# Relevance Factor VAE: Learning and Identifying Disentangled Factors

arXiv: https://arxiv.org/abs/1902.01568 · HTML: https://mlsanity.com/paper/1902.01568/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Feb 2019.

## Built on by (most cited first, 20 of 20)

- [Disentangled Representation Learning](https://mlsanity.com/paper/2211.11695/) (2022)
- [EditVAE: Unsupervised Part-Aware Controllable 3D Point Cloud Shape Generation](https://mlsanity.com/paper/2110.06679/) (2021)
- [Benchmarks, Algorithms, and Metrics for Hierarchical Disentanglement](https://mlsanity.com/paper/2102.05185/) (2021)
- [Graph-based Unsupervised Disentangled Representation Learning via Multimodal Large Language Models](https://mlsanity.com/paper/2407.18999/) (2024)
- [Review of Disentanglement Approaches for Medical Applications -- Towards Solving the Gordian Knot of Generative Models in Healthcare](https://mlsanity.com/paper/2203.11132/) (2022)
- [Closed-Loop Unsupervised Representation Disentanglement with β-VAE Distillation and Diffusion Probabilistic Feedback](https://mlsanity.com/paper/2402.02346/) (2024)
- [There Was Never a Bottleneck in Concept Bottleneck Models](https://mlsanity.com/paper/2506.04877/) (2025)
- [Modular Machine Learning: An Indispensable Path towards New-Generation Large Language Models](https://mlsanity.com/paper/2504.20020/) (2025)
- [Analyzing Generative Models by Manifold Entropic Metrics](https://mlsanity.com/paper/2410.19426/) (2024)
- [DAReN: A Collaborative Approach Towards Reasoning And Disentangling](https://mlsanity.com/paper/2109.13156/) (2021)
- [Break The Spell Of Total Correlation In betaTCVAE](https://mlsanity.com/paper/2210.08794/) (2022)
- [Transferring disentangled representations: bridging the gap between synthetic and real images](https://mlsanity.com/paper/2409.18017/) (2024)
- [RENs: Relevance Encoding Networks](https://mlsanity.com/paper/2205.13061/) (2022)
- [Rethinking Disentanglement under Dependent Factors of Variation](https://mlsanity.com/paper/2408.07016/) (2024)
- [Independence Constrained Disentangled Representation Learning from Epistemological Perspective](https://mlsanity.com/paper/2409.02672/) (2024)
- [PH-VAE: A Polynomial Hierarchical Variational Autoencoder Towards Disentangled Representation Learning](https://mlsanity.com/paper/2502.02856/) (2025)
- [Unveiling and Steering Connectome Organization with Interpretable Latent Variables](https://mlsanity.com/paper/2505.13011/) (2025)
- [Disentanglement of Sources in a Multi-Stream Variational Autoencoder](https://mlsanity.com/paper/2510.15669/) (2025)
- [Causal Representation Learning on High-Dimensional Data: Benchmarks, Reproducibility, and Evaluation Metrics](https://mlsanity.com/paper/2603.17405/) (2026)
- [Deep Attention Reweighting: Post-Hoc Attention-Based Feature Aggregation in CNNs for Disentangling Core and Spurious Features under Spurious Correlations](https://mlsanity.com/paper/2605.20732/) (2026)

## Abstract (arXiv)

We propose a novel VAE-based deep auto-encoder model that can learn disentangled latent representations in a fully unsupervised manner, endowed with the ability to identify all meaningful sources of variation and their cardinality. Our model, dubbed Relevance-Factor-VAE, leverages the total correlation (TC) in the latent space to achieve the disentanglement goal, but also addresses the key issue of existing approaches which cannot distinguish between meaningful and nuisance factors of latent variation, often the source of considerable degradation in disentanglement performance. We tackle this issue by introducing the so-called relevance indicator variables that can be automatically learned from data, together with the VAE parameters. Our model effectively focuses the TC loss onto the relevant factors only by tolerating large prior KL divergences, a desideratum justified by our semi-parametric theoretical analysis. Using a suite of disentanglement metrics, including a newly proposed one, as well as qualitative evidence, we demonstrate that our model outperforms existing methods across several challenging benchmark datasets.