The “LUSt: Learning a Universal Similarity Function” project has been funded under Horizon Europe and will run for two years (2025-2026). This grant supports a postdoctoral fellowship at CTU in Prague with Giorgos Tolias as an advisor and a collaboration with the University of Amsterdam and assistant professor Pascal Mettes. The project’s objective is to perform research on similarity learning based on the similarities of the constituent parts of multimedia items. The ultimate goal is creating a foundational model capable of accommodating all modalities — visual, audio, text — and supporting a broad range of similarity types, including uni-modal, cross-modal, and multi-modal scenarios.