Introduces cause-aware error detectors using ASR latent representations to classify token-level failures into perception, comprehension, and deletion, enabling targeted multi-turn clarifications that double recall on domain-shift errors and cut WER by up to 30%.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Proactive for Uncertainty: Cause-Aware Error Diagnosis and Interactive Clarification for Spoken Dialogue Systems
Introduces cause-aware error detectors using ASR latent representations to classify token-level failures into perception, comprehension, and deletion, enabling targeted multi-turn clarifications that double recall on domain-shift errors and cut WER by up to 30%.