Training-free k-reciprocal re-ranking improves state-of-the-art musical version identification
Under review
Paper, code and data: coming soon.
PhD student, ADASP team, LTCI, Télécom Paris
I work on machine learning for music, specifically musical version identification (also called cover song identification): finding every recording of the same underlying work in a large audio collection. My PhD runs under a CIFRE agreement with Claimy, a Paris music-tech company working on royalty recovery.
Under review
Paper, code and data: coming soon.