I studied Computer Science at Warwick university, then, after a few years outside academia, I joined Edinburgh to take MScs in Informatics and Neuroinformatics and a PhD in computational neuroscience, looking at where self-motion cues are processed and integrating, in the human brain (using fMRI).

In 2014, after a bit of travelling I went to Kampala (Uganda) to lecture ML/AI at Makerere University. I’m now a Research Fellow at the University of Sheffield in the department of Computer Science in the Machine Learning group.

My work encompasses Differential Privacy and its applications to Gaussian process (GP) regression and classification, bounds on attacks to GP classifiers by adversarial examples, a kernel for regression over integrals and a method for tracking bees using retroreflective tags.

My work is in particular now focused on modelling air pollution in Kampala, using data from a network of low-cost sensors. I’m currently investigating probabilistically handling the calibration of the sensors using mobile units. This system will soon be incorporated into a pipeline providing live predictions for policy makers and stakeholders in the city.

List of current projects,

  • We’re developing new tools to allow data to be anonymised, through the framework of differential privacy (see paper in AISTATS on Gaussian Processes and Differential Privacy).
  • Air pollution Kampala – a real time model of the air pollution (PM2.5) across Kampala (see talk and a blog post about my last visit there). We’re hoping to soon run validation experiments and roll out the monitoring to the whole city.
  • I’m currently working on finding bounds on adversarial attacks against GP classifiers.
  • A paused project: Clustering patients in the MND database.
  • Dialysis: Predicting a few days ahead for clinical decision support.
  • Other projects: Bee tracking, integral kernel, etc…

Email me: m.t.smith@sheffield.ac.uk

Previous work

My first MSc looked at how to infer features of distal synapses using the variance in the flow of current. My second MSc research focused on the network-level modelling of the head direction system and hands-on work with tetrode recordings of place cells in rat, in an attempt to understand if different sources of information are integrated in the dorsal presubiculum. For my PhD I switched to using fMRI in humans, but remained focused on the area of sensory integration in the head-direction and navigation systems. In particular how we integrate visual and auditory self-motion cues.

I spent most of 2014 lecturing at Makerere University, Kampala, Uganda. There I became involved in the field of Development Informatics, and have several on-going research topics; covering air pollution, nutrition-data, automated microscopy, traffic collision data and malaria distribution prediction. A variety of machine learning methods have been applied (for example Gaussian Process models for the model of malaria distribution). More details about some of these projects can be found at the Artificial Intelligence in the Developing World (AI-DEV) group’s website.

In 2015 we organised the Data @ Sheffield event, as part of the open data science initiative.

As part of an innovate UK collaboration we built the scikic inference tool, which provided both a conversation interface and a backend API for inferring demographic and lifestyle features about individuals. It is hoped it will be a useful tool to demonstrate the power of machine learning. In the future we hope to develop a user-centric data model for the analysis and storage of user data, with the motivation that personalised medicine and associated research requires access to user data.


We hold annual Data Science Africa events in June/July in East Africa, as part of the wider data science Africa network.

Selected Publications
  • Smith, M. T., Livingstone, M., Comont, R. A method for low-cost, low-impact insect tracking using retroreflective tags, Methods in Ecology and Evolution. (in review)
  • Smith, M. T., Álvarez, M. A., Lawrence, N. D. Gaussian Process Regression for Binned Data. Statistics and Computing. (in review)
  • Smith, M. T., Álvarez, M. A., Lawrence, N. D. Improved Differentially Private Regression and Classification with Sparse Gaussian Processes. Journal of Machine Learning Research. (in review)
  • Smith, M.T., Álvarez, M. A., Zwiessele, M., Lawrence, N. D. (2018). Differentially Private Gaussian Processes, AI STATS 2018. (link)
  • Mubangizi, M., Andrade-Pacheco, R., Smith, M. T., Quinn, J. A., & Lawrence, N. (2014). Malaria surveillance with multiple data sources using Gaussian process models. 1st International Conference on the Use of Mobile ICT in Africa.
  • Wutte, M. G., Smith, M. T., Flanagin, V. L., & Wolbers, T. (2011). Physiological signal variability in hMT+ reflects performance on a direction discrimination task. Frontiers in psychology, 2.
  • Bett, D., Stevenson, C. H., Shires, K. L., Smith, M. T., Martin, S. J., Dudchenko, P. A., & Wood, E. R. (2013). The Postsubiculum and Spatial Learning: The Role of Postsubicular Synaptic Activity and Synaptic Plasticity in Hippocampal Place Cell, Object, and Object-Location Memory. The Journal of Neuroscience, 33(16), 6928-6943.
  • Feldwisch-Drentrup, H., Barrett, A. B., Smith, M. T., & van Rossum, M. C. (2012). Fluctuations in the open time of synaptic channels: An application to noise analysis based on charge. Journal of Neuroscience Methods, 210(1), 15-21.
Talks and invited presentations
  • Smith, M. T., Bagonza, J. (2016). Crowd Sourced Transcription of Kampala’s Traffic Collision Data. Data Science Africa 2016. (talk) (video)
  • Smith, M. T., Zwiessele, M., Lawrence, N. (2016). Differentially Private Gaussian Processes. NIPS workshop PMPML16. (poster & spotlight) (link)
  • Smith, M. T. (2017). Low cost air quality monitoring in Uganda. Data Science Africa 2017. (talk) (link)
  • Smith, M. T., Álvarez, M. A., Zwiessele, M., Lawrence, N. D. (2017). Sparse Gaussian Processes for Differentially Privacy. TI3 Project workshop. (poster)
  • Smith, M. T., Álvarez, M. A., Zwiessele, M., Lawrence, N. D. (2017). Differentially Private Gaussian Processes. Advances in Data Science 2017. (talk, poster & extended abstract)
  • Smith, M. T., Álvarez, M. A., Zwiessele, M., Lawrence, N. D. (2017). Improved Differential Privacy using Inducing Variables. Gaussian Process Approximations Workshop, Amazon Berlin. (invited talk)
  • Smith, M. T., Álvarez, M. A., Fotheringham, J. (2018). Short-term Modelling of Clinical Variables in Hemodialysis Patients. BAYES COMP 2018. (Poster)
  • Smith, M. T., Ssematimba, J., Álvarez, M. A., Bainomugisha, E. (2018). Air quality monitoring in Uganda using motorbike taxis. Data Science Africa 2018. (talk)
  • Smith, M. T., Ssematimba, J., Álvarez, M. A., Bainomugisha, E. (2018). Gaussian Process Models for Low Cost Air Quality Monitoring. Advances in Data Science 2018. (talk & extended abstract)
  • Smith, M. T. (2018). Methods for high dimensional DP for GP regression and classification. CFE CMStatistics 2018. (Invited talk) (slides)