My research interests lie in areas of computer science and biomedical engineering with practical applications to healthcare delivery.
Specifically, I'm interested in
Clinical Machine Learning, Biological Signals, and Health-data Analytics,
Explainabilityand reliable Validationin Medical Image Analysis, and
- Uses of
techin Public Healthresearch.
Below are some examples of projects I've worked
on over the years.
(click headings to expand)
Brain-Computer Interfaces for optimal co-operation between human
and AI agents. (University of Essex)
I am currently working as a Senior Research Officer at the prestigious Brain-Computer Interfaces and Neural Engineering Lab at the University of Essex, working on the BARI project: a US/UK Bilateral Academic Research Initiative.
This is a large, DSTL-funded international project led by Prof. Riccardo Poli, involving partners from the University of Essex, University of Oxford, Harvard Medical School, University of Massachusetts Medical School, University of Southern California, and University of California (Berkeley).The aim of the project is to investigate optimal collaboration in teams consisting of both human and artificial agents, in order to optimise human-AI collaborative decision-making. My role has been to implement, conduct, and analyse EEG-based experiments, and investigate the role of behavioural and neuro-physiological biomarkers relating to confidence and trust when performing group decisions in non-trivial scenarios.
Linear Dynamical Systems and signal fusion for the prediction
and management of secondary depression
I worked as a Senior Research Officer at the University of Essex, on the NEVERMIND project: "Neurobehavioural predictive and personalised modelling of depressIve symptoms during primary somatic Diseases with ICT-enabled self-management procedures".
This was a large, Horizon2020-funded international project involving University of Essex, University of Pisa, Madrid Polytechnic University, University of Turin, University of Lisbon, Karolinska Institute, and a number of private companies).The project involved the use of intelligent tools and systems enabling depression self management in patients with secondary depression. Under the supervision of Dr Luca Citi and Prof. Riccardo Poli, I was responsible for researching appropriate computational models on the basis of a sophisticated Decision Support System, for providing suitable, real-time inference of patients’ mental state, enabling optimal interventions through personalised models and real-time feedback.
Clinical measures of uncertainty, appropriate validation, and explainability in medical image analysis.
This reflects the topics covered by my DPhil thesis, at the University of Oxford.
My DPhil research focused on intelligent image analysis and segmentation in medical images. The work focused on three themes:
- Expressing clinical measures of uncertainty that complement statistical measures, and guide segmentation algorithms towards outcomes that are clinically optimal as well as statistically optimal.
- Explore the properties of extant validation operators and metrics, and propose improved metrics that are more reliable in the presence of fuzzy and probabilistic algorithms and ground truth sets.
- Characterise modes of segmentation failure that can be detected during validation, providing a layer of explainability to medical image segmentation algorithms.
Wearable technologies for ‘quantitative self’ analysis and
I, together with three friends, co-founded Sentimoto Ltd in 2013 during my DPhil at the University of Oxford. Sentimoto was a company dealing with wearables for older adults and their families. We created a prototype device, as well as software for existing wearables, and used it to analyse signals to detect markers of wellbeing in older adults, to support their health and enable sharing insights with family members.
We won several awards for our work, including the UnLtd Fast Growth award in 2014. I eventually resigned as co-director from the company in 2015 to focus on finishing my DPhil. The company successfully exited to Babylon Health shortly after.
Low-cost monitoring using SMS-signals to predict handpump health
in rural Africa.
This was a project I got involved in during my DPhil (unrelated to my studies), in collaboration with researchers from the Department of Geography at the University of Oxford. The project aimed to improve access to drinking water in rural Kenya, where most people still rely on handpumps for their water needs.
The project involved installing low-cost sensorts to the handles of handpumps, which then sent SMS messages for analysis (and thus bypassing the need for an internet infrastructure for data collection). My team was responsible for creating a web interface for the retrieval, visualisation, and analysis of the information from the SMS signals.In partnership with local businesses, the project allowed local authorities not only to monitor handpumps effectively and efficiently, but also helped with predictions of failure. This ensured repair mechanics could be dispatched as quickly as possible at the point of failure, minimizing disruptions to water supply in the affected areas.