The knowledge discovery research is comprised of two main areas:
1. Pattern Recognition and Machine Learning (Kuncheva). The aim of this research theme is to develop internationally leading research on classifier ensembles and their applications, through individual research (a second edition of a monograph is being prepared) and domestic and international collaborations. Encouraged by the results from a recent collaboration with a medical team from Swansea, we seek to expand the applied research of the (virtual) group into other health domains such as ageing, obesity and mental health.
2. Artificial Intelligence and Intelligent Agents, AI:IA (Teahan). This research theme addresses theoretical and applied research into artificial intelligence (AI) and intelligent autonomous agent systems, with specific focus on evolutionary algorithms, natural language processing and information retrieval. Four PhD and 1 MPhil student completed in the REF period and currently there are 6 PhD students with projects in: Chinese and Arabic Natural Language Processing; information retrieval for peer to peer networks and motion capture data; agent-based modelling for Malaria; and conversational agents for e-learning. We were the first to develop a framework (jGE) for using Grammatical Evolution (GE) to evolve Java programs (this was prior to GEVA being produced by the Dublin GE research group). jGE has been downloaded by researchers from the USA, UK, Netherlands, Poland, Brazil and Colombia. A paper published at IJCAI’11 describes a modification to GE that significantly outperforms other GE and Genetic Programming implementations on standard benchmarking tests.