Synthesizing neural networks for learning in games

Authors:
Robert G. Price;Scott D. Goodwin
Affiliations:
University of Windsor, Windsor, Ontario, Canada;University of Windsor, Windsor, Ontario, Canada
Venue:
Future Play '08 Proceedings of the 2008 Conference on Future Play: Research, Play, Share
Year:
2008

Citing 4
Cited 1

Multitask Learning

Machine Learning - Special issue on inductive transfer
Evolving neural networks through augmenting topologies

Evolutionary Computation
Real-time evolution of neural networks in the NERO video game

AAAI'06 proceedings of the 21st national conference on Artificial intelligence - Volume 2
Real-time neuroevolution in the NERO video game

IEEE Transactions on Evolutionary Computation

Universal design for learning (UDL1) and living (UDL2) in virtual reality-based treatments for children with autism

Proceedings of the 6th International Conference on Rehabilitation Engineering & Assistive Technology

Quantified Score

Hi-index	0.00

Visualization

Abstract

Many games have started to employ learning techniques to make them more realistic or interesting. Usually though, this learning is done before the game ships, and it cannot compensate for any exploits a character discovers. One reason for this is that game publishers do not want to risk having the non-player characters making odd decisions in games that learn. In this paper we propose an approach that can be used to quickly jump-start the learning process in a game that uses a neural network to learn. We create different environments that might occur in a game, analyse them and come up with a starting point that allows the agents to quickly be able to accomplish their goals, which in our case is navigating through a random board.