Hugging Face Trending Papers

The Role of Disfluencies in Speech Translation

Read the original on Hugging Face Trending Papers →

Current speech translation systems, including SpeechLLMs, are trained on cleaned text and tend to strip disfluencies like filled pauses and false starts rather than translate them. We show this comes at a cost: disfluencies carry meaning that gets lost when speech is cleaned up.

Summary generated by The Flow from the publisher's feed. The full article lives at Hugging Face Trending Papers.