Using Explainable-AI to leverage biomedical datasets

Add to your list(s) Download to your calendar using vCal

• Joel Mitchell
• Wednesday 12 May 2021, 12:00-13:00
• online via zoom.

If you have a question about this talk, please contact Dr. Marc Haber.

Combined with powerful computational resources and an ever-growing library of machine learning algorithms the bioinformatician has a plethora of exploratory options. Some of these, however, can be black boxes. Explainable-AI (XAI) aims to be interpretable, permitting understanding of the mechanism behind the data-based prediction. In this talk we discuss one such algorithm, iterative Random Forests (iRF), with an application to the UK Coronavirus Cancer Monitoring Project (UKCCMP) dataset. The UKCCMP was set up to enable clinicians to make informed choices in cancer patient care during the COVID -19 pandemic; to do this, patient data from cancer centres across the United Kingdom was collected and analysed. iRF is, at its a core, a random forest classifier. However, it also recovers interactions by intersecting sets associated with decision paths leading to class one leaf nodes. These recovered interactions may then generate hypotheses for future testing. We will look at how iRF performs at predicting patient outcome in the UKCCMP dataset and discuss methodologies around applying it.

This talk is part of the CCB seminars series.

This talk is included in these lists:

• online via zoom

Note that ex-directory lists are not shown.