1 paper · 1 filter
Siyuan Zhang, Jian Zong, Junyu Wang +7
While LALMs show promise on audio question answering, they fail to focus on question-relevant segments of audio and provide a clear, checkable reasoning process when dealing with c…