Research covers diverse areas bridging the gap between various disciplines. The Intelligent Systems Laboratory addresses both the advance of fundamental theories and the design of real-world applications.
Research within the Intelligent Systems group includes:
The ability to identify individual animals is a prerequisite for many questions in behavioural ecology, cognitive research, conservation monitoring, and wildlife epidemiology. With the increasing availability of remote audiovisual recording devices, such as camera or video traps, standardized data collection has become much easier, in particular in the wild. The core objective is to study, develop and evaluate computer vision systems that exploit the structure of animal skin patterns for a robust biometric identification of individuals in natural habitats. The project will primarily focus on great apes and elephant. For further information please visit SAISBECO.
Research includes machine learning, pattern recognition, web technologies, data mining, bioinformatics, semantic image analysis and natural intelligent systems.
A further major project in animal biometrics is concerned with an automatic recognition of individual penguins in a colony. For further details please visit Spot the Penguin.
Data Mining is concerned with the automated analysis of large scale data in order to extract the useful knowledge hidden in it. Our approach to dealing with data has essentially remained unchanged for the past 25 centuries: we categorise it, divide it into small chunks, then build indexes and catalogues so we can find what we want. But with terabytes of data becoming available every second in areas such as finance, medicine and commerce, this is no longer a viable strategy.
As a result there is an urgent need to find new ways of solving the following types of problems. In finance, methods are needed to predict stock trends, based not only on traditional statistical indicators but, potentially, incorporating all the information that a human trader might use such as social and political trends and current events. Financial institutions need to generate credit scores before making loans, whether to individual home buyers or major corporations. Retailers need to mine associations from store card and online transaction data in order to customise advertising and promotions to customers. Banks, companies, universities, governments and other organisations are increasingly concerned with online security and the need to detect intrusions and anomalies.