Experiential Robot Learning with Accelerated Neuroevolution

TL;DR

This paper introduces the Accelerated Neuroevolution algorithm, an alternative to gradient-based methods, for physical robot learning tasks, demonstrating its effectiveness on simulated and real robotic tasks with promising results.

Contribution

The paper presents a novel neuroevolution algorithm tailored for experiential robot learning, capable of efficiently training robots without requiring differentiability.

Findings

Successfully trained agents on Flappy Bird simulation

Achieved task completion on a physical NAO robot

Demonstrated low-generation learning efficiency

Abstract

Derivative-based optimization techniques such as Stochastic Gradient Descent has been wildly successful in training deep neural networks. However, it has constraints such as end-to-end network differentiability. As an alternative, we present the Accelerated Neuroevolution algorithm. The new algorithm is aimed towards physical robotic learning tasks following the Experiential Robot Learning method. We test our algorithm first on a simulated task of playing the game Flappy Bird, then on a physical NAO robot in a static Object Centering task. The agents successfully navigate the given tasks, in a relatively low number of generations. Based on our results, we propose to use the algorithm in more complex tasks.