Takeshi Homma

Improving the Naturalness of Simulated Conversations for End-to-End Neural Diarization

2022 · arXiv (Cornell University)

This paper investigates a method for simulating natural conversation in the model training of end-to-end neural diarization (EEND). Due to the lack of any annotated real conversational dataset, EEND is usually pretrained on a large-scale …