Using Analytic AI to improve Behavioural Research

Speakers: 

• Dr Janna Hastings, Assistant Professor of Medical Knowledge and Decision Support, Universities of St. Gallen and Zurich
• Professor Susan Michie, Co-Director of BR-UK, University College London
• Professor Robert West, Professor Emeritus, University College London

About the webinar

This webinar is the second in a series of three sessions on the use of AI in behavioural research. The first was on generative AI and focused on its use to answer research questions, structure text and generate ideas for interventions. This session is about analytical AI and focuses on its use in classifying knowledge, and reasoning and making predictions from data. This form of AI is widely used in society, for example, tailoring advertising to customers, detecting fraud, weather forecasting and diagnosing diseases from medical images. In research, it is used for a wide variety of purposes, including data integration, pattern recognition and prediction, and natural language processing.

The webinar will cover the use of ontologies and machine learning to build tools to identify studies relevant to a research question and capture key information in reports intervention evaluations, and to use those data to make predictions of outcomes in novel intervention scenarios. It will present the Human Behaviour-Change project, a collaboration between computer and behavioural scientists. 

It will include a demonstration of a prototype tool trained using information extracted from approximately 500 reports of randomised trials evaluating smoking cessation interventions that can make predictions of smoking cessation outcomes based on user-inputted specifications of intervention content, mode of delivery, setting and study population. It will consider how this technology can be developed in different behavioural science domains.

There will be plenty of time for questions and discussion.