President and co-founder at Prompt AI
Professor in the Division of Computer Science and EECS at University of California Berkeley
How media typically covers Trevor Darrell
Trevor Darrell as author
Vision-Language-Action models can be trained to reject impossible instructions by detecting false premises, clarifying ambiguities, and grounding alternatives in perception, improving false-premise detection by 97.56% over baselines.
“Author of the research paper on Vision-Language-Action models for robotics”