Alan Blair's Research Interests

  • Mobile robots

    • I work as a program leader within the ARC Centre of Excellence for Autonomous Systems, where I am participating in a number of projects including feature-based SLAM for underwater robots, reinforcment learning for flying and walking robots, and applications of information geometry to decentralized data fusion for robot navigation and planning. I have also had some involvement with the F180 League and the Sony Legged League of Robocup, as well as Robocup Junior.

  • Co-evolutionary learning

    •  What factors contribute to the success or failure of a co-evolutionary learning system? It seems that certain features of the task domain and training environment may enhance the learning process by preventing collusive suboptimal equilibria in the meta-game of learning [5]. I have studied these issues in the context of backgammon [4], other interactive games [18], simulated robotics [16] and computer-aided education [20], and am now experimenting with massively parallel models of cognition in the form of a society of interacting agents.

  • Dynamic language processing

    •  Language tasks can be performed by either symbolic systems or dynamical systems. Each brings its own particular bias to the task, with regard to learning and induction. A better understanding of these biases may provide valuable insights into the structure of human language and the way it is processed. I have tried to take a few small steps in this direction by studying connectionist approaches to computational phonology [2], analysing the dynamical behaviour of recurrent neural networks trained to induce formal languages from examples [1,6], developing new architectures for language processing tasks [30], and studying the evolution of communication [24].

    Journal Articles

     [1] S. Chalup & A.D. Blair, 2003. Incremental training of first order recurrent neural networks to predict a context-sensitive language, Neural Networks 16, 955-972.

    [2] A.D. Blair & J. Ingram, 2003. Learning to predict the phonological structure of English loanwords in Japanese, Applied Intelligence 19, 101-108.

    [3] M. Boden & A.D. Blair, 2003. Learning the dynamics of embedded clauses, Applied Intelligence 19, 51-63.

    [4] J.B. Pollack & A.D. Blair, 1998. Co-evolution in the successful learning of Backgammon strategy, Machine Learning 32, 225-240.

    [5] A.D. Blair & J.B. Pollack, 1997. What makes a good co-evolutionary learning environment? Australian Journal of Intelligent Information Processing Systems 4, 1997, 166-175.

    [6] A.D. Blair & J.B. Pollack, 1997. Analysis of dynamical recognizers, Neural Computation 9(5), 1997, 1127-1142.

    [7] A.D. Blair, 1995. Adelic path space integrals, Reviews in Mathematical Physics 7(1), 1995, 21-49.

    Book Chapters

    [8] J. Wiles, A.D. Blair & M. Bodén, 2001. Representation Beyond Finite States: Alternatives to Push-Down Automata, in J.F. Kolen & S.C. Kremer (Eds.) A Field Guide to Dynamical Recurrent Networks, IEEE Press, 129-142.

    [9] E. Sklar, A.D. Blair & J.B. Pollack, 2001. Training intelligent agents using human data collected on the Internet, in J.Liu, N. Zhong, Y. Tang & P. Wang (Eds.) Agent Engineering, World Scientific, 201-226.

    Conference Proceedings - Coevolution, Learning and Robotics

    [10] J. Thomas, A. Blair & N. Barnes, 2003. Towards an efficient optimal trajectory planner for multiple mobile robots, Proceedings of the 2003 International Conference on Intelligent Robots and Systems, 2291-2296.

    [11] T. Ord & A. Blair, 2002. Exploitation and peacekeeping: introducing more sophisticated interactions to the Iterated Prisoner's Dilemma, Proceedings of the 2002 Congress on Evolutionary Computation, 1606--1611.

    [12] S. Versteeg & A. Blair, 2001. Getting the job done in a hostile environment, Proceedings of the 14th Australian Joint Conference on Artificial Intelligence (LNAI 2256).

    [13] D. Shaw, N. Barnes & A. Blair, 2001. Creating Characters for Dynamic Stories in Interactive Games, Proceedings of the International Conference on Application Development of Computer Games in the 21st Century.

    [14] J. Wiles, H. Chenery, J. Hallinan, A. Blair, A. & D. Naumann, 2000. Effects of damage to the CDM Stroop model, Cognitive Science in Australia 2000: Proceedings of the Fifth Biennial Conference of the Australasian Cognitive Science Society.

