Self-Supervised Multi-View Learning for Disentangled Music Audio Representations

TL;DR

This paper introduces a self-supervised multi-view learning framework for music audio that effectively disentangles shared and private representations, improving the quality of learned audio features without labeled data.

Contribution

It presents a novel SSL method that explicitly separates shared and private audio features, addressing entanglement issues in previous approaches.

Findings

Effective disentanglement of shared and private representations demonstrated

Improved robustness and generalization in music audio tasks shown

Case study confirms the method's effectiveness in controlled settings

Abstract

Self-supervised learning (SSL) offers a powerful way to learn robust, generalizable representations without labeled data. In music, where labeled data is scarce, existing SSL methods typically use generated supervision and multi-view redundancy to create pretext tasks. However, these approaches often produce entangled representations and lose view-specific information. We propose a novel self-supervised multi-view learning framework for audio designed to incentivize separation between private and shared representation spaces. A case study on audio disentanglement in a controlled setting demonstrates the effectiveness of our method.