In the presentation, Dr Dickens will present some of the models he has worked on including: neuroscientific models of human decision making, crowdsourcing and mixture of expert models for source fusion, and models of dynamic language use in online forums.
Dr Luke Dickens is a Lecturer in Machine Learning and Data Science in the Dept. of Information Studies at UCL. His research expertise includes statistical inference techniques, probabilistic modelling, human behavioural modelling and machine learning for decision support. Before coming to UCL, Dr Dickens worked as a research associate at Imperial College London on a number of projects including his own EPSRC doctoral prize fellowship ‘Reinforcement learning for science’. He has also worked in a number of successful cross-disciplinary research collaborations including with social-scientists, developing linguistic models to detect shifts in social identity in written material; with medical scientists, investigating technologies for the artificial pancreas; with neuroscientists, investigating reinforcement learning models of human behaviour; and with bioengineers, exploring the use of vision technologies for localisation and navigation. In each of these collaborations his overarching research agenda has been the validation and use of machine learning models for science.
All members of staff are welcome to attend and RAs and postgraduate students are particularly encouraged to attend. Any enquiries should be directed to Dr Jim Ang. (C.S.Ang@kent.ac.uk)