The Option-Critic Architecture

TL;DR

The paper introduces the option-critic architecture, enabling autonomous learning of temporally extended actions in reinforcement learning through policy gradient methods, improving scalability and efficiency.

Contribution

It develops a novel framework that learns options' internal policies and termination conditions simultaneously without extra rewards or subgoals.

Findings

Effective in discrete environments

Demonstrates flexibility in continuous settings

Improves learning efficiency

Abstract

Temporal abstraction is key to scaling up learning and planning in reinforcement learning. While planning with temporally extended actions is well understood, creating such abstractions autonomously from data has remained challenging. We tackle this problem in the framework of options [Sutton, Precup & Singh, 1999; Precup, 2000]. We derive policy gradient theorems for options and propose a new option-critic architecture capable of learning both the internal policies and the termination conditions of options, in tandem with the policy over options, and without the need to provide any additional rewards or subgoals. Experimental results in both discrete and continuous environments showcase the flexibility and efficiency of the framework.