How to Peel with a Knife: Aligning Fine-Grained Manipulation with Human Preference
cs.RO, cs.AI, cs.CV, cs.LG, cs.SY, eess.SY
Submitted: 2026-03-03
Updated: 2026-11-02
Code: https://github.com/empriselab/gen3
Project page: https://toruowo.github.io/peel
Terminology
Sources
- Sim-to-Real Reinforcement Learning for Vision-Based Dexterous Manipulation on Humanoids
- DexGraspNet: A Large-Scale Robotic Dexterous Grasp Dataset for General Objects Based on Simulation
- DextrAH-G: Pixels-to-Action Dexterous Arm-Hand Grasping with Geometric Fabrics
- $\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization
- Contrastive Preference Learning: Learning from Human Feedback without RL
- PEBBLE: Feedback-Efficient Interactive Reinforcement Learning via Relabeling Experience and Unsupervised Pre-training
- FDPP: Fine-tune Diffusion Policy with Human Preference
- Reactive Diffusion Policy: Slow-Fast Visual-Tactile Policy Learning for Contact-Rich Manipulation
- Vegetable Peeling: A Case Study in Constrained Dexterous Manipulation
- Task and Joint Space Dual-Arm Compliant Control
- SAM 2: Segment Anything in Images and Videos
- Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots
- Learning Visuotactile Skills with Two Multifingered Hands
- Offline Reinforcement Learning with Implicit Q-Learning
- Residual Off-Policy RL for Finetuning Behavior Cloning Policies
- Self-Improving Vision-Language-Action Models with Data Generation via Residual RL
- GR-RL: Going Dexterous and Precise for Long-Horizon Robotic Manipulation
Related papers
- Rethinking Actions: Continuity-Aware Priors for World-Model Planning
- MOSAIC: Modular Obstacle and Spatial-Aware Framework for Manipulation In Cluttered Scenes
- SAGAS: Semantic-Aware Graph-Assisted Stitching for Offline Temporal Logic Planning
- DexSeed: Expanding Human Hand-Object Interactions into Robot Demonstrations for Dexterous Policy Learning
- LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning
- CIALLO: Learning Shared Intent Representations for Cross-Embodiment Manipulation