# Evaluation of Latent Space Disentanglement in the Presence of Interdependent Attributes

arXiv: https://arxiv.org/abs/2110.05587 · HTML: https://mlsanity.com/paper/2110.05587/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 10 of its 16 references that are in the map.
- First posted on arXiv 11 Oct 2021.

## Builds on (most cited first, 10 of 10)

- [Auto-Encoding Variational Bayes](https://mlsanity.com/paper/1312.6114/) (2013)
- [Variational Inference of Disentangled Latent Concepts from Unlabeled Observations](https://mlsanity.com/paper/1711.00848/) (2017)
- [Neural Audio Synthesis of Musical Notes with WaveNet Autoencoders](https://mlsanity.com/paper/1704.01279/) (2017)
- [Isolating Sources of Disentanglement in Variational Autoencoders](https://mlsanity.com/paper/1802.04942/) (2018)
- [Learning Deep Disentangled Embeddings with the F-Statistic Loss](https://mlsanity.com/paper/1802.05312/) (2018)
- [Theory and Evaluation Metrics for Learning Disentangled Representations](https://mlsanity.com/paper/1908.09961/) (2019)
- [Music FaderNets: Controllable Music Generation Based On High-Level Features via Low-Level Feature Modelling](https://mlsanity.com/paper/2007.15474/) (2020)
- [MIDI-VAE: Modeling Dynamics and Instrumentation of Music with Applications to Style Transfer](https://mlsanity.com/paper/1809.07600/) (2018)
- [Musical Composition Style Transfer via Disentangled Timbre Representations](https://mlsanity.com/paper/1905.13567/) (2019)
- [Is Disentanglement enough? On Latent Representations for Controllable Music Generation](https://mlsanity.com/paper/2108.01450/) (2021)

## Built on by (most cited first, 1 of 1)

- [Latte: Cross-framework Python Package for Evaluation of Latent-Based Generative Models](https://mlsanity.com/paper/2112.10638/) (2021)

## Abstract (arXiv)

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.