LongEval at CLEF 2025：信息检索模型性能的纵向评估
信息检索
2025-03-12 v1
摘要
本文介绍了 CLEF 2025 conference 的第三届 LongEval 实验室，继续探索信息检索（IR）中时间持久性挑战。该实验室包含两个任务，旨InList researchers 提供反映用户查询和文档相关性随时间演变的测试数据。通过评估模型性能如何随测试数据与训练数据的时间差异而退化，LongEval 旨在推进对 IR 系统中时间动态的理解。2025 版旨在吸引 IR 和 NLP 社区共同应对开发能够在 web 搜索和科学检索领域维持检索质量的自适应模型。
引用
@article{arxiv.2503.08541,
title = {LongEval at CLEF 2025: Longitudinal Evaluation of IR Model Performance},
author = {Matteo Cancellieri and Alaa El-Ebshihy and Tobias Fink and Petra Galuščáková and Gabriela Gonzalez-Saez and Lorraine Goeuriot and David Iommi and Jüri Keller and Petr Knoth and Philippe Mulhem and Florina Piroi and David Pride and Philipp Schaer},
journal= {arXiv preprint arXiv:2503.08541},
year = {2025}
}
备注
Accepted for ECIR 2025. To be published in Advances in Information Retrieval - 47th European Conference on Information Retrieval, ECIR 2025, Lucca, Italy, April 6-10, 2025, Proceedings