Metraipolis › archive › dc_296c22011fd4

Fairness Beyond a Single Run: Training-Seed Variability in Speech LLM Adaptation — arXiv:2609.38976

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.38976
License (as archived): CC0-1.0
Archived: 2026-10-07T13:13:36.671Z
SHA-256 of the archived text: c7706a2d250d7b3b1957c0592a29564fb8b96643abad999113b4e547dec634b7
Stable URL of this page: https://metraipolis.com/documento/dc_296c22011fd4

How to cite

Fairness Beyond a Single Run: Training-Seed Variability in Speech LLM Adaptation — arXiv:2609.38976 — arXiv (en), https://arxiv.org/abs/2609.38976 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_296c22011fd4 (Metraipolis archive dc_296c22011fd4, sha256 c7706a2d250d7b3b1957c0592a29564fb8b96643abad999113b4e547dec634b7)

BibTeX

@misc{dc296c22011fd4,
  title = {Fairness Beyond a Single Run: Training-Seed Variability in Speech LLM Adaptation — arXiv:2609.38976},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.38976},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_296c22011fd4; sha256 c7706a2d250d7b3b1957c0592a29564fb8b96643abad999113b4e547dec634b7},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.38976. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Demographic fairness gaps in automatic speech recognition are almost always reported from a single training run. We fine-tune the Q-former projector and LoRA adapters of a speech LLM at five audio compression factors and six random seeds, holding the encoder, base decoder, data and decoding fixed, and evaluate every run on Common Voice and Fair-Speech. At 460 h of clean LibriSpeech, the seed moves fairness metrics more than compression does on most demographic axes. A balanced 3x3 decomposition attributes 85.3% of the variation in Fair-Speech ethnicity normalized gap to the seed against 8.3% to compression (p = 0.009), though compression explains more on age and gender. Held-out LibriSpeech word error rate spreads by 0.04 points across those seeds while Common Voice spreads by 8.57, so these are not failed runs, and the effect survives controlling for accuracy and dropout. Scaling and diversifying the adaptation set to 960 h damps the effect but does not remove it. On Fair-Speech ethnicity, two single-run systems must differ by more than 0.30 in normalized gap to exceed seed variability.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.38976 — 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 2026-10-07T12:44:51.778Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).