mlsanityData to 14 Oct 2026

DeGuV: Depth-Guided Visual Reinforcement Learning for Generalization and Interpretability in Manipulation

arXiv 2509.04970 ↗First posted 5 Sep 2025cs.RO cs.AIcited by 0

On the map: Deep reinforcement learning methods → RL for robot locomotion and manipulation

Abstract (arXiv)

Reinforcement learning (RL) agents can learn to solve complex tasks from visual inputs, but generalizing these learned skills to new environments remains a major challenge in RL application, especially robotics. While data augmentation can improve generalization, it often compromises sample efficiency and training stability. This paper introduces DeGuV, an RL framework that enhances both generalization and sample efficiency. In specific, we leverage a learnable masker network that produces a mask from the depth input, preserving only critical visual information while discarding irrelevant pixels. Through this, we ensure that our RL agents focus on essential features, improving robustness under data augmentation. In addition, we incorporate contrastive learning and stabilize Q-value estimation under augmentation to further enhance sample efficiency and training stability. We evaluate our proposed method on the RL-ViGen benchmark using the Franka Emika robot and demonstrate its effectiveness in zero-shot sim-to-real transfer. Our results show that DeGuV outperforms state-of-the-art methods in both generalization and sample efficiency while also improving interpretability by highlighting the most relevant regions in the visual input

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in RL for robot locomotion and manipulation, one of 15 directions in Deep reinforcement learning methods. Builds on 36 of its 40 references that are in the map. First posted on arXiv 5 Sep 2025.

Builds on

36 of 40 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025Dream to Generalize: Zero-Shot Model-Based Reinforcement Learning for Unseen Visual Distractions5 Jun
  2. 2024Learning to Manipulate Anywhere: A Visual Generalizable Framework For Reinforcement Learning22 Jul
  3. 2024ViSaRL: Visual Reinforcement Learning Guided by Human Saliency16 Mar

Other references · 33

  1. 2018Representation Learning with Contrastive Predictive Codingcited by 7,777
  2. 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,885
  3. 2020CURL: Contrastive Unsupervised Representations for Reinforcement Learningcited by 760
  4. 2019Solving Rubik's Cube with a Robot Handcited by 734
  5. 2020Simple Copy-Paste is a Strong Data Augmentation Method for Instance Segmentationcited by 596
  6. 2020Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixelscited by 585
  7. 2020robosuite: A Modular Simulation Framework and Benchmark for Robot Learningcited by 511
  8. 2020Reinforcement Learning with Augmented Datacited by 468
  9. 2021Mastering Visual Continuous Control: Improved Data-Augmented Reinforcement Learningcited by 372
  10. 2016Target-driven Visual Navigation in Indoor Scenes using Deep Reinforcement Learningcited by 366
  11. 2020Agent57: Outperforming the Atari Human Benchmarkcited by 303
  12. 2016Learning to Navigate in Complex Environmentscited by 198
21 more
  1. 2021End-to-End Urban Driving by Imitating a Reinforcement Learning Coachcited by 194
  2. 2018A Study on Overfitting in Deep Reinforcement Learningcited by 174
  3. 2020Generalization in Reinforcement Learning by Soft Data Augmentationcited by 137
  4. 2023Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learningcited by 133
  5. 2021Contrastive Behavioral Similarity Embeddings for Generalization in Reinforcement Learningcited by 123
  6. 2021Stabilizing Deep Q-Learning with ConvNets and Vision Transformers under Data Augmentationcited by 105
  7. 2021Q-attention: Enabling Efficient Learning for Vision-based Robotic Manipulationcited by 85
  8. 2023Decomposing the Generalization Gap in Imitation Learning for Visual Robotic Manipulationcited by 82
  9. 2022Pre-Trained Image Encoder for Generalizable Visual Reinforcement Learningcited by 60
  10. 2022Look Closer: Bridging Egocentric and Third-Person Views with Transformers for Robotic Manipulationcited by 49
  11. 2022Mask-based Latent Reconstruction for Reinforcement Learningcited by 39
  12. 2021Domain Adaptation In Reinforcement Learning Via Latent Unified State Representationcited by 35
  13. 2022Look where you look! Saliency-guided Q-networks for generalization in visual Reinforcement Learningcited by 34
  14. 2020Robotic Grasping using Deep Reinforcement Learningcited by 28
  15. 2022Human-level Atari 200x fastercited by 28
  16. 2023RL-ViGen: A Reinforcement Learning Benchmark for Visual Generalizationcited by 28
  17. 2021Fast and Data-Efficient Training of Rainbow: an Experimental Study on Ataricited by 10
  18. 2023Ignorance is Bliss: Robust Control via Information Gatingcited by 10
  19. 2023Taming the Panda with Python: A Powerful Duo for Seamless Robotics Programming and Integrationcited by 10
  20. 2023MaDi: Learning to Mask Distractions for Generalization in Visual Deep Reinforcement Learningcited by 10
  21. 2021An A* Curriculum Approach to Reinforcement Learning for RGBD Indoor Robot Navigationcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Zero-Shot Visual Generalization in Robot Manipulation
  2. 2023Visual-Policy Learning through Multi-Camera View to Single-Camera View Knowledge Distillation for Robot Manipulation Tasks
  3. 2022Masked World Models for Visual Control
  4. 2026FOCUS: From Privileged States to RGB-D with Controlled Modality Switching and Representation Alignment ↗
  5. 2024MENTOR: Mixture-of-Experts Network with Task-Oriented Perturbation for Visual Reinforcement Learning
  6. 2025PEEK: Guiding and Minimal Image Representations for Zero-Shot Generalization of Robot Manipulation Policies
  7. 2025Salience-Invariant Consistent Policy Learning for Generalization in Visual Reinforcement Learning
  8. 2025ManiVID-3D: Generalizable View-Invariant Reinforcement Learning for Robotic Manipulation via Disentangled 3D Representations