Metraipolis › archive › dc_ad1982f5d809
Preference-Aware Pareto Steering for Multi-Task Speech LLMs — arXiv:2610.27200
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.27200
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-21T07:08:49.464Z
- SHA-256 of the archived text
a670162c4e4853b8d80a26330486e488c62d2f22873e0aa9f09aaa1b1c279ad9- Stable URL of this page
https://metraipolis.com/documento/dc_ad1982f5d809
How to cite
Preference-Aware Pareto Steering for Multi-Task Speech LLMs — arXiv:2610.27200 — arXiv (en), https://arxiv.org/abs/2610.27200 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ad1982f5d809 (Metraipolis archive dc_ad1982f5d809, sha256 a670162c4e4853b8d80a26330486e488c62d2f22873e0aa9f09aaa1b1c279ad9)
BibTeX
@misc{dcad1982f5d809,
title = {Preference-Aware Pareto Steering for Multi-Task Speech LLMs — arXiv:2610.27200},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.27200},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ad1982f5d809; sha256 a670162c4e4853b8d80a26330486e488c62d2f22873e0aa9f09aaa1b1c279ad9},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27200. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Speech large language models (LLMs) perform multiple tasks, including automatic speech recognition (ASR), timestamp prediction (TS), and speech translation (ST). Joint training of speech LLMs creates task trade-offs, yet standard training with fixed loss weights does not generally control the resulting trade-off according to deployment preferences. To address this, we propose Endpoint-Anchored Preference-Aware Sampling Uniformly along the Pareto Front (EA-SURF), an a priori method that adapts scalarization weights so that the resulting task trade-off reflects the desired preference. EA-SURF retains the standard weighted-loss training pipeline, tracking the current model's performance relative to each task-specific reference and adapting the scalarization weights as training proceeds. We theoretically show that EA-SURF can match a desired preference, with the update becoming stationary at the target under checkable regularity conditions. Across six three-objective speech LLM requests, EA-SURF empirically converges toward the requested targets and increases endpoint-inclusive normalized hypervolume by 17.8% over fixed loss weighting. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL) ; Sound (cs.SD). Pagina del lavoro: https://arxiv.org/abs/2610.27200 — 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 2026-10-21T07:05:58.101Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).