j.hogan@qut.edu.au

Username: j.hogan@qut.edu.au

Participant details

	Network Member	James M. Hogan
	Organisation/Institution	Queensland University of Technology
	Department/Centre	School of Software Engineering and Data Communications
	Lab
	Website	www.fit.qut.edu.au/~hogan
	Research area	Natural Language Processing Artificial Intelligence
	Research keywords

Research Interests

James M. Hogan's research interests are centred upon machine learning and its application to cognitive science and bioinformatics problems, with subsidiary interests in software engineering. In particular, his work is concerned with learning in visual domains with associated text, in the tradition of the Berkeley L0 project. His most significant contribution in this area is a connectionist synthesis of bottom up and top down attentional systems as a model of spatial relations acquisition and processing. This work is now being extending through a more realistic model of top-down attentional processing, in collaboration with researchers from the Johns Hopkins University. Additional contributions in the field of visual processing have centred upon the classification of facial expressions through the use of novel neural network receptive fields - and subsequently within the support vector machine framework. Within the pure text processing domain, Dr. Hogan is active in the development of a Hidden Markov Model approach to authorship attribution, with conceptually related work in the use of novel kernels for the identification of promoter motifs with bacterial genomes. More recently, he has taken a leading role in the development of software internationalisation work within Australia.