SynIRgy: Synthetic Data and Simulation Synergy for Information Retrieval

Manel Slokom, Alejandro Bellogín, Mayank Singh, Andrea Barraza-Urbina

Abstract: The increasing complexity of information access systems necessitates innovative evaluation methods beyond real-world user data, which is often biased, incomplete, or difficult to obtain due to privacy concerns. Simulation and synthetic data methods provide controlled and reproducible environments for training, evaluating, refining, and benchmarking algorithms in information retrieval and personalization.