CUNI-ÚFAL at the WMT26 Terminology Translation Task

Aleš Manuel Papáček, Tomáš Sourada


Abstract
This paper describes the CUNI-ÚFAL submission to the WMT26 Terminology Translation Task. We present a model-agnostic pipeline that selects document-relevant terminology for a general-purpose LLM, instantiated with GPT-5.6-Sol for both tracks. For Track1 (explicit dictionary) we compare glossary placement, inline annotation, dictionary filtering, and terminology-aware revision: filtering the dictionary per document improves judged coverage and, slightly, translation quality. Inline term insertion buys a small coverage gain at the cost of fluency. For Track2 (in-domain seed bitexts) we compare presenting them as translation history, extracting a dictionary and reusing the Track1 pipeline, and their hybrid. Lacking development references, we evaluated with surface- and lemma-based term matching plus an LLM judge. Among submitted systems in the official evaluation, our Track1 submission ties for the highest COMET-22 score and obtains the best XCOMET-XXL and MetricX-24 XL scores, and our primary Track2 submission achieves the highest term success.
Anthology ID:
2026.wmt-main.120
Volume:
Proceedings of the Eleventh Conference on Machine Translation
Month:
October
Year:
2026
Address:
Budapest, Hungary
Editors:
Barry Haddow, Tom Kocmi, Philipp Koehn, Christof Monz
Venues:
WMT | WS
SIG:
SIGMT
Publisher:
Association for Computational Linguistics
Note:
Pages:
1849–1859
Language:
URL:
https://aclanthology.org/2026.wmt-main.120/
DOI:
Bibkey:
Cite (ACL):
Aleš Manuel Papáček and Tomáš Sourada. 2026. CUNI-ÚFAL at the WMT26 Terminology Translation Task. In Proceedings of the Eleventh Conference on Machine Translation, pages 1849–1859, Budapest, Hungary. Association for Computational Linguistics.
Cite (Informal):
CUNI-ÚFAL at the WMT26 Terminology Translation Task (Papáček & Sourada, WMT 2026)
Copy Citation:
PDF:
https://aclanthology.org/2026.wmt-main.120.pdf