Metraipolis › archive › dc_ca36107aae1a

Refined Assessment for Translation Evaluation: Rethinking Machine Translation Evaluation in the Era of Human-Level Systems — arXiv:2610.33749

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.33749
License (as archived)
CC0-1.0
Archived
2026-10-27T08:11:24.929Z
SHA-256 of the archived text
bb18138adac43a53fea1905380ab00bed4881db23f01e7901f13f3776fe6c422
Stable URL of this page
https://metraipolis.com/documento/dc_ca36107aae1a

How to cite

Refined Assessment for Translation Evaluation: Rethinking Machine Translation Evaluation in the Era of Human-Level Systems — arXiv:2610.33749 — arXiv (en), https://arxiv.org/abs/2610.33749 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ca36107aae1a (Metraipolis archive dc_ca36107aae1a, sha256 bb18138adac43a53fea1905380ab00bed4881db23f01e7901f13f3776fe6c422)
BibTeX
@misc{dcca36107aae1a,
  title = {Refined Assessment for Translation Evaluation: Rethinking Machine Translation Evaluation in the Era of Human-Level Systems — arXiv:2610.33749},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.33749},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ca36107aae1a; sha256 bb18138adac43a53fea1905380ab00bed4881db23f01e7901f13f3776fe6c422},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.33749. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

As machine translation systems approach human-level quality, traditional evaluation methodologies struggle to detect subtle translation errors. We critically examine limitations in current gold-standard approaches (MQM and ESA), including inconsistencies from variable annotator expertise, excessive categorization complexity, coarse severity granularity, accuracy bias over fluency, and time constraints. To address this issue, we introduce a high-quality dataset consisting of human evaluations for English-Russian translations from WMT24, created by professional linguists. We show that expert assessments without time pressure yield substantially different results from standard evaluations. To enable consistent and rich annotation by these experts, we developed the RATE (Refined Assessment for Translation Evaluation) protocol. RATE provides a streamlined error taxonomy, expanded severity ratings, and multidimensional scoring balancing accuracy and fluency, facilitating deeper analysis of MT outputs. Our analysis, powered by this expert dataset, reveals that state-of-the-art MT systems may have surpassed human translations in accuracy while still lagging in fluency - a critical distinction obscured by existing accuracy-biased metrics. Our findings highlight that advancing MT evaluation requires not only better protocols but crucially, high-quality annotations from skilled linguists.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.33749 — 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 2026-10-27T07:56:00.207Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).