DeepAI AI Chat
Log In Sign Up

Information Theoretically Aided Reinforcement Learning for Embodied Agents

by   Guido Montufar, et al.
Max Planck Society

Reinforcement learning for embodied agents is a challenging problem. The accumulated reward to be optimized is often a very rugged function, and gradient methods are impaired by many local optimizers. We demonstrate, in an experimental setting, that incorporating an intrinsic reward can smoothen the optimization landscape while preserving the global optimizers of interest. We show that policy gradient optimization for locomotion in a complex morphology is significantly improved when supplementing the extrinsic reward by an intrinsic reward defined in terms of the mutual information of time consecutive sensor readings.


page 5

page 9

page 13

page 15

page 16

page 17


Influence-Based Multi-Agent Exploration

Intrinsically motivated reinforcement learning aims to address the explo...

Emergence of Locomotion Behaviours in Rich Environments

The reinforcement learning paradigm allows, in principle, for complex be...

Cliff Diving: Exploring Reward Surfaces in Reinforcement Learning Environments

Visualizing optimization landscapes has led to many fundamental insights...

Learning from Outside the Viability Kernel: Why we Should Build Robots that can Fall with Grace

Despite impressive results using reinforcement learning to solve complex...

ES Is More Than Just a Traditional Finite-Difference Approximator

An evolution strategy (ES) variant recently attracted significant attent...

High-Dimensional Control Using Generalized Auxiliary Tasks

A long-standing challenge in reinforcement learning is the design of fun...

Evaluating Agents without Rewards

Reinforcement learning has enabled agents to solve challenging tasks in ...