Berkeley Proposes Data Inoculation: RL to Improve Robot Demo Quality During Collection
A UC Berkeley team including Sergey Levine proposed Data Inoculation, a method that uses reinforcement learning to improve the quality of human demonstration data at collection time, arguing that data quality matters as much as or more than the model in robot learning.
2026-11-03 ~ 2026-11-04 · 2 related posts
- Berkeley's Data Inoculation: Using RL During Data Collection for Better Robot Demos — berkeley_ai · 2026-11-03
- Berkeley researchers use RL to improve robot data collection with Data Inoculation — berkeley_ai · 2026-11-04