Improving cross-lingual representation for semantic retrieval with code-switching
The result's identifiers
Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F26%3AXMMRBREF" target="_blank" >RIV/00216208:11320/26:XMMRBREF - isvavai.cz</a>
Result on the web
<a href="http://dx.doi.org/10.1016/j.knosys.2025.113919" target="_blank" >http://dx.doi.org/10.1016/j.knosys.2025.113919</a>
DOI - Digital Object Identifier
<a href="http://dx.doi.org/10.1016/j.knosys.2025.113919" target="_blank" >10.1016/j.knosys.2025.113919</a>
Alternative languages
Result language
angličtina
Original language name
Improving cross-lingual representation for semantic retrieval with code-switching
Original language description
Semantic Retrieval (SR) has become an indispensable part of the FAQ system in the task-oriented question-answering (QA) dialogue scenario. The demand for a cross-lingual smart customer service system for e-commerce platforms and specific business scenarios has been increasing recently Most previous studies directly exploit cross-lingual pre-trained models (PTMs) for multilingual knowledge retrieval, while some also incorporate continual pre-training before fine-tuning PTMs on downstream tasks. However, no matter which schema is used, the previous work ignores to inform PTMs of some features of the downstream task, i.e. train their PTMs without providing any signals related to the downstream task (e.g., SR). To this end, in this work, we propose an Alternative Cross-lingual PTM for SR via code-switching. We are the first to utilize the code-switching approach for cross-lingual SR. Besides, we introduce the novel code-switched continual pre-training instead of directly using the PTMs on the SR tasks. The experimental results show that our proposed approach consistently outperforms the previous SOTA methods on SR and semantic textual similarity (STS) tasks with three business corpora and four open datasets in 20+ languages. Our approach outperforms the strongest baseline over 3.7 points on these datasets on average. The code is available at https://github.com/miradel51/codemix_ptm. © 2025
Czech name
—
Czech description
—
Classification
Type
J<sub>SC</sub> - Article in a specialist periodical, which is included in the SCOPUS database
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
—
Continuities
—
Others
Publication year
2025
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data specific for result type
Name of the periodical
Knowledge-Based Systems
ISSN
0950-7051
e-ISSN
—
Volume of the periodical
325
Issue of the periodical within the volume
2025
Country of publishing house
US - UNITED STATES
Number of pages
32
Pages from-to
113919
UT code for WoS article
—
EID of the result in the Scopus database
2-s2.0-105009512698