Explainable AI for Life Science

Overview

Artificial intelligence is inspiring potential new applications across the life sciences – from diagnosing disease in medical images to predicting clinical outcomes from tabular health records. But as AI models grow more powerful, they also grow more opaque. In high-stakes domains such as life sciences and healthcare, a prediction alone is not enough: we need to understand why a model made it, whether it can be trusted, and whether it learned the right thing or a shortcut.

This workshop provides an introduction to explainable AI (xAI) and its role in the life sciences. Through interactive lectures and hands-on notebooks, participants will apply explanation methods such as SHAP, feature importance, saliency maps, and Grad-CAM to real medical datasets – including brain stroke prediction from tabular data and classification of chest X-ray images.

Key takeaways

By the end of this workshop, you will:

• Understand use cases of AI in life sciences and how explainable AI (xAI) helps to make a step towards trustworthy AI in the life sciences.
• Describe the main categories of xAI methods and explain the differences between them.
• Evaluate model explanations critically, recognising that xAI is a tool – not a guarantee – and that explanations must be validated by domain experts.
• Gain hands-on experience in generating, interpreting, and critically validating AI explanations on real-life science problems.

Who is this for?

This workshop is intended for:

• Researchers, computational biologists, and data scientists working in life science, biomedical research, or healthcare
• AI/ML practitioners interested in biological or medical applications
• Anyone with an interest in trustworthy, interpretable, and explainable AI in science

Prerequisites

• Basic familiarity with Python is expected
• Experience with machine learning is preferable, but no prior experience with explainable AI is required