# Next-Turn: Duration-Aware Streaming Endpoint Detection via Time-to-Next-Speech-Onset Prediction

arXiv: https://arxiv.org/abs/2606.18094 · HTML: https://mlsanity.com/paper/2606.18094/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Jun 2026.

## Built on by (most cited first, 1 of 1)

- [FD-VAD: Semantic Endpoint Detection for Streaming Full-Duplex Speech](https://mlsanity.com/paper/2609.35791/) (2026)

## Abstract (arXiv)

Endpoint detection (EPD) is essential for natural turn-taking in streaming speech systems. However, reliably determining the endpoint of an utterance is challenging because speakers often pause mid-utterance due to hesitations and disfluencies. Semantic EPD has emerged as a promising direction to address this issue but is hindered by ambiguous supervision and strict streaming constraints. We propose Next-Turn that uses the time-to-next-speech-onset as the training objective, where targets are derived directly from speech timestamps and require no additional annotation. Experiments show that the proposed method outperforms conventional acoustic and recent semantic EPD baselines, achieving a 25.9% absolute improvement in endpoint accuracy within 320 ms over the strongest baseline. In addition, joint training with the duration-aware objective complements standard binary EPD, with gains that increase monotonically with increasing pauses.