Personalised Similarity Search

Day - Time: 17 September 2025, h.15:00

Place: Area della Ricerca CNR di Pisa - Room: C-29 (Aula Faedo)

Speakers

• Matús Sikyna (Faculty of Informatics, Masaryk University, Brno, Czech Republic)

Referent

Nicola Messina

Abstract

The importance of similarity search has become prominent in the fast-evolving vector databases, which apply content embedding techniques on complex data to produce and manage large collections of high-dimensional vectors. Processing of such data is only possible by using a similarity function for storage, structure, and retrieval. However, if multiple users access the collection, their views on similarity can differ as similarity, in general, is subjective and context-dependent. In this seminar, we elaborate on the problem of a similarity search engine implementation, where users use a common index but search with personalised views of similarity, implemented by a possibly different similarity model. We start with an introduction into the SISAP 2024 and 2025 papers on this topic with regards to the range search: theoretical foundations behind the approach, the role of the relevance feedback and metric learning, practical scenarios of usage. Then, we discuss the current research of the personalised similarity search for the kNN search.