Monkey See, Can Monkey Do? A Benchmark for Evaluating Robot Skill Learning by Observation
cs.RO
Submitted: 2026-09-08
Updated: 2026-10-15
Comments: 31 pages, 8 tables, 11 figures. In Proceedings of CoRL 2026
Code: https://github.com/haosulab/MPlib
Project page: https://roboreel.github.io
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Robot Learning from Human Videos: A Survey
- Learning from Observation: A Survey of Recent Advances
- Human2Robot: Learning Robot Actions from Paired Human-Robot Videos
- H2R: A Human-to-Robot Data Augmentation for Robot Pre-training from Videos
- XSkill: Cross Embodiment Skill Discovery
- Learning Generalizable Robot Policy with Human Demonstration Video as a Prompt
- PhysBrain: Human Egocentric Data as a Bridge from Vision Language Models to Physical Intelligence
- One-Shot Imitation under Mismatched Execution
- R3M: A Universal Visual Representation for Robot Manipulation
- Masked Visual Pre-training for Motor Control
- Motion Tracks: A Unified Representation for Human-Robot Transfer in Few-Shot Imitation Learning
- Point Policy: Unifying Observations and Actions with Key Points for Robot Manipulation
- R+X: Retrieval and Execution from Everyday Human Videos
- Keypoint Action Tokens Enable In-Context Imitation Learning in Robotics
- RH20T: A Comprehensive Robotic Dataset for Learning Diverse Skills in One-Shot
- UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
- HowTo100M: Learning a Text-Video Embedding by Watching Hundred Million Narrated Video Clips
- Assembly101: A Large-Scale Multi-View Video Dataset for Understanding Procedural Activities
- Ego-Exo4D: Understanding Skilled Human Activity from First- and Third-Person Perspectives
- Scaling Egocentric Vision: The EPIC-KITCHENS Dataset
Related papers
- FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
- MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
- RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
- ARC: A Reasoning Recipe for Robot Foundation Models
- HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
- APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies