Multi-Level Variational Autoencoder: Learning Disentangled Representations from Grouped Observations

TL;DR

The paper introduces ML-VAE, a deep probabilistic model that learns disentangled, semantically meaningful representations from grouped data, enabling manipulation and generalization to unseen groups with minimal supervision.

Contribution

ML-VAE is a novel deep probabilistic model that disentangles representations at both group and observation levels, improving interpretability and generalization.

Findings

Learns semantically meaningful disentangled representations

Enables manipulation of latent factors

Generalizes to unseen groups

Abstract

We would like to learn a representation of the data which decomposes an observation into factors of variation which we can independently control. Specifically, we want to use minimal supervision to learn a latent representation that reflects the semantics behind a specific grouping of the data, where within a group the samples share a common factor of variation. For example, consider a collection of face images grouped by identity. We wish to anchor the semantics of the grouping into a relevant and disentangled representation that we can easily exploit. However, existing deep probabilistic models often assume that the observations are independent and identically distributed. We present the Multi-Level Variational Autoencoder (ML-VAE), a new deep probabilistic model for learning a disentangled representation of a set of grouped observations. The ML-VAE separates the latent representation into semantically meaningful parts by working both at the group level and the observation level, while retaining efficient test-time inference. Quantitative and qualitative evaluations show that the ML-VAE model (i) learns a semantically meaningful disentanglement of grouped data, (ii) enables manipulation of the latent representation, and (iii) generalises to unseen groups.