Saved in:
Bibliographic Details
Main Authors: Lee, Daniel, Sharma, Harsh, Han, Jieun, Jeong, Sunny, Oh, Alice, Shwartz, Vered
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.20451
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916712416280576
author Lee, Daniel
Sharma, Harsh
Han, Jieun
Jeong, Sunny
Oh, Alice
Shwartz, Vered
author_facet Lee, Daniel
Sharma, Harsh
Han, Jieun
Jeong, Sunny
Oh, Alice
Shwartz, Vered
contents Translating knowledge-intensive and entity-rich text between English and Korean requires transcreation to preserve language-specific and cultural nuances beyond literal, phonetic or word-for-word conversion. We evaluate 13 models (LLMs and MT models) using automatic metrics and human assessment by bilingual annotators. Our findings show LLMs outperform traditional MT systems but struggle with entity translation requiring cultural adaptation. By constructing an error taxonomy, we identify incorrect responses and entity name errors as key issues, with performance varying by entity type and popularity level. This work exposes gaps in automatic evaluation metrics and hope to enable future work in completing culturally-nuanced machine translation.
format Preprint
id arxiv_https___arxiv_org_abs_2504_20451
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Team ACK at SemEval-2025 Task 2: Beyond Word-for-Word Machine Translation for English-Korean Pairs
Lee, Daniel
Sharma, Harsh
Han, Jieun
Jeong, Sunny
Oh, Alice
Shwartz, Vered
Computation and Language
Information Retrieval
Translating knowledge-intensive and entity-rich text between English and Korean requires transcreation to preserve language-specific and cultural nuances beyond literal, phonetic or word-for-word conversion. We evaluate 13 models (LLMs and MT models) using automatic metrics and human assessment by bilingual annotators. Our findings show LLMs outperform traditional MT systems but struggle with entity translation requiring cultural adaptation. By constructing an error taxonomy, we identify incorrect responses and entity name errors as key issues, with performance varying by entity type and popularity level. This work exposes gaps in automatic evaluation metrics and hope to enable future work in completing culturally-nuanced machine translation.
title Team ACK at SemEval-2025 Task 2: Beyond Word-for-Word Machine Translation for English-Korean Pairs
topic Computation and Language
Information Retrieval
url https://arxiv.org/abs/2504.20451