1 paper · 1 filter
Yunrui Cai, Dingdong Wang, Lingwei Meng +3
Joint multi-talker ASR with speaker attribution and timestamps is particularly challenging when the speaker count is unknown and the output must follow a strict structure. We obser…