Auto-Encoding Variational Bayes Research Publication
Diederik P. Kingma and Max Welling submitted 'Auto-Encoding Variational Bayes' on December 20, 2013. The paper presented a stochastic variational inference and learning method for directed probabilistic models with continuous latent variables, including a reparameterized variational-lower-bound estimator that can be optimized with standard stochastic gradient methods.
Historical classification
This verified record belongs to LINEAiGE's artificial-intelligence history layer beginning in 1956.
Why it matters
A documented generative latent-variable modeling milestone that combines variational inference with a learned approximate inference, or recognition, model and a reparameterization-based optimization method.
People, organizations & technologies
PEOPLE
Diederik P. Kingma, Max Welling
TECHNOLOGIES
Auto-Encoding Variational Bayes, Variational inference, Latent-variable models
Evidence & verification
STATE
verified
CONFIDENCE
high
LAST VERIFIED
2026-09-10T09:04:42+08:00
The original arXiv record establishes the December 20, 2013 first-submission date, authors, and the paper's stochastic variational inference, reparameterized lower-bound estimator, and approximate-inference-model claims. This record uses that first-submission date strictly as a research-publication milestone. It does not infer a software or product release date, institutional affiliation, geography, ancestry, later influence, causality, or relationships, and it does not retroactively assign terminology beyond what is needed to identify the documented method.
Sources
Auto-Encoding Variational Bayes
arXiv
PUBLISHED
2013-12-20
Trace the chronology
Record identity
LINEAiGE IDauto-encoding-variational-bayes-2013