Broadly, my research interests are centred around the translation of abstract mathematics and computational methods to maximise the use of rich, multimodal and complex biomedical data. Through deep collaboration with clinicians, healthcare professionals and with patient input, I develop new and exploit already available methods to learn from structured (e.g. physiological time-series) and unstructured (e.g. free-text clinical notes, images) electronic health records and biomedical data. However, benefits of technological advances to patients and clinicians may be missed if the ethical challenges posed by the adoption of artificial intelligence in the healthcare system are not considered and addressed systematically. I am exploring the ethical implications of algorithms on clinical decision support tools and means to mitigate data and technology-induced biases.
I strongly support Open Science and Outreach. I am committed to sharing software and data, where possible, for reproducibility and benefit to a wider research community. I am also a member of the Patient and Public Involvement network of the Oxford Health Biomedical Research Centre and work closely with patients and lay members of the public to co-design research questions and better understand the impact of new technologies and induced disparities on patients and their families. I organise workshops and datathons on machine learning for biomedical data, such as the Oxford-Turing Workshop on Deep and Frequent Phenotyping in Neurodegeneration and Natural Language Processing for UK-CRIS medical records.