Learning Representations for New Sound Classes With Continual Self-Supervised Learning
This addresses the challenge of updating sound recognition models with new classes in a continual learning context, where labeled data is scarce, though it is incremental as it builds on existing self-supervised and continual learning techniques.
The paper tackles the problem of continually incorporating new sound classes into a recognition system without relying on labeled data, proposing a self-supervised representation learning framework that achieves similar performance to distillation-based methods while being robust to forgetting.
In this paper, we work on a sound recognition system that continually incorporates new sound classes. Our main goal is to develop a framework where the model can be updated without relying on labeled data. For this purpose, we propose adopting representation learning, where an encoder is trained using unlabeled data. This learning framework enables the study and implementation of a practically relevant use case where only a small amount of the labels is available in a continual learning context. We also make the empirical observation that a similarity-based representation learning method within this framework is robust to forgetting even if no explicit mechanism against forgetting is employed. We show that this approach obtains similar performance compared to several distillation-based continual learning methods when employed on self-supervised representation learning methods.