Posterior Collapse as a Phase Transition in Variational Autoencoders

📅 2025-10-01
📈 Citations: 0
Influential: 0
📄 PDF

career value

175K/year
🤖 AI Summary
Posterior collapse in variational autoencoders (VAEs) is commonly attributed to optimization failure, obscuring its fundamental origins. Method: We formulate posterior collapse as a phase transition phenomenon jointly driven by data structure and model hyperparameters, drawing on statistical physics principles; we analyze the stability of the KL divergence between the approximate posterior and prior via variational inference, and identify a critical point where this divergence exhibits discontinuity—marking the phase boundary between collapse and effective latent inference. Contribution/Results: Our framework systematically characterizes the phase transition behavior on both synthetic and real-world datasets, precisely identifying hyperparameter thresholds that prevent collapse. This work establishes a novel theoretical foundation for understanding trainability, representational capacity, and latent-space geometry in deep generative models, offering actionable guidance for VAE design and hyperparameter selection.

Technology Category

Application Category

📝 Abstract
We investigate the phenomenon of posterior collapse in variational autoencoders (VAEs) from the perspective of statistical physics, and reveal that it constitutes a phase transition governed jointly by data structure and model hyper-parameters. By analyzing the stability of the trivial solution associated with posterior collapse, we identify a critical hyper-parameter threshold. This critical boundary, separating meaningful latent inference from collapse, is characterized by a discontinuity in the KL divergence between the approximate posterior and the prior distribution. We validate this critical behavior on both synthetic and real-world datasets, confirming the existence of a phase transition. Our results demonstrate that posterior collapse is not merely an optimization failure, but rather an emerging phase transition arising from the interplay between data structure and variational constraints. This perspective offers new insights into the trainability and representational capacity of deep generative models.
Problem

Research questions and friction points this paper is trying to address.

Posterior collapse constitutes a phase transition in VAEs
Critical hyper-parameter threshold separates meaningful inference from collapse
Phase transition arises from data structure and variational constraints interplay
Innovation

Methods, ideas, or system contributions that make the work stand out.

Identified critical hyper-parameter threshold for collapse
Characterized phase transition via KL divergence discontinuity
Revealed collapse as data-structure interaction phase transition
🔎 Similar Papers
Z
Zhen Li
Department of Complexity Science and Engineering, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, 277-8561, Chiba, Japan
F
Fan Zhang
Department of Human and Engineered Environmental Studies, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, 277-8561, Chiba, Japan
Z
Zheng Zhang
Department of Human and Engineered Environmental Studies, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, 277-8561, Chiba, Japan
Y
Yu Chen
Department of Human and Engineered Environmental Studies, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, 277-8561, Chiba, Japan