First demonstration that LLM-generated reward functions using tactile sensing can outperform a human-engineered baseline for real-world in-hand rotation.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Text2Touch: Tactile In-Hand Manipulation with LLM-Designed Reward Functions
First demonstration that LLM-generated reward functions using tactile sensing can outperform a human-engineered baseline for real-world in-hand rotation.