    [15] E. Sklar, A.D. Blair, P. Funes & J.B. Pollack, 1999. Training intelligent agents using human Internet data, Proceedings of the First Asia-Pacific Conference on Intelligent Agent Technology, World Scientific, 354-363.

    [16] A.D. Blair & E. Sklar, 1999. Exploring evolutionary learning in a simulated hockey environment, Congress on Evolutionary Computation, 197-203.

    [17] N.Ireland & A.D. Blair, 1999. Target signal selection for a neural network based financial classifier, ICSC Symposium on Soft Computing in Financial Markets.

    [18] A.D. Blair, E. Sklar & P. Funes, 1999. Co-evolution, determinism and robustness, X. Yao et al. (Eds.): Proceedings of the Second Asia-Pacific Conference on Simulated Evolution And Learning (SEAL'98) LNCS 1585, 389-396.

    [19] A.D. Blair & E. Sklar, 1998. The evolution of subtle manoeuvres in simulated hockey, Proceedings of the Fifth Conference on Simulation of Adaptive Behavior, Zurich, 1998, 280-285.

    [20] E. Sklar, A.D. Blair & J.B. Pollack, 1998. Co-evolutionary learning: machines and humans schooling together, in: G. Ayala, ed., Proceedings of Workshop on Current Trends and Applications of Artificial Intelligence in Education, ITESM, Mexico, 98-105.

    [21] A.D. Blair, 1999. Co-evolutionary learning - lessons for human education? Proceedings of the Fourth Conference of the Australasian Cognitive Science Society, Newcastle, Australia, 1997.

    [22] J.B. Pollack & A.D. Blair, 1997. Why did TD-Gammon work? Advances in Neural Information Processing Systems 9, 1997, 10-16.

    [23] J.B. Pollack, A.D. Blair & M. Land, 1997. Coevolution of a Backgammon player, Proceedings of the Fifth International Conference on Artificial Life, MIT Press, 1997, 92-98.

    Conference Proceedings - Dynamic Language Processing

    [24] B. Tonkes, A.D. Blair & J. Wiles, 2000. Evolving learnable languages, In S.A. Solla, T.K. Leen & K.-R. Muller (Eds), Advances in Neural Information Processing Systems 12, MIT Press, 66 - 72.

    [25] S. Chalup & A.D. Blair, 1999. Hill climbing in recurrent neural networks for learning the anbncn language, Proceedings of the Sixth International Conference on Neural Information Processing, 508-513.

    [26] M. Bodén, J. Wiles, B. Tonkes & A.D. Blair, 1999. Learning to predict a context-free language: Analysis of dynamics in recurrent hidden units, Proceedings of ICANN'99, Edinburgh, 359-364.

    [27] B. Tonkes, A.D. Blair & J. Wiles, 1999. A paradox of neural encoders and decoders or Why don't we talk backwards? X. Yao et al. (Eds.): Proceedings of the Second Asia-Pacific Conference on Simulated Evolution And Learning (SEAL'98) LNCS 1585, 357-364.

    [28] A.D. Blair & J. Ingram, 1998. Loanword formation: a neural network approach, Proceedings of the Fourth Meeting of the ACL Special Interest Group in Computational Phonology, Montreal, 1998, 45-54.

    [29] B. Tonkes, A.D. Blair & J. Wiles, 1998. Inductive bias in context-free language learning, Proceedings of the Ninth Australian Conference on Neural Networks, Brisbane, Australia.

    [30] A.D. Blair & J.B. Pollack, 1998. Quasi-orthogonal maps for dynamic language recognition, Proceedings of the Fourth International Conference on Neural Information Processing, Springer, 1997, 1065-1067.

    [31] A.D. Blair, 1995. Two layer digital RAAM, Proceedings of the Seventeenth Annual Conference of the Cognitive Science Society, Pittsburgh, PA, 1995, 478-481.

    Technical Reports and Thesis

    [32] S.K. Chalup & A.D. Blair, 2000. First Order Recurrent Neural Networks Learn to Predict a Mildly Context-Sensitive Language, Department of Computer Science and Software Engineering, The University of Newcastle, Technical Report 2000-06, ISBN 0-7259-1109-3.

    [33] A.D. Blair, 1997. Scaling up RAAMs, Brandeis University Computer Science Technical Report CS-97-192.

    [34] A.D. Blair, 1994. Path Integrals on Ultrametric Spaces, Doctoral Dissertation, MIT, 1994.

    Return to Alan Blair's home page