Announcement_7

Starting 2026 with a good news, our paper Mitigating data replication in text-to-audio generative diffusion models through anti-memorization guidance co-written with Francisco Messina, Francesca Ronchini, Paolo Bestagini and Fabio Antonacci, has been accepted at the 2026 IEEE International Conference on Acoustics, Speech, and Signal Processing in Barcelona! In this paper, we explore how to use post-training techniques to avoid the generation of audio tracks memorized from the training set.

here you can find the preprint of the paper https://arxiv.org/abs/2505.07615 and also the code is fully available at https://polimi-ispl.github.io/anti-memorization-tta/