Gabriel Davis Jones

Research Overview

I lead Oxford Digital Health Labs, where our research is centred on the practical development and application of data science and artificial intelligence in healthcare.

Next Generation Assessment of Fetal Wellbeing

In the MRC DPFS project "Next Generation Assessment of Fetal Wellbeing using Artificial Intelligence," we collaborate with the University of Oxford's Department of Computer Science and the Royal Women's Hospital in Melbourne, Australia.

• Dataset: Managing data from over 200,000 pregnancies, including clinical and biosignal data (e.g., cardiotocography and ultrasound).
• Objective: Developing AI-driven tools to detect high-risk pregnancies, targeting complications like stillbirths and a range of adverse outcomes.
• Validation: Models are validated on diverse populations, leveraging datasets from various geographical and healthcare contexts.
• Global Impact: Our work includes collaboration with Professor Jane Hirst and CareMother India, where we assess models on an expansive Indian dataset, aiming for universally applicable tools to improve maternal and newborn healthcare worldwide.

Research and Development in Computerised CTG

I lead Research and Development in Computerised CTG, focusing on advancing the Dawes-Redman algorithm, a pioneering system developed to analyse fetal heart rate patterns during pregnancy.

• What is the Dawes-Redman Algorithm?
  This algorithm employs statistical and rule-based methods to provide objective, automated assessments of CTG recordings, enabling early detection of fetal distress. It is widely recognised for improving the interpretation of fetal heart rate and uterine contraction data, thereby supporting clinicians in making timely and accurate decisions.

• Current Focus:

  • Advancing the algorithm using state-of-the-art machine learning techniques.
  • Incorporating additional biosignals and clinical data to enhance prediction accuracy for adverse outcomes.
  • Validating these updates on diverse populations to ensure robust applicability across varied clinical settings.

• Clinical Impact:
  The goal is to create enhanced tools that reduce stillbirths, improve outcomes for high-risk pregnancies, and provide universal standards for CTG interpretation.

Oxford Martin School Programme on Global Epilepsy

I also lead research in the Oxford Martin School Programme on Global Epilepsy, focusing on developing AI tools for epilepsy diagnosis and treatment in resource-limited settings.

• Geographical Focus: Kenya, Ghana, Tanzania, South Africa, and India.
• Collaboration: Working with Professors Arjune Sen, Timothy Denison, and Sloan Mahone to leverage existing big data.
• Tools: Creating culturally and geographically tailored clinical tools for regions bearing 75% of the global epilepsy burden.

AI Safety and Large Language Models

Our group places significant emphasis on AI safety, particularly examining publicly available large language models (LLMs) such as ChatGPT and Gemini to evaluate their:

• Robustness and Performance: In both public and private healthcare settings.
• Reliability and Bias: Conducting rigorous testing with real-world clinical data.
• Ethical Considerations: Addressing patient privacy, inequity, and algorithmic bias.

This work aims to enhance the safe integration of advanced AI technologies into clinical practice.

Contributions to Global Healthcare

In addition, I am a subject matter expert at the Bill and Melinda Gates Foundation, contributing to initiatives that leverage AI to improve healthcare outcomes globally. I am also the director of AI and Big Data Analytics at the Oxford Centre for Global Epilepsy.

Collaboration Opportunities

We welcome engagement from individuals interested in computer science, artificial intelligence, data science, medicine, and global health.

Our team offers various projects and is open to inquiries from prospective collaborators.