Introduces TPI-Train dataset and TPI-Bench to mitigate semantic shortcut learning in SLMs by enforcing acoustic cue prioritization for third-party interruption handling.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Audio-Mind introduces a conditional, auditable agentic framework for audio understanding that preserves frontend judgment and acquires bounded external evidence only when needed, reporting 80.4% on MMAR and 82.8% on MSU-Bench.
citing papers explorer
-
Still Between Us? Evaluating and Improving Voice Assistant Robustness to Third-Party Interruptions
Introduces TPI-Train dataset and TPI-Bench to mitigate semantic shortcut learning in SLMs by enforcing acoustic cue prioritization for third-party interruption handling.
-
Audio-Mind: An Auditable Agentic Framework for Audio Understanding
Audio-Mind introduces a conditional, auditable agentic framework for audio understanding that preserves frontend judgment and acquires bounded external evidence only when needed, reporting 80.4% on MMAR and 82.8% on MSU-Bench.