Metraipolis › archive › dc_83f271a9f29c

Soft Posterior Speaker Injection for Multi-Talker Speech Recognition — arXiv:2609.01287

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.01287
License (as archived)
CC0-1.0
Archived
2026-10-17T07:19:02.447Z
SHA-256 of the archived text
36d88c869c76f1558e392b7f79869653768c8b43c10d2323851d2c9078d054e6
Stable URL of this page
https://metraipolis.com/documento/dc_83f271a9f29c

How to cite

Soft Posterior Speaker Injection for Multi-Talker Speech Recognition — arXiv:2609.01287 — arXiv (en), https://arxiv.org/abs/2609.01287 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_83f271a9f29c (Metraipolis archive dc_83f271a9f29c, sha256 36d88c869c76f1558e392b7f79869653768c8b43c10d2323851d2c9078d054e6)
BibTeX
@misc{dc83f271a9f29c,
  title = {Soft Posterior Speaker Injection for Multi-Talker Speech Recognition — arXiv:2609.01287},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.01287},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_83f271a9f29c; sha256 36d88c869c76f1558e392b7f79869653768c8b43c10d2323851d2c9078d054e6},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01287. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multi-talker automatic speech recognition (MT-ASR) remains challenging in the presence of overlapping speech. Hard segmentation introduces irreversible errors, whereas serialized output training (SOT) avoids explicit segmentation but does not condition a pretrained encoder on speaker activity. We propose Soft Posterior Speaker Injection (SPSI). A Soft Posterior Head predicts per-frame speaker posteriors $\hat{\mathbf{P}}$ and injects them into Whisper through Multi-layer Feature-wise Linear Modulation (MFLM) and Speaker Memory Prompts (SMP). The benefit of SPSI is largest where overlap is heaviest and under domain transfer. On controlled two-speaker LibriSpeech overlap, SPSI reduces concatenated minimum-permutation word error rate (cpWER) from $61.5\%$ to $60.0\%$ in the high-overlap bin, and from $51.9\%$ to $51.0\%$ on the full set, relative to SOT. By contrast, Speaker CE, SD-CTC, SA-DiCoW, and Pipeline (oracle/est.\ VAD) do not outperform SOT. Freeze-posterior overlap-heavy adaptation reduces held-out LibriCSS cpWER from $42.3\%$ to $36.8\%$ on sessions $8$--$9$, a $5.5$-point gain over SOT. The source code is available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.MM).
Categorie arXiv: Sound (cs.SD) ; Multimedia (cs.MM).
Pagina del lavoro: https://arxiv.org/abs/2609.01287 — 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 2026-10-17T07:07:19.991Z dalla pagina https://arxiv.org/catchup/cs.MM/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).