Get Started
Research questionCan stationary-point ELBO values be expressed as entropy sums across variational generative models?The ELBO is optimized in unsupervised latent-variable learning, yet its value at convergence can be difficult to interpret across different generative models. A common entropy-based form could clarify what stationary-point values represent.
Machine Learning
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.On the Equality of the ELBO to a Sum of Entropies at Stationary Points of LearningThe result concerns models with one set of latent variables and one set of observed variables, using exponential-family distributions and typically satisfying a parameterization criterion. It covers finite datasets, any stationary point including saddles, and well-behaved variational families; the stated class includes standard Gaussian variational autoencoders. The decomposition involves the variational entropy, the negative prior entropy, and the expected negative entropy of the observable distribution.research paper · Sep 3, 2026
Related questions
How can VAEs reparameterize latent distributions on manifolds with nontrivial topology?How can generative models capture kinetic dynamics on Lie groups from partial endpoint observations without flattening the manifold?When does changing the entropy in attention produce a genuinely different weighting profile rather than merely rescale temperature?When are Double Machine Learning estimators inadmissible under structure-agnostic models?
Home
Topics
Search
Library