UCIL: An Unsupervised Class Incremental Learning Approach for Sound Event Detection

TL;DR

This paper introduces an unsupervised class-incremental learning method for sound event detection, enabling models to adapt to new sound classes over time while maintaining performance on previously learned classes.

Contribution

It presents a novel unsupervised framework with distillation, sample selection, and exemplar update strategies tailored for incremental sound event detection.

Findings

Effective in integrating new sound classes without performance loss

Provides insights into continual learning methods for real-world audio

Evaluates on DCASE 2023 dataset demonstrating practical applicability

Abstract

This work explores class-incremental learning (CIL) for sound event detection (SED), advancing adaptability towards real-world scenarios. CIL's success in domains like computer vision inspired our SED-tailored method, addressing the unique challenges of diverse and complex audio environments. Our approach employs an independent unsupervised learning framework with a distillation loss function to integrate new sound classes while preserving the SED model consistency across incremental tasks. We further enhance this framework with a sample selection strategy for unlabeled data and a balanced exemplar update mechanism, ensuring varied and illustrative sound representations. Evaluating various continual learning methods on the DCASE 2023 Task 4 dataset, we find that our research offers insights into each method's applicability for real-world SED systems that can have newly added sound classes. The findings also delineate future directions of CIL in dynamic audio settings.