Spoken Function Calling: A New Perspective on Spoken Language Understanding for Large Audio Language Models
The paper reframes spoken language understanding as structured function calling for audio models.
The authors propose Spoken Function Calling to replace looser SLU rule definitions with clearer structured rules. They build SFC-Bench with a multi-agent synthesis pipeline and test both LLMs and large audio language models. In their experiments, SFC improves semantic extraction accuracy over traditional SLU, and post-training further strengthens LALM performance. ArXiv · AI/CL/LG's note
The authors propose Spoken Function Calling to replace looser SLU rule definitions with clearer structured rules. They build SFC-Bench with a multi-agent synthesis pipeline and test both LLMs and large audio language models. In their experiments, SFC improves semantic extraction accuracy over traditional SLU, and post-training further strengthens LALM performance. ArXiv · AI/CL/LG's note
score 4