Title: Computational Red Teaming: Can Evolution and Learning Augment Human Behaviour?

Hussein A. Abbass

Professor, University of New South Wales,

Australia.

 

Abstract: Computational red teaming (CRT) is a field that attempts to create a form of artificial intelligence (AI) whereby intelligence is measured as the ability of a computer environment to challenge humans. Creating a computer environment to challenge human thinking, design or planning is far from being a trivial task. Artificial Neural Networks and Evolution (Neuroevolution) are fundamental algorithms for creating intelligence in CRT. Neuroevolution has many advantages including the natural distributed architecture of a neural network enables calculations to be carried out in parallel; the interaction between evolution and learning occurs with minimum bias; and prior knowledge can be encoded through simple procedures, with evolution providing natural mechanisms to either maintain these knowledge if they were useful or replacing them if they were not. In this talk, I demonstrate a series of studies designed to compare human performance against computer performance in a synthetic setting. Here, humans and computers are compared to distil similarities and differences of behaviours. We demonstrate that CRT overcomes human bias. It is able to generate behaviours that are sometimes similar to human behaviours and in many cases other behaviours that are very different from human behaviours albeit being plausible ones. The talk will conclude with the challenges in CRT.

 

Biography: Prof. Dr. Hussein Abbass is a Professor of Information Technology at the University of New South Wales, Canberra Campus, Australia. He is a Fellow of the Operational Research Society (UK), a fellow of the Australian Computer Society, and a Senior Member of the IEEE.
Prof Abbass is an Associate Editor of the IEEE Transactions on Evolutionary Computation, the IEEE Computational Intelligence Magazine, the American Institute of Mathematical Sciences Journal of Industrial and Management Optimization, and three other journals. Prof Abbass's work integrates cognitive science, operations research and artificial intelligence. He was the General Chair of the 2012 IEEE World Congress on Computational Intelligence, to be held in Brisbane June 2012, and is the Premier and Largest event by the IEEE Computational Intelligence Society.