Search papers, labs, and topics across Lattice.
3
0
6
9
LLMs can effectively aggregate diverse speech quality metrics, even outperforming specialized models when labeled data is scarce.
Gaze is a surprisingly effective cue for resolving the cocktail party problem, boosting audio-visual speech enhancement by over 23% in SI-SDR.
Speech quality assessment is skewed: male listeners consistently give higher scores than female listeners, and standard MOS models learn and perpetuate this bias.
