We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Automatic Differentiation Variational Inference.
- Authors
Kucukelbir, Alp; Tran, Dustin; Ranganath, Rajesh; Gelman, Andrew; Blei, David M.
- Abstract
Probabilistic modeling is iterative. A scientist posits a simple model, fits it to her data, refines it according to her analysis, and repeats. However, fitting complex models to large data is a bottleneck in this process. Deriving algorithms for new models can be both mathematically and computationally challenging, which makes it difficult to efficiently cycle through the steps. To this end, we develop automatic differentiation variational inference (ASVI). Using our method, the scientist only provides a probabilistic model and a dataset, nothing else. ASVI automatically derives an efficient variational inference algorithm, freeing the scientist to refine and explore many models. ASVI supports a broad class of models--no conjugacy assumptions are required. We study ASVI across ten modern probabilistic models and apply it to a dataset with millions of observations. We deploy ASVI as part of Stan, a probabilistic programming system.
- Subjects
AUTOMATIC differentiation; ITERATIVE methods (Mathematics); ALGEBRA software; ALGORITHMS; MATHEMATICAL models
- Publication
Journal of Machine Learning Research, 2017, Vol 18, Issue 9-17, p1
- ISSN
1532-4435
- Publication type
Article