Tip-of-the-Tongue Retrieval

Tip-of-the-tongue (TOT) retrieval addresses the challenge of helping users find specific information like movies or songs when they cannot recall specific identifiers like titles. People express information needs during TOT states using combinations of content descriptions, contextual memories, and comparative information. The work includes developing synthetic query generation methods using large language models and human elicitation techniques to overcome data collection challenges. TREC evaluation tracks have been established to systematically evaluate TOT retrieval systems across domains like movies, celebrities, and landmarks.

Publications

Y. He, T.-E. Kim, F. Diaz, J. Arguello, B. Mitra
SIGIR 2025
J. Arguello, S. Bhargav, F. Diaz, T.-E. Kim, Y. He, E. Kanoulas, B. Mitra
TREC 2024
J. Arguello, S. Bhargav, F. Diaz, E. Kanoulas, B. Mitra
TREC 2023
J. Arguello, A. Ferguson, E. Fine, B. Mitra, H. Zamani, F. Diaz
CHIIR 2021