Favour Nyikosa

DPhil Student, Engineering, Machine Learning Research Group

Favour Nyikosa’s research is in machine learning with an emphasis on Bayesian optimisation methods for sequential data with applications in online portfolio selection and domains using robotic sensors. Prior to this, he studied Advanced Computing at Imperial College, London, with a thesis titled "Global Optimisation using Gentlest Ascent Dynamics and Saddle Point Stability of Functions on Differentiable Manifolds".

He obtained his undergraduate degree in Computer Science from the University of Zambia and was a Network Engineer at Vodafone's Africonnect, then a Software Engineer on the SmartCare Electronic Health System project by the United States' government and its affiliates in Zambia. Favour is interested in using data to facilitate pragmatic and intelligent decision making in uncertain environments.


Working Paper

Rizvi, S., van Heerden, E., Salas, A., Nyikosa, F., Roberts, S., Osborne, M. and Rodriguez, E. (2017). Identifying sources of discrimination risk in the life cycle of machine intelligence applications under new European Union regs.
Nyikosa, F (2017). Iteratively-projected Bayesian optimisation.
Nyikosa, F (2017). A framework for adaptive Bayesian optimisation.
Nyikosa, F., Osborne, M. and Roberts, S (2015). Adaptive Bayesian optimisation for online portfolio selection (NIPS).
Rizvi, S.A.A., Roberts, S.J., Osborne, M.A., and Nyikosa, F. (2017). A novel approach to forecasting financial volatility with Gaussian process envelopes.
Rizvi, S.A.A., Roberts, S.J., Osborne, M.A. and Nyikosa, F. (2017). Predicting dynamic renyi entropy using gaussian processes to estimate financial information flows.