Deep Generative Models for Decision-Making and Control

TL;DR

This paper explores the integration of deep generative models with decision-making and control, analyzing their limitations in reinforcement learning and proposing solutions using advanced inference techniques.

Contribution

It investigates the empirical shortcomings of deep model-based RL and introduces novel planning strategies inspired by generative modeling inference techniques.

Findings

Identified key limitations of current model-based RL methods.

Proposed new planning strategies using generative inference techniques.

Enhanced understanding of how generative models can improve decision-making.

Abstract

Deep model-based reinforcement learning methods offer a conceptually simple approach to the decision-making and control problem: use learning for the purpose of estimating an approximate dynamics model, and offload the rest of the work to classical trajectory optimization. However, this combination has a number of empirical shortcomings, limiting the usefulness of model-based methods in practice. The dual purpose of this thesis is to study the reasons for these shortcomings and to propose solutions for the uncovered problems. Along the way, we highlight how inference techniques from the contemporary generative modeling toolbox, including beam search, classifier-guided sampling, and image inpainting, can be reinterpreted as viable planning strategies for reinforcement learning problems.