Two Methods for Wild Variational Inference

Qiang Liu, Yihao Feng

Abstract: Variational inference provides a powerful tool for approximate probabilistic inference on complex, structured models. Typical variational inference methods, however, require to use inference networks with computationally tractable probability density functions. This largely limits the design and implementation of variational inference methods. We consider wild variational inference methods that do not require tractable density functions on the inference networks, and hence can be applied in more challenging cases. As an example of application, we treat stochastic gradient Langevin dynamics (SGLD) as an inference network, and use our methods to automatically adjust the step sizes of SGLD to maximize its convergence speed, significantly outperforming the hand-designed step size schemes.

Conflicts: cs.dartmouth.edu

Keywords: Theory