Embodied Task Agents Can Perform Fine Manipulation through Tactile In-Context Learning

Key points

• We investigate this capability through Tactile-ETA, a system that incorporates tactile experience into an embodied task agent through in-context learning.1
• Successful expert demonstrations show how contact changes during manipulation.1
• We pair each recorded robot movement with its camera views and bilateral tactile observations.1
• These findings demonstrate the feasibility of using tactile experience in context for fine manipulation and suggest a path toward extending embodied task agents to additional sensory inputs.1

Sources (1)

• [1]Embodied Task Agents Can Perform Fine Manipulation through Tactile In-Context Learning
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 20, 04:27 AM

  Can embodied task agents use tactile experience to perform fine manipulation without task-specific model training?

  We investigate this capability through Tactile-ETA, a system that incorporates tactile experience into an embodied task agent through in-context learning.

Extractive summary: sentences quoted from the sources.