Infinite-dimensional gradient-based descent for alpha-divergence minimisation

05/20/2020
by   Kamélia Daudel, et al.
0

This paper introduces the (α, Γ)-descent, an iterative algorithm which operates on measures and performs α-divergence minimisation in a Bayesian framework. This gradient-based procedure extends the commonly-used variational approximation by adding a prior on the variational parameters in the form of a measure. We prove that for a rich family of functions Γ, this algorithm leads at each step to a systematic decrease in the α-divergence. Our framework recovers the Entropic Mirror Descent (MD) algorithm with improved O(1/N) convergence results and provides an alternative to the Entropic MD that we call the Power descent and for which we prove convergence to an optimum. Moreover, the (α, Γ)-descent allows to optimise the mixture weights of any given mixture model without any information on the underlying distribution of the variational parameters. This renders our method compatible with many choices of parameters updates and applicable to a wide range of Machine Learning tasks. We demonstrate empirically on both toy and real-world examples the benefit of using the Power descent and going beyond the Entropic MD framework, which fails as the dimension grows.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
09/26/2019

The f-Divergence Expectation Iteration Scheme

This paper introduces the f-EI(ϕ) algorithm, a novel iterative algorithm...
research
06/09/2021

Mixture weights optimisation for Alpha-Divergence Variational Inference

This paper focuses on α-divergence minimisation methods for Variational ...
research
03/09/2021

Monotonic Alpha-divergence Minimisation

In this paper, we introduce a novel iterative algorithm which carries ou...
research
10/02/2022

Improved Stein Variational Gradient Descent with Importance Weights

Stein Variational Gradient Descent (SVGD) is a popular sampling algorith...
research
07/17/2022

Mean field Variational Inference via Wasserstein Gradient Flow

Variational inference (VI) provides an appealing alternative to traditio...
research
10/16/2012

Sparse Q-learning with Mirror Descent

This paper explores a new framework for reinforcement learning based on ...

Please sign up or login with your details

Forgot password? Click here to reset