RMIT ECML Group

cathamAI and Robotics

Oct 23, 2013 (3 years and 9 months ago)

141 views

RMIT ECML Group

Vic Ciesielski (Group Leader)

Paul Boxer, James Brusey, Tom Gamble, Ken Eckermann,

Andrew Innes, Brian Lam, Xiaodong Li,Tom Loveard,


Dylan Mawhinney, Michael Matkovic,Tania Binos,

Jeff Riley, Andy Song, Laura Thompson, Yaniv Bernstein



RESEARCH OBJECTIVES

To gain a better understanding of
EC systems by applying them to
real world problems.

THEORETICAL PROJECTS
AND APPROACHES


Texture discrimination in Genetic
Programming


Fine grained parallel Genetic Algorithms


Genetic Programming methods for
classification problems


Parameter control in co
-
evolutionary
Genetic Algorithms


Preserving diversity in Genetic
Programming.


SAMPLE PROJECTS

Genetic Programming to
evolve programs to control
RoboCup soccer players.







Texture discrimination using
Genetic Programming








Varying population density in
parallel Genetic Algorithms

ACHIEVEMENTS & RESULTS


Papers in CEC2002:


Texture classifiers generated by Genetic Programming,

Andy Song, Vic Ciesielski, Hugh Williams


Prevention of early convergence in Genetic Programming by replacement of similar programs



Vic Ciesielski, Dylan Mawhinney



The effects of varying population density in a fine
-
grained parallel Genetic Algorithm





Xiaodong Li, Michael Kirley


vc,xiaodong,szhengan@cs.rmit.edu.au RMIT, Department of Computer Science, GPO BOX 2476V, Melbourne, Australia 3001

PRACTICAL PROJECTS
AND APPROACHES


RoboCup soccer players evolved in
Genetic Programming


Evolving fuzzy rules for RoboCup
soccer players using Genetic Algorithm


Analysis and classification of beer foam
texture using Genetic Programming


Qualitative spatial reasoning to learn
naïve physics using Bayesian networks


Reinforcement learning for physical
robot behaviours


Web mining for adaptive web sites.


Malt classification from texture


Cephalometric analysis of cranial X
-
rays


Congress on Evolutionary Computation

GALESIA


Neural
Network
Council

IEEE