Alex started his doctoral studies at The Alan Turing Institute in October 2016. Alex is registered at the University of Edinburgh under the supervision of Chris Williams. His research interests include recurrent neural networks, state space models, generative models, multi-task learning and approximate inference.
Alex is working on multi-task dynamical systems, sequence models which explicitly model the inter-sequence variation (e.g. of different people, organisations etc.) within a dataset. This endows models such as linear dynamical systems and RNNs with the ability to adapt to particular individuals either via end-user control, or probabilistic inference. Applications include personalised drug response modelling and customisable walking style for motion capture sequences.