Metraipolis › archive › dc_8085486bdbd4

BanglaTurn: A Benchmark and Whisper-Based Model for End-of-Turn Detection in Bangla Speech — arXiv:2609.29371

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.29371
License (as archived)
CC0-1.0
Archived
2026-10-08T04:49:29.294Z
SHA-256 of the archived text
70cba43162c9a141b4fdf9f3acb33ed77fa7403cda65f541dfd78dab6336061c
Stable URL of this page
https://metraipolis.com/documento/dc_8085486bdbd4

How to cite

BanglaTurn: A Benchmark and Whisper-Based Model for End-of-Turn Detection in Bangla Speech — arXiv:2609.29371 — arXiv (en), https://arxiv.org/abs/2609.29371 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8085486bdbd4 (Metraipolis archive dc_8085486bdbd4, sha256 70cba43162c9a141b4fdf9f3acb33ed77fa7403cda65f541dfd78dab6336061c)
BibTeX
@misc{dc8085486bdbd4,
  title = {BanglaTurn: A Benchmark and Whisper-Based Model for End-of-Turn Detection in Bangla Speech — arXiv:2609.29371},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.29371},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8085486bdbd4; sha256 70cba43162c9a141b4fdf9f3acb33ed77fa7403cda65f541dfd78dab6336061c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.29371. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

This paper presents BanglaTurn, a corpus for end-of-turn detection in Bangla conversational speech, and a model trained on it. The corpus holds 35,374 samples of 3 to 15 s of podcast speech, labelled for turn state by combining speaker diarization with an LLM pass, with every label then checked by a human annotator. The model pairs a Whisper encoder with task-specific classification heads. On a class-balanced test set drawn from a held-out podcast, it reaches 84.33% accuracy (95% CI 80.3 to 88.1) against 69.28% for the Smart-Turn v3 baseline, and lowers the false negative rate from 51.57% to 7.55% at the cost of a higher false positive rate. We report what encoder layer fine-tuning, multi-scale pooling and INT8 quantization each contribute, and latency stays within 165 to 191 ms end to end on CPU.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-25 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.29371 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T04:27:20.049Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).