Home
- » Keynote speaker

Keynote speaker

Professor Hussein A. Abbass

University of New South Wales

http://www.cs.adfa.edu.au/~abbass/

Speaker bio: Hussein Abbass is currently a Professor and Chair of Information Technology at the School of Information Technology and Electrical Engineering, University of New South Wales, the Australian Defence Force Academy in Canberra, Australia.
He is the Director of the University Defence and Security Applications Research Centre and the Director of the Artificial Life and Adaptive Robotics Laboratory. He is the Chair of the Australian Computer Society National Committee on Complex Systems, the chair of the IEEE-CIS task force on Artificial Life and Complex Adaptive Systems, and a Chief Investigator on the Australian Research Council (ARC) Centre for Complex Systems (ACCS). He holds an Advisory Professor at Vietnam National University, Ho-Chi Minh city, and held visiting positions at Imperial College London and University of Illinois. He is on the editorial board for two journals IJICC and IJASS. His main research interests include evolutionary games, learning (data mining) and optimization, ensemble learning, and multi-agent systems. He has 170+ refereed publications and his research is funded by the Australian Research Council (ARC), Eurocontrol, and other government organisations and industry.

Abstract:

Multi-objective search normally results in a set of efficient solutions called the non-dominated set. Research has focused on inventing new algorithms and heuristics to find this set - with a myriad of evolutionary-based heuristics invented on this topic alone. However, many practitioners wonder: why do we need to generate a set of efficient solutions when at the end of the day, any traditional decision making process would require a single solution? So why do we need to generate many when we only need one?

In this talk, I argue with evidence that there are many application areas in intelligent and complex systems where all non-dominated solutions are needed. This will be demonstrated through four different application areas. The first is concerned with the advantages of combining the non-dominated set into an ensemble of learning machines [4]. The second demonstrates the use of non-dominance as a measure for complexity in evolutionary robotics and embodied cognition [3]. The third uses the non-dominated set to explore the fitness landscape of conflict in wargaming [2]. The fourth demonstrates the use of the non-dominated set for risk assessment and conflict detection in air traffic management [1].

References:

[1] Alam S., Shafi K., Abbass H.A. and Barlow M., An Ensemble Approach for Conflict Detection in Free Flight by Data Mining, Transportation Research Part C. In Press.

[2] Yang A., Abbass H.A., and Sarker H (2006). Characterizing Warfare in Red Teaming, IEEE Transactions on Systems, Man, Cybernetics, Part B, 36(2), pp. 268-285.

[3] Teo J. and Abbass H.A. (2005) Multi-objectivity and Complexity in Embodied Cognition. IEEE Transactions on Evolutionary Computation, 9(4), pp. 337-360.

[4] Abbass H.A. (2003) Pareto Neuro-Ensemble. The 16th Australian Joint Conference on Artificial Intelligence (AI'03), Perth, Lecture Notes in Artificial Intelligence LNAI 2903, Springer-Verlag, pp. 554-566.