Metraipolis › archive › dc_ed274c1944f4

Efficient One-to-Many Translation with Joint Multi-Stream Diffusion — arXiv:2609.16312

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.16312
License (as archived): CC0-1.0
Archived: 2026-10-15T18:32:48.911Z
SHA-256 of the archived text: 63162bba5e6e38a59f46f0011f27a3fc737af02ec171d039ba5442f10ce7712a
Stable URL of this page: https://metraipolis.com/documento/dc_ed274c1944f4

How to cite

Efficient One-to-Many Translation with Joint Multi-Stream Diffusion — arXiv:2609.16312 — arXiv (en), https://arxiv.org/abs/2609.16312 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ed274c1944f4 (Metraipolis archive dc_ed274c1944f4, sha256 63162bba5e6e38a59f46f0011f27a3fc737af02ec171d039ba5442f10ce7712a)

BibTeX

@misc{dced274c1944f4,
  title = {Efficient One-to-Many Translation with Joint Multi-Stream Diffusion — arXiv:2609.16312},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.16312},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ed274c1944f4; sha256 63162bba5e6e38a59f46f0011f27a3fc737af02ec171d039ba5442f10ce7712a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.16312. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

One-to-many machine translation (MT) is computationally expensive for autoregressive (AR) systems, which suffer from linear latency scaling with both sequence length and the number of target languages. We explore how diffusion can enable multilingual translation with a discrete diffusion framework that refines all target languages in parallel, achieving sublinear latency scaling with the number of targets, and supports deployment as a single unified model to replace multiple independent systems. Conditioned on a continuous semantic anchor rather than source tokens, our framework supports zero-shot transfer to unseen source languages without retraining, maintaining approximately $75\%$ of its supervised translation quality on zero-shot sources. We investigate the quality-latency frontier and find that with accelerated sampling, it achieves comparable supervised quality to AR baselines with a $2 \times$ speedup and $11.9\%$ better zero-shot BLEU. These results highlight the potential of joint multi-stream diffusion as a practical and flexible alternative for efficient one-to-many translation.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-16 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.16312 — 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 2026-10-15T18:03:09.614Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-16?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).