School of Design, Engineering & Computing

‘Complexity is not a problem’ says computational expert

19 August 2011

Professor Bogdan Gabrys speaks at INFER workshop.

The current explosion of global data is laying the foundations for a new decade of more advanced predictive analytics according to Bournemouth University Professor Bogdan Gabrys.

With the world’s ‘digital universe’ set to expand to 44 times as large in 2020 than it was in 2009, Professor Gabrys has encouraged his fellow experts in computational intelligence to move away from the engineering maxim that ‘simple is beautiful’ by embracing the challenges of the more biological concept that ‘complexity is not a problem’.

Professor Gabrys, Director of the University’s Smart Technology Research Centre, was speaking at the latest workshop organised as part of INFER, a major EU-funded project involving researchers from organisations in the UK, Poland and Germany including experts from BU’s Smart Technology Research Centre. INFER is focusing on pervasively adaptive software systems for the development of an open modular platform applicable in various commercial settings and industries. The main innovation of the project is a novel type of environment in which the “fittest” predictive model will emerge - either autonomously or by user high-level goal-related assistance and feedback.

The latest workshop, hosted by Bournemouth University, gave researchers and practitioners an opportunity to meet and discuss the advances and challenges in designing and building Adaptive Prediction Systems.

“It is time we moved away from the approach that ‘simple is beautiful’ and towards more complex solutions because we now have the computers and computational power available to tackle much more complex problems,” say Professor Gabrys who opened the workshop as one of four keynote speakers with a presentation on ‘Building Robust and Adaptive Predictive Systems’.

“Our problem in the past was having too little data but over the last couple of decades it’s changed to the situation where companies have now become data rich but information poor,” he continued. “They’re asking themselves ‘where is my needle in the haystack? - we’ve got a lot of data but what’s useful, what should I react to and what should I do with that information?’ That’s our challenge.”

Professor Gabrys was joined on the programme by Dr Joao Gama of the University of Porto in Portugal who spoke on ‘Challenges and issues in learning from data streams’; Dr Rogerio de Lemos of the University of Kent who focused on ‘Architecting Resilience in Self-Adaptive Software Systems and Dr Albert Bifet of the University of Waikato, New Zealand, who gave further insight on ‘MOA: a Real-time Analytics Open Source Framework’.

The 1.55 Million Euro INFER project is funded by the European Commission within the Marie Curie Industry-Academia Partnerships & Pathways (IAPP) programme and will run until June 2014. Through extended secondments, the project offers researchers the possibility to move sector and country in order to provide, absorb and implement new knowledge in a professional industrial-academic environment.

INFER Partners include Evonik Industries from Germany, one of the world's leading companies in the process industry and Research & Engineering Centre (REC) from Poland, a highly innovative software engineering company. The University’s Smart Technology Research Centre is an interdisciplinary and integrative centre conducting research in the field of automated intelligent technologies.