Smart Note · Series
Robot Data
Data sources, collection methods, and data systems for robot learning.
10 articles

Robot Data EP01: Yang Gao and Chuan Wen on Learning Robot Policies from Internet Videos with ATM
The ATM authors discuss point trajectories, cross-embodiment learning and using human videos to reduce the need for action-labeled robot data.

Robot Data, Episode 2: Cheng Chi and Zhenjia Xu on UMI, a Physical API for Robot Training
A 2024 conversation with Cheng Chi and Zhenjia Xu about UMI’s handheld hardware, data collection, policy learning and generalization.

Robot Data: Xuxin Cheng on Immersive Robot Teleoperation with Open-TeleVision
Xuxin Cheng discusses stereo vision, active head movement, motion retargeting and learning robot skills from immersive teleoperation.

Robot Data, Episode 4: Chen Wang on MimicPlay, DexCap and Learning from Human Motion
The full interview on hierarchical planning, human play, wearable motion capture, retargeting, dexterous control and the differences between DexCap and UMI.

Robot Data EP05: Zipeng Fu on HumanPlus and Mobile ALOHA—From Synthetic Data to Real Data, and Combining the Two
Zipeng Fu discusses how HumanPlus and Mobile ALOHA use simulation, human motion and real-world teleoperation data to train robots.

Robot Data, Episode 6: Yecheng Jason Ma on Eureka, DrEureka and LLMs for Simulation and Reinforcement Learning
A full conversation on automatic reward design, domain randomization, sim-to-real transfer, reinforcement learning and VIP.

Robot Data, Episode 7: Hang Zhao and Ziwen Zhuang on Parkour and Autonomous Driving
The complete discussion of locomotion, simulation, reinforcement learning, RGB perception and lessons from autonomous-driving data loops.

Robot Data, Episode 8: Chuang Gan and Zhou Xian on RoboGen, Generative Simulation and Differentiable Physics (With a Research Review)
A full interview on generating robot demonstrations, learning skills with differentiable simulation, and the research behind key physics platforms.

Robot Data, Episode 9: Lirui Wang on Heterogeneous Robot Data and Policy Composition
The complete PoCo interview on combining simulation, human video, real robots and tactile information, followed by HPT, fleet learning and a personal reflection.

Robot Data, Episode 10: A Full Review of Season 1
Revisiting all nine Season 1 interviews: human videos, simulation, motion capture, teleoperation and the composition of heterogeneous robot data.
No matching articles.