Dr Ivan Jordanov
Reader in Computational Intelligence
I am a Reader and leader of the Computational Intelligence Research Group at the School of Computing. I have a PhD in Computer Aided Optimization and an MSc in Applied Mathematics from the Technical University of Sofia; and a BSc in Mechanical and Electrical Engineering from the Naval University, Bulgaria. I joined the University of Portsmouth in 2003 after 3 years with De Montfort University as a Senior Lecturer; 2 years with the University of Wales Institute, Cardiff as a Senior Researcher; and 16 years with the Technical University Sofia as Associate Professor.
My main research interests and activities are in the field of computational intelligence:
Machine Learning, Neural Networks and Heuristic Global Optimisation - Investigation, modeling and development of intelligent hybrid optimization methods that encompasses heuristic and deterministic approaches for global optimization. These include metaheuristics based on so-called low-discrepancy sequences of points, genetic algorithms and local search optimizers. The developed methods are employed for supervised Neural Network (NN) learning.
Pattern Recognition - Investigation and development of Intelligent Machine Vision System that can grab an image, digitise it and pre-process the data, incorporate segmentation and feature based extraction and dimensionality reduction with PCA, ICA and LDA, and apply supervised and unsupervised NN learning for solving pattern recognition, image identification and classification problems.
COMPUTATIONAL INTELLIGENCE (MACHINE LEARNING, NEURAL NETWORKS, HEURISTIC GLOBAL OPTIMISATION AND GENETIC ALGORITHMS)
Investigation, modelling and development of intelligent hybrid optimization methods that encompasses heuristic and deterministic approaches for global search and optimization. These include metaheuristics based on so-called Low-discrepancy sequences of points, Genetic algorithms and Local search optimizers. The methods investigate the trade-offs between exploration and exploitation and the efficiency of the global search.
This research includes all aspects of analysis, design, modelling and investigation of computationally intelligent pattern recognition machine vision system – data acquisition and data representation, statistical pre-processing, feature extraction and dimensionality reduction (PCA, ICA and LDA), decision making (neural networks) and system evaluation. The developed methods are employed for Neural Network learning of Image Processing and Pattern recognition problems. Different training paradigms and performance metrics taking into account data features and data distribution are tested and evaluated for solving identification, mapping and classification problems.