# Bangla speech model improves detection of when speakers finish

_Published Friday, September 25, 2026 at 7:08 AM EDT · AI, Science · Latest · Tier 2 — Notable_

Researchers report that a model trained to detect when a Bangla speaker has finished talking reached 84.33% accuracy on a held-out podcast test set, compared with 69.28% for the Smart-Turn v3 baseline. Its false negative rate fell from 51.57% to 7.55%, which could help a voice system avoid missing the moment to respond.

The model uses a Whisper speech encoder with classification components trained on 35,374 labelled podcast samples. Its measured processing time was 165 to 191 milliseconds on a CPU. The lower false negative rate came with a higher false positive rate, meaning the system more often risks treating an unfinished turn as complete.

## Sources

- [cs.CL updates on arXiv.org](https://arxiv.org/abs/2609.29371)

---
Canonical: https://techandbusiness.org/newswire/WtRc7XS1XjsSdSflQRxOBS
Published: 2026-09-25T11:08:06.898Z
Story chronology: 2026-09-25T04:00:00.000Z
Retrieved: 2026-09-25T12:34:33.018Z
Publisher: Tech & Business (techandbusiness.org)