ICASSP 2025

ICASSP25 opening ceremony

A decade of DCASE and incremental learning

We presented some of our recent and not so recent work at ICASSP 2025 in Hyderabad this year: domain-incremental learning for audio classification, and an attempt at class-incremental learning for SELD.

The not-so-recent category was the overview of DCASE challenge and workshops, which I was honored to present to a full room.

Annamaria presenting the paper at ICASSP2025

Links to the full papers are available on the Publications page.