Business Systems and Knowledge Modelling Lab

Media Release

Crime prediction and public health

Global research conference highlights latest in Intelligent Information and Engineering Systems.

One of the largest information technology conferences ever held in Melbourne - the International Conference on Knowledge-based Intelligent Information and Engineering Systems (KES 2005) - is being hosted by La Trobe University's School of Business in September.

The conference involves some 700 delegates from more than 40 countries, many from the Asia-Pacific region. Intelligent Information and Engineering systems, says the conference's general chairperson, Dr Rajiv Khosla, cover applications in an enormous range of endeavour, from finance, banking, tourism, manufacturing and automation to health care, bioinformatics, medical and general diagnosis - as well as security and defence.

News about the latest developments is being delivered by researchers from 400 universities and from industry bodies world-wide, such as the Australia's CSIRO and DSTO; Daewoo from South Korea; Hitachi and NTT, Japan; and Vimtech, Spain.

Dr Khosla, Associate Professor in the School of Business on the University's main Melbourne campus at Bundoora, says about 100 technical sessions demonstrate 'how far intelligent systems have come in translating research into commercial applications in practically all areas of business, commerce and engineering.'

While the main conference venue is at Hilton on the Park in Melbourne, prominent researchers from Harvard University, Tokyo University, Rensselaer Polytechnic Institute in New York, and other institutions are also presenting seminars at the La Trobe University Business Intelligence Institute - Business Systems and Knowledge Modelling (BII-BSKM) Research Laboratory.

Business Systems and Knowledge Modelling Laboratory

The laboratory, an externally funded research group, is headed by Dr Khosla. It comprises six postgraduate researchers, research assistants and 15 external research affiliates and collaborators from industry and from institutions in Japan, USA, and Europe.

The laboratory engages in applied research and develops IT products and prototypes for commercial applications. Some examples are:

Emotionally Intelligent Smart Recruitment Systems: 'Emotions,' says Dr Khosla, 'form an important component of human behaviour and decision making. The aim of this research is to design and develop a smart sales recruitment and benchmarking tool which will provide HR managers and recruitment agencies with psychological profiles of selling behaviour and the emotional state of sales candidates.'

The technique involves analysis of video images of the candidates. The work also has implications for 'web personalisation' - e-shopping, e-tourism and internet based decision support systems in general. An earlier version of this system was commercialised seven years ago.

On-line Environmental Benchmarking System for Consumer Manufacturing: The aim of this project is to help consumers make choices that might encourage them to buy more environmentally-friendly products. 'Such systems can also be useful for policy makers and regulators in determining industry and product benchmarking, and might lead to cleaner manufacturing processes.' he says.

Besides Dr Khosla, other key figures in this research project include Dr Clare D'Souza from La Trobe University and Dr Mehdi Taghian, now at Deakin University.

Face Detection and Annotation in Natural and Complex Settings: Finding faces in crowds and analysing facial expressions - such as in recent video footage from London underground train stations - is an important problem to be solved for large-scale security screening world-wide.

Dr Khosla's team is developing a system which uses co-operating image processing agents to hone in on certain critical parts of faces.

Automated Diagnostic System for Analysis of Serious Pathological Conditions: Demand for rapid analysis of specimens for diagnostic medicine is increasing dramatically. Dr Khosla says most automated diagnostic systems are used with stained specimens, a process that can interfere with cell characteristics.

'We have developed an accurate method of identifying unstained human cell images for serious pathological conditions such as breast cancer, where a pathologist looks at the grade or the appearance of potential cancer cells under a microscope, and in Chronic Myeloid Leukaemia, where the morphology of the white blood cells changes dramatically.'

Details about the La Trobe University Business Systems and Knowledge Modelling Laboratory can be found on http://www.latrobe.edu.au/bskm.

- La Trobe Bulletin, September 2005