Unpicking Data at the Seams: Understanding Disentanglement in VAEs
- CMLCoGe

Disentanglement, or identifying statistically independent factors of the data, is relevant to much of machine learning, from controlled data generation and robust classification to efficient encoding and improving our understanding of the data itself. Disentanglement arises in several generative paradigms including Variational Autoencoders (VAEs), Generative Adversarial Networks and diffusion models. Prior work takes a step towards understanding disentanglement in VAEs by showing diagonal posterior covariance matrices promote orthogonality between columns of the decoder's Jacobian. Building on this, we close the gap in our understanding of disentanglement by showing how if follows from such orthogonality and equates to factoring the data distribution into statistically independent components.
View on arXiv@article{allen2025_2410.22559, title={ Unpicking Data at the Seams: Understanding Disentanglement in VAEs }, author={ Carl Allen }, journal={arXiv preprint arXiv:2410.22559}, year={ 2025 } }