Nov 20, 2016 · Our method, termed variational boosting, iteratively refines an existing variational approximation by solving a sequence of optimization ...

We propose a black-box variational inference method to approximate intractable distributions with an increasingly rich approximating class.

Abstract. We propose a black-box variational inference method to approximate intractable distributions with an increasingly rich approximating class.

We present a black-box variational inference (BBVI) method to approximate in- tractable posterior distributions with an increasingly rich approximating ...

This work iteratively refines an existing variational approximation by solving a sequence of optimization problems, allowing the practitioner to trade ...

Our method, variational boosting, iteratively refines an existing variational approximation by solving a sequence of optimization problems, allowing a trade-off ...

Our method, variational boosting, iteratively refines an existing variational approximation by solving a sequence of optimization problems, allowing a trade-off ...

Introducing a new component requires initialization of com- ponent parameters. When our component distributions are mixtures of Gaussians, we found that the ...

Dive into the research topics of 'Variational boosting: Iteratively refining posterior approximations'. Together they form a unique fingerprint. Sort by; Weight ...

Variational Boosting: Iteratively Refining Posterior Approximations. A. Miller, N. Foti, and R. Adams. ICML, volume 70 of Proceedings of Machine Learning ...