At IBM Research I’ve been working on research in Bayesian Deep Learning (BDL), Bayesian Optimisation and Computer Vision. Part of my work has involved leading the development of IBM’s upcoming Bayesian Optimisation product, which has been covered in a number of articles across publications ranging from HPC Wire to Forbes.

I’ve also been involved in the aiScope project to help transform UK agriculture, in which I’ve been developing Computer Vision-based technologies and integrating these with geospatial big data systems to help improve data management and develop robust forecasting tools for smart agriculture.

My publications at IBM Research involve work on improving the scalability of Probabilistic Backpropagation and on investigating the impact of the quality of BDL model uncertainty estimates within safety critical systems. Upcoming publications include work on big data approaches for mobile signal strength prediction (in which we investigate the use of BDL methods) and a novel method for multi-fidelity Bayesian optimisation (links to be provided when available).

Other projects I’ve been involved with at IBM include:

• Improving efficiency of ports via advanced predictive modelling techniques

• Bayesian Optimisation for optimising chemical reactor yield