JANA: Jointly Amortized Neural Approximation of Complex Bayesian Models
Paper in proceeding, 2023

This work proposes “jointly amortized neural approximation” (JANA) of intractable likelihood functions and posterior densities arising in Bayesian surrogate modeling and simulation-based inference. We train three complementary networks in an end-to-end fashion: 1) a summary network to compress individual data points, sets, or time series into informative embedding vectors; 2) a posterior network to learn an amortized approximate posterior; and 3) a likelihood network to learn an amortized approximate likelihood. Their interaction opens a new route to amortized marginal likelihood and posterior predictive estimation - two important ingredients of Bayesian workflows that are often too expensive for standard methods. We benchmark the fidelity of JANA on a variety of simulation models against state-of-the-art Bayesian methods and propose a powerful and interpretable diagnostic for joint calibration. In addition, we investigate the ability of recurrent likelihood networks to emulate complex time series models without resorting to hand-crafted summary statistics.

Author

Stefan T. Radev

Heidelberg University

Marvin Schmitt

University of Stuttgart

Valentin Pratz

Heidelberg University

Umberto Picchini

Chalmers, Mathematical Sciences, Applied Mathematics and Statistics

Ullrich Köthe

Heidelberg University

Paul Christian Bürkner

University of Stuttgart

Proceedings of Machine Learning Research

26403498 (eISSN)

Vol. 216 1695-1706

39th Conference on Uncertainty in Artificial Intelligence, UAI 2023
Pittsburgh, USA,

Deep learning and likelihood-free Bayesian inference for stochastic modelling

Swedish Research Council (VR) (2019-03924), 2020-01-01 -- 2023-12-31.

Chalmers AI Research Centre (CHAIR), 2020-01-01 -- 2024-12-31.

Subject Categories

Probability Theory and Statistics

More information

Latest update

9/19/2024