Creation of Autonomous Artificial Intelligent Agents using Novelty Search method of fitness function optimization
The medium maze results for NS agent
Abstract
Search for Novelty is an universal method of biological life evolution through introduction of the random beneficial mutations to the genetic code of the organism. In this work we describe experiment of applying Novelty Search method of fitness function optimization combined with Neuro-Evolution of Augmenting Topologies (NEAT) algorithm to produce Autonomous Artificial Intelligent Agents capable to solve spatial navigation task in complex maze environment. The AI Agents produced by the neuro-evolution method described by the NEAT algorithm evolve through gradual complexification of their internal neural network by augmenting its topologies.\n\nFinally, we consider how to apply studied optimization methods and evolutionary algorithms to create ensembles of compact modular AI systems trained using vocabulary of terms describing real world settings. And how to control such ensembles by specialized supervisors knowledgeable about the task and the operating environment.\n\nWe provide the complete source code for implementing the NEAT algorithm, including Novelty Search and Objective-Based optimization methods, in the GO programming language.
Iaroslav Omelianenko
CTO/Research Director at NewGround LLC
My research interests include genetic algorithms/neuroevolution, synthetic cognitive systems, smart environment, and cooperative robotics.