ADU-Bench is a new audio dialogue benchmark with 20,715 items across general talk, twelve skills, nine languages, and four ambiguity types, and it shows current audio-language models perform poorly on prosodic ambiguity.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Benchmarking Open-ended Audio Dialogue Understanding for Large Audio-Language Models
ADU-Bench is a new audio dialogue benchmark with 20,715 items across general talk, twelve skills, nine languages, and four ambiguity types, and it shows current audio-language models perform poorly on prosodic ambiguity.