What failures should you collect? Comparing four data sources for robot reward models
mangahomanga · x · 2026-11-06
Opening a research thread on what failures to collect for training robot reward models in manipulation.
The team analyzes failure data from four sources — kinesthetic teaching, policy rollouts, human videos, and real2sim — and their impact on reward quality and downstream real-world RL. Collection costs differ hugely: teleoperation is high-effort while human videos are low-effort.
Related event: JHU Study Compares Four Failure Data Sources for Robot Reward Models(7 posts)→
More from Embodied
- Policy trained entirely in imagination matches PPO with 7.9× fewer simulator transitions — EpisodeYang · 2026-11-06
- Apple's mystery robot reportedly uses a Mac Mini as its planner brain — micoolcho · 2026-11-06
- Microdock v2 lets robot sit while docked, cutting servo power to zero — natesiggard · 2026-11-06
- Building an Asimov Robot in 22 Hours — and It Worked Out of the Box — whurley · 2026-11-06
- NEO-VLA drops the vision encoder: native VLA pretraining pays off on RoboTwin — liuziwei7 · 2026-11-06
- Gold Tesla Cybercab and Optimus shown at UP Summit — DMaguireARK · 2026-11-06