Prosody-based token-level goal/detail classification, combined with in-context LLM prompting, disambiguates robot instructions better than text-only processing.
We highlight speech instruction disambiguation as a key research area in robotics and curate a novel dataset of 1,540 ambiguous utterances, to advance future work
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
-
Enhancing Speech Instruction Understanding and Disambiguation in Robotics via Speech Prosody
Prosody-based token-level goal/detail classification, combined with in-context LLM prompting, disambiguates robot instructions better than text-only processing.