SEUPD@CLEF: Team CLOSE on Temporal Persistence of IR Systems’ Performance

This paper presents the work of the CLOSE group, a team of students from the University of Padua, Italy, for the Conference and Labs of the Evaluation Forum (CLEF) LongEval LAB 2023 Task 1 [1]. Our work involved developing an Information Retrieval (IR) system that can handle changes in data over time while maintaining high performance. We first introduce the problem as stated by CLEF and then describe our system, explaining the different methodologies we implemented. We provide the results of our experiments and analyze them based on the choices we made regarding various techniques. Finally, we propose potential avenues for future improvement of our system.