Metraipolis › archive › dc_3468488951e5

Task-Conditional Flow Matching for Balanced Multilingual Text Embedding Adaptation — arXiv:2608.05785

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.05785
License (as archived)
CC0-1.0
Archived
2026-10-25T21:27:13.727Z
SHA-256 of the archived text
638e06968c5022c92f4d8464204d8d140717dc9dd44143aa5846748a9e24869d
Stable URL of this page
https://metraipolis.com/documento/dc_3468488951e5

How to cite

Task-Conditional Flow Matching for Balanced Multilingual Text Embedding Adaptation — arXiv:2608.05785 — arXiv (en), https://arxiv.org/abs/2608.05785 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3468488951e5 (Metraipolis archive dc_3468488951e5, sha256 638e06968c5022c92f4d8464204d8d140717dc9dd44143aa5846748a9e24869d)
BibTeX
@misc{dc3468488951e5,
  title = {Task-Conditional Flow Matching for Balanced Multilingual Text Embedding Adaptation — arXiv:2608.05785},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.05785},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3468488951e5; sha256 638e06968c5022c92f4d8464204d8d140717dc9dd44143aa5846748a9e24869d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.05785. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Multilingual text embedding models are commonly adapted using a single training objective across diverse tasks, despite different tasks requiring fundamentally different optimization strategies. We introduce Task-Conditional Flow Matching (TCFM), a multilingual embedding adaptation framework that selectively applies Flow Matching to translation tasks while optimizing retrieval, classification, and pair-classification tasks with objectives better aligned to their learning dynamics. TCFM further combines teacher-guided representation preservation with a three-stage curriculum to enable stable adaptation. Evaluated on the Indic Massive Text Embedding Benchmark, TCFM establishes a new state-of-the-art, consistently improving embedding quality across a diverse set of multilingual tasks and generalizing across embedding model families. We will publicly release the codebase and datasets upon acceptance of the paper.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-07 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.05785 — 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 2026-10-25T21:24:38.205Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).