作者位次 #2/4ACL 2026Findings· Findings of ACL
Don’t Just Listen, Try Planning: Graph-based Retrieval-Generation Agent for Long-form Audio Meeting Understanding
Quanwei Tang · Dong Zhang · Shoushan Li · Guodong Zhou
摘要
Long-form audio meeting understanding (LAMU) is gaining attention, but dedicated question answering (QA) datasets are lacking. Previous tailored speech QA and existing Speech LLMs suffer from acoustic information loss and poor long-term dependency capture. We construct the LongAudioQA dataset and...