Health Intelligence

Health Intelligence

Publications

News & features

News & features

Overview

Breakthrough advances in AI and machine learning (ML) have led to ambitious visions of how new systems can help revolutionize healthcare. These range from new approaches to understanding health risks, predicting disease progression, and creating personalized health interventions for improved patient outcomes; through to the development of innovative tools to support the practices of healthcare professionals, and reduce spending.

To realize this tremendous potential requires the development of machine learning applications that are effective, trustworthy and implementable in real healthcare contexts. Tackling this challenging ambition is the Healthcare Intelligence team at Microsoft Research Cambridge – a multi-disciplinary group of machine learning experts, social scientists, designers and engineers. Working jointly with experts in healthcare, privacy and compliance, the team develops human-centred ML solutions that can transform care pathways and improve health outcomes. This includes a strong focus on the development of “responsible AI” technology to ensure that the types of solutions we create will benefit all people, and align with their values and ethics. We approach this by following Microsoft’s AI Principles of: fairness, reliability & safety, inclusiveness, privacy & security, accountability and transparency.

We are particularly interested in the following areas:

busy hospital corridor

Hospital care

Managing patient flows along healthcare pathways is a critical concern for hospitals in both operational and clinical terms. Working with key clinical partners, we are developing Machine-Learning based systems that will offer intelligent insights about these operational and clinical events. The research adopts a human centred machine learning perspective that combines deep clinical knowledge, ethnographic understanding of clinical work practices and machine learning expertise to develop ML insights that are actionable in real world clinical contexts.

close up of lab test tubes

Immunomics

The adaptive immune system is an incredibly complex and powerful defense against disease. With Microsoft Immunomics we are developing state-of-the-art deep generative models trained on vast amounts of T-cell repertoire data, learning not only to predict but to understand disease. By decoding the immune system in this way,  we can transform precision diagnostics and treatment options for many diseases.


Dr. Raj Jena using Innereye in a hospital setting

Medical imaging

In Project InnerEye, we are developing ML-based productivity tools for the automatic, quantitative analysis of radiological images – in particular, our work turns radiological images into measuring devices. Our focus is on the following applications: 1. extraction of targeted radiomics measurements for quantitative radiology, 2. efficient contouring for radiotherapy planning, and 3. precise surgery planning and navigation.

someone holding their head

Mental health

One in four of us will be affected by a mental health condition at some point in our lives. Mental health presents one of the most challenging and under-investigated domains of machine learning research. With the great need for effective mental healthcare services combined with breakthrough advances in machine learning and AI, Project Talia explores how a human-centric approach to machine learning can meaningfully assist in the detection, diagnosis, monitoring, and treatment of mental health problems.

coronavirus graphic

Pandemic preparedness

COVID-19 has mobilized communities around the world to work together on a range of critical issues required to understand and respond to the current pandemic. Scientific research is vital to both this near and long-term pandemic preparedness.

wet alb, hands using pipet

Synthetic biology

The ability to program biology could enable fundamental breakthroughs in medicine and many other industries. The Station B project builds on over a decade of research at Microsoft on understanding and programming information processing in biological systems, in collaboration with several leading universities, and aims to improve all phases of workflow typically used for programming biological systems.

People