Search papers, labs, and topics across Lattice.
Chinese University of Hong Kong
1
0
3
Speech QA models achieve peak performance at 4.17 Hz, revealing the critical role of frame rate and representation alignment in bridging the gap between speech and text reasoning.
