0

Latent Spectroscopy: Posterior Collapse as a Feature

We show that, in linear Gaussian VAEs, posterior collapse is a form of latent feature selection. Feature importance is set by each latent coordinate's contribution to reconstruction, itself given by the corresponding PCA eigenvalue.

Open
Preview
Year
2026
ArXiv
arxiv.org/abs/2605.22691
Hosting
Excerpt onlyCC-BY-NC-4.0

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2605.22691CC-BY-NC-4.0
TL;DR
Semantic Scholar
Attribution policy →

Abstract

We show that, in linear Gaussian VAEs, posterior collapse is a form of latent feature selection. Feature importance is set by each latent coordinate's contribution to reconstruction, itself given by the corresponding PCA eigenvalue. Varying the regularizer strength β reveals a ranked spectrum of collapse events, with thresholds set by the utility/PCA spectrum. A mode-by-mode analysis identifies the scale-invariant signal fraction as an order parameter obeying Landau scaling near collapse. WorldClim experiments confirm both the utility-threshold calibration and the predicted near-collapse scaling.