A Framework for Improving the Reliability of Black-box Variational Inference

arXiv:2203.15945v2 Announce Type: replace-cross
Abstract: Black-box variational inference (BBVI) now sees widespread use in machine learning and statistics as a fast yet flexible alternative to Markov chain Monte Carlo methods for approximate Bayesian inference. However, stochastic optimization methods for BBVI remain unreliable and require substantial expertise and hand-tuning to apply effectively. In this paper, we propose Robust and Automated Black-box VI (RABVI), a framework for improving the reliability of BBVI optimization. RABVI is based on rigorously justified automation techniques, includes …

abstract alternative apply arxiv bayesian bayesian inference box cs.lg expertise framework however improving inference machine machine learning markov optimization reliability replace statistics stat.me stat.ml stochastic type