Timezone: »

Robust, Accurate Stochastic Optimization for Variational Inference
Akash Kumar Dhaka · Alejandro Catalina · Michael Andersen · Måns Magnusson · Jonathan Huggins · Aki Vehtari

Thu Dec 10 09:00 AM -- 11:00 AM (PST) @ Poster Session 5 #1477

We examine the accuracy of black box variational posterior approximations for parametric models in a probabilistic programming context. The performance of these approximations depends on (1) how well the variational family approximates the true posterior distribution, (2) the choice of divergence, and (3) the optimization of the variational objective. We show that even when the true variational family is used, high-dimensional posteriors can be very poorly approximated using common stochastic gradient descent (SGD) optimizers. Motivated by recent theory, we propose a simple and parallel way to improve SGD estimates for variational inference. The approach is theoretically motivated and comes with a diagnostic for convergence and a novel stopping rule, which is robust to noisy objective functions evaluations. We show empirically, the new workflow works well on a diverse set of models and datasets, or warns if the stochastic optimization fails or if the used variational distribution is not good.

Author Information

Akash Kumar Dhaka (Aalto University)
Alejandro Catalina (Aalto University)
Michael Andersen (Technical University of Denmark)
Måns Magnusson (Aalto University)
Jonathan Huggins (Boston University)
Aki Vehtari (Aalto University)

More from the Same Authors