UvA-DARE (University of Amsterdam)· 2013· Preprint
Auto-Encoding Variational Bayes
- 15,629citations
- 2013year
Short summary
A novel stochastic variational inference algorithm enables efficient learning in directed probabilistic models with continuous latent variables and large datasets, even when posteriors are intractable.
AI-generated from the title and abstract; the full text is not read.
Key points
- Introduces a stochastic variational inference and learning algorithm for directed probabilistic models.
- The algorithm scales to large datasets and handles intractable posterior distributions.
- A reparameterization technique yields a lower bound estimator optimizable via stochastic gradient methods.
- Efficient posterior inference is achieved by fitting an approximate inference model to the intractable posterior.
AI-generated from the title and abstract; the full text is not read.
Abstract
How can we perform efficient inference and learning in directed probabilistic models, in the presence of continuous latent variables with intractable posterior distributions, and large datasets? We introduce a stochastic variational inference and learning algorithm that scales to large datasets and, under some mild differentiability conditions, even works in the intractable case. Our contributions are two-fold. First, we show that a reparameterization of the variational lower bound yields a lower bound estimator that can be straightforwardly optimized using standard stochastic gradient methods. Second, we show that for i.i.d. datasets with continuous latent variables per datapoint, posterior inference can be made especially efficient by fitting an approximate inference model (also called a recognition model) to the intractable posterior using the proposed lower bound estimator. Theoretical advantages are reflected in experimental results.
The authors' abstract, as published at the source. UvA-DARE (University of Amsterdam), 2013 · DOI ↗
Continue with a free account
Ask the paper: 3 free questions a day about this paper; save it, get its citation, new summaries every day for your field. Takeaways are Premium.
Continue free on the webSign in with Google or Apple; no card needed. You come back to this paper.
On your phone:
Field: Artificial Intelligence
Artificial IntelligenceComputer Science