Automatic Differentiation Variational Inference
This addresses the challenge for scientists and practitioners in probabilistic modeling by automating inference, reducing mathematical and computational barriers, though it builds on existing variational methods.
The paper tackles the bottleneck of fitting complex probabilistic models to large data by developing automatic differentiation variational inference (ADVI), which automatically derives efficient variational inference algorithms from models and datasets, enabling scientists to explore many models without manual derivations. It demonstrates ADVI across ten models and scales to millions of observations, integrating into Stan for practical use.
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 (ADVI). Using our method, the scientist only provides a probabilistic model and a dataset, nothing else. ADVI automatically derives an efficient variational inference algorithm, freeing the scientist to refine and explore many models. ADVI supports a broad class of models-no conjugacy assumptions are required. We study ADVI across ten different models and apply it to a dataset with millions of observations. ADVI is integrated into Stan, a probabilistic programming system; it is available for immediate use.