Computational Intelligence Research Group

Download Report

Transcript Computational Intelligence Research Group

Computational Intelligence
Research Group
Principal investigators:
Prof Andries Engelbrecht
([email protected])
Mr Bryton Masiye
([email protected])
Mr Nelis Franken
([email protected])
http://cirg.cs.up.ac.za
Main Research Focus
 Conducts theoretical and applied research in
computational intelligence
 Application areas:



Optimization
Classification
Pattern recognition
 Research areas:







2

Artificial neural networks
Swarm intelligence
Evolutionary computation
Artificial immune systems
Bioinformatics
Swarm robotics (MAS)
Data and text mining
Image analysis
CIRG Members
 Staff members

Andries Engelbrecht, Bryton Masiye, Nelis
Franken
 Graduate students

PhD students: 9
 Masters students: 35
 B.Sc-Hons project students: 4
3
Research outputs
4
Year
Journal
articles
Conference
papers
Book
chapters
Books
2001
2
3
1
-
2002
1
7
-
1
2003
-
10
-
-
2004
4
2
1
-
2005
6
9
-
1
Graduandi





PhD students: 3
M.Sc students: 10
M.IT students: 1
M.BA students: 1
B.Sc-Hons students: 29
Funding
 NRF Focus area grants

Explorative Computelligence, 2003-2004,
R380000
 CiClops, 2005-2009, R1000000
 BMW surface scanning via BE@UP

Prototype, 2002-2003, R109000
 Production system, 2004-2005, R561770
International Collaboration
 With



Prof J-P Muller, CIRAD, France
Prof A Salman, University of Kuwait
Dr MGH Omran, Arab Open University, Kuwait
 Visitors



Prof W Duch, Nicholaus Copernicus University,
Poland
Prof J-P Muller, CIRAD, France
Prof M Wolldridge, University of Liverpool, UK
Research Detail
 Artificial neural networks:

New training strategies
 Model selection
 Learning function derivatives
 Artificial immune systems:

Analysis of overfitting
 New classifier algorithms
 Data clustering in changing environments
Research Detail (cont.)
 Swarm Intelligence:

Particle swarm optimization:
•
•
•
•
•
•
•
•
•
New optimization algorithms
Niching to locate multiple solutions
Multi-objective optimization
Theoretical analysis
Training game-playing agents
Tracking dynamically changing objectives
Constrained optimization
Training support vector machines
Methods for discrete search spaces
Research Detail (cont.)
 Swarm intelligence (cont.):

Ant Colony Optimization:
• Routing in ad hoc mobile networks
• Foraging models for robot swarms

Swarm robotics:
• Social-based coordination methods
• Coevolutionary trained behaviours
• Blackboard systems
Research Detail (cont.)
 Evolutionary computation:

Analysis of coevolution
 Coevolutionary training of game agents
 Evolving decision and model trees
 New differential evolution algorithms
 Bioinformatics:

DNA sequence alignment
 RNA secondary structure prediction
 Molecular docking
Research Detail (cont.)
 Image analysis:

Surface anomaly detection
 New image segmentation methods
 Analysing video feeds
 Gesture recognition and tracking
 Data, Text, Image Mining:

Data sampling methods
 Mining continuous-valued classes
 Genetic programming for mining
Researc Projects








BMW surface scanning
Cilib and CiClops
CyberSaint
Scorpio
Agere
Gesture recognition, device free pointing
Counting parasitic bees
Protein visualiser
Future
 Centre of Excellence in AI