CUNI at WMT23 General Translation Task: MT and a Genetic Algorithm
The result's identifiers
Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F23%3A10475903" target="_blank" >RIV/00216208:11320/23:10475903 - isvavai.cz</a>
Result on the web
<a href="https://aclanthology.org/2023.wmt-1.8/" target="_blank" >https://aclanthology.org/2023.wmt-1.8/</a>
DOI - Digital Object Identifier
<a href="http://dx.doi.org/10.18653/v1/2023.wmt-1.8" target="_blank" >10.18653/v1/2023.wmt-1.8</a>
Alternative languages
Result language
angličtina
Original language name
CUNI at WMT23 General Translation Task: MT and a Genetic Algorithm
Original language description
This paper describes the submissions of different Charles University teams to WMT23 General translation task (English to Czech and Czech to Ukrainian translation directions). Our main submission, CUNI-GA, is a result of applying a novel n-best list reranking and modification method on translation candidates produced by the two other submitted systems, CUNI-Transformer and CUNI-DocTransformer (document-level translation only used for the $en rightarrow cs$ direction). Our method uses a genetic algorithm and MBR decoding to search for optimal translation under a given metric (in our case, a weighted combination of ChrF, BLEU, COMET22-DA, and COMET22-QE-DA). Based on the automatic metrics, our submissions are first in the constrained track and competitive with the best unconstrained systems according to some of the metrics.
Czech name
—
Czech description
—
Classification
Type
D - Article in proceedings
CEP classification
—
OECD FORD branch
10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)
Result continuities
Project
Result was created during the realization of more than one project. More information in the Projects tab.
Continuities
P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)
Others
Publication year
2023
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data specific for result type
Article name in the collection
Proceedings of the Eighth Conference on Machine Translation
ISBN
979-8-89176-041-7
ISSN
—
e-ISSN
—
Number of pages
9
Pages from-to
119-127
Publisher name
Association for Computational Linguistics
Place of publication
Stroudsburg, PA, USA
Event location
Singapore
Event date
Dec 6, 2023
Type of event by nationality
WRD - Celosvětová akce
UT code for WoS article
—