024

Learning to Manipulate Anywhere: A Visual Generalizable Framework For Reinforcement Learning
Zhecheng Yuan, Tianming Wei, Shuiqi Cheng, Gu Zhang, Yuanpei Chen, Huazhe Xu
Reinforcement Learning LeArning Abstract Visual Reinforcement Learning Visuo Motor Control Model Generalizability Visual Generalization

June 30, 2024

DEAR: Disentangled Environment and Agent Representations for Reinforcement Learning without Reconstruction
Ameya Pore, Riccardo Muradore, Diego Dall'Alba
Reinforcement Learning Full State Reconstruction Environment Feature Disentangled Representation Robotic Task Visual Reinforcement Learning Disentangled Haze Agent Representation

June 10, 2024

Sim-To-Real Transfer for Visual Reinforcement Learning of Deformable Object Manipulation for Robot-Assisted Surgery
Paul Maria Scheikl, Eleonora Tagliabue, Balázs Gyenes, Martin Wagner, Diego Dall'Alba, Paolo Fiorini, Franziska Mathis-Ullrich
Formality Transfer Visual Reinforcement Learning Robot Assisted Surgery Deformable Object Manipulation Visuomotor Policy Sim to Real Cognitive Robotics Real Robotic System

May 27, 2024

A Recipe for Unbounded Data Augmentation in Visual Reinforcement Learning
Abdulaziz Almuzairee, Nicklas Hansen, Henrik I. Christensen
Data Augmentation Strong Generalization Q Learning Complete Recipe Visual Reinforcement Learning Visual Generalization

May 20, 2024

Learning Future Representation with Synthetic Observations for Sample-efficient Reinforcement Learning
Xin Liu, Yaran Chen, Dongbin Zhao
Synthetic Data Visual Reinforcement Learning Sample Efficient Reinforcement Learning Downstream Policy Learning Self Supervised Reinforcement Learning Future Latent Auxiliary Function

April 19, 2024

Zero-Shot Stitching in Reinforcement Learning using Relative Representations
Antonio Pio Ricciardi, Valentino Maiorca, Luca Moschella, Riccardo Marin, Emanuele Rodolà
Reinforcement Learning Visual Reinforcement Learning Relative Representation

March 16, 2024

ViSaRL: Visual Reinforcement Learning Guided by Human Saliency
Anthony Liang, Jesse Thomason, Erdem Bıyık
Reinforcement Learning Visual Representation Human Saliency Visual Reinforcement Learning Saliency Guided Training Robot

February 6, 2024

Read to Play (R2-Play): Decision Transformer with Multimodal Game Instruction
Yonggang Jin, Ge Zhang, Hao Zhao, Tianyu Zheng, Jarvi Guo, Liuyu Xiang, Shawn Yue, Stephen W. Huang, Zhaofeng He, Jie Fu
Reinforcement Learning Decision Transformer Read V Visual Reinforcement Learning Contextual Cue Task Guidance Multimodal Game

February 5, 2024

Understanding What Affects the Generalization Gap in Visual Reinforcement Learning: Theory and Empirical Evidence
Jiafei Lyu, Le Wan, Xiu Li, Zongqing Lu
Theoretical Understanding Generalization Gap Visual Reinforcement Learning Empirical Evidence Policy Generalization

January 24, 2024

TraKDis: A Transformer-based Knowledge Distillation Approach for Visual Reinforcement Learning with Application to Cloth Manipulation
Wei Chen, Nicolas Rojas
Reinforcement Learning Knowledge Distillation Application Proficiency Robot Perception Visual Reinforcement Learning Active Vision Cloth Manipulation Robotic Cloth Manipulation

January 17, 2024

An Efficient Generalizable Framework for Visuomotor Policies via Control-aware Augmentation and Privilege-guided Distillation
Yinuo Zhao, Kun Wu, Tianjiao Yi, Zhiyuan Xu, Xiaozhu Ju, Zhengping Che, Qinru Qiu, Chi Harold Liu, Jian Tang
Policy Learning Visual Reinforcement Learning Visuomotor Policy Model Generalizability Cut and Paste Privileged Knowledge Distillation

Visual Reinforcement Learning

Papers

The Surprising Ineffectiveness of Pre-Trained Visual Representations for Model-Based Reinforcement Learning

MENTOR: Mixture-of-Experts Network with Task-Oriented Perturbation for Visual Reinforcement Learning

Sliding Puzzles Gym: A Scalable Benchmark for State Representation in Visual Reinforcement Learning

Generalizing Consistency Policy to Visual RL with Prioritized Proximal Experience Regularization

Learning Occlusion-aware Decision-making from Agent Interaction via Active Perception

MOOSS: Mask-Enhanced Temporal Contrastive Learning for Smooth State Evolution in Visual Reinforcement Learning

Integrating Saliency Ranking and Reinforcement Learning for Enhanced Object Detection

CURLing the Dream: Contrastive Representations for World Modeling in Reinforcement Learning

Pretrained Visual Representations in Reinforcement Learning

Learning to Manipulate Anywhere: A Visual Generalizable Framework For Reinforcement Learning

DEAR: Disentangled Environment and Agent Representations for Reinforcement Learning without Reconstruction

Sim-To-Real Transfer for Visual Reinforcement Learning of Deformable Object Manipulation for Robot-Assisted Surgery

A Recipe for Unbounded Data Augmentation in Visual Reinforcement Learning

Learning Future Representation with Synthetic Observations for Sample-efficient Reinforcement Learning

Zero-Shot Stitching in Reinforcement Learning using Relative Representations

ViSaRL: Visual Reinforcement Learning Guided by Human Saliency

Read to Play (R2-Play): Decision Transformer with Multimodal Game Instruction

Understanding What Affects the Generalization Gap in Visual Reinforcement Learning: Theory and Empirical Evidence

TraKDis: A Transformer-based Knowledge Distillation Approach for Visual Reinforcement Learning with Application to Cloth Manipulation

An Efficient Generalizable Framework for Visuomotor Policies via Control-aware Augmentation and Privilege-guided Distillation