Chun-Yi Kuan

Instruction-level feedback from audio-aware LLMs can drastically enhance the accuracy of multi-event audio generation, bridging a critical gap in current models.

Audio-language models are blind to negation, with performance on negated sound concepts plummeting to below chance levels.

Semantic-level uncertainty estimation methods significantly enhance the reliability of audio-aware language models, outperforming traditional approaches in critical reasoning tasks.