Evaluation of Latent Space Disentanglement in the Presence of Interdependent Attributes

Abstract

Controllable music generation with deep generative models has become increasingly reliant on disentanglement learning techniques. However, current disentanglement metrics, such as mutual information gap (MIG), are often inadequate and misleading when used for evaluating latent representations in the presence of interdependent semantic attributes often encountered in real-world music datasets. In this work, we propose a dependency-aware information metric as a drop-in replacement for MIG that accounts for the inherent relationship between semantic attributes.

Submitted to the Late-Breaking Demo Session of the 22nd International Society for Music Information Retrieval Conference