Result Details

A Comparative Study of Text Retrieval Models on DaReCzech

ŠTĚTINA, J.; FAJČÍK, M.; HRADIŠ, M.; ŠTEFÁNIK, M. A Comparative Study of Text Retrieval Models on DaReCzech. Recent Advances in Slavonic Natural Language Processing, 2024, no. 7, p. 85-100.
Type
journal article
Language
English
Authors
Štětina Jakub, Bc., FIT (FIT)
Fajčík Martin, Ing., Ph.D., DCGM (FIT)
Hradiš Michal, Ing., Ph.D., UAMT (FEEC), DCGM (FIT)
Štefánik Michal
Abstract

This article presents a comprehensive evaluation of off-the-shelf document retrieval models: Splade, Plaid, Plaid-X, SimCSE, Contriever, OpenAI ADA and Gemma2 chosen to determine their performance on the Czech retrieval dataset DaReCzech. The primary objective of our experiments is to estimate the quality of modern retrieval approaches in the Czech language. Our analyses include retrieval quality, speed, and memory footprint. Secondly, we analyze whether it is better to use the model directly in Czech text, or to use machine translation into English, followed by retrieval in English. Our experiments identify the most effective option for Czech information retrieval. The findings revealed notable performance differences among the models, with Gemma22 achieving the highest precision and recall, while Contriever performing poorly. Conclusively, SPLADE and PLAID models offered a balance of efficiency and performance.

Keywords

Information Retrieval; Evaluation; Comparison; Czech Language; Performance Assessment; Document Retrieval; Model Analysis

URL
Published
2024
Pages
85–100
Journal
Recent Advances in Slavonic Natural Language Processing, no. 7, ISSN
EID Scopus
2-s2.0-105004656879
BibTeX 
@article{BUT193747,
  author="Jakub {Štětina} and Martin {Fajčík} and Michal {Hradiš} and  {}",
  title="A Comparative Study of Text Retrieval Models on DaReCzech",
  journal="Recent Advances in Slavonic Natural Language Processing",
  year="2024",
  number="7",
  pages="85--100",
  issn="2336-4289",
  url="https://www.scopus.com/inward/record.uri?eid=2-s2.0-105004656879&partnerID=40&md5=952c98d094a89128a0781f11c0c3b8b8"
}
Projects
Multilingual and Cross-cultural interactions for context-aware, and bias-controlled dialogue systems for safety-critical applications, EU, HORIZON EUROPE, start: 2024-01-01, end: 2026-12-31, running
semANT - Semantic Document Exploration, MK, NAKI III – program na podporu aplikovaného výzkumu v oblasti národní a kulturní identity na léta 2023 až 2030, DH23P03OVV060, start: 2023-03-01, end: 2027-12-31, running
Departments
Fakulta informačních technologií (FIT)
Back to top