1 paper
Linkai Peng, Baorian Nuchged, Kaiqi Fu +1
Full-duplex speech models listen and speak at once, promising always-on assistants. Yet they must also decide when they should speak. Human listeners speak when addressed or when t…