Course: Evaluation of AI Algorithms for Biomedical Research and Sex and Gender Biases in AI

Rapid advances in artificial intelligence (AI) are transforming biomedical research, but these systems can internalize and amplify historical biases present in clinical trial data, annotations, and experimental design. Without rigorous evaluation, we risk developing and using models that misdiagnose, recommend ineffective treatments, and perpetuate existing inequalities.

For this reason, the Barcelona Supercomputing Centre (BSC) is organizing a four-day course to address sex and gender biases throughout the entire lifecycle of AI systems applied to biomedical research, from project conceptualization and data generation to model training, showcasing practical cases and providing tools to identify, evaluate, and mitigate these biases.

What Will You Learn?

Participants will understand key concepts and why they matter in biomedical research; identify specific biases in biomedical data and understand their impact on research quality; critically evaluate AI models using contextualized metrics and frameworks; and design bias-aware biomedical AI projects from the outset.

The course is structured into four integrated blocks, each day with a different focus:

• Monday 26 – Fundamentals and Concepts (4h): Part 1: 14:00–16:00, Break: 16:00–16:30, Part 2: 16:30–18:30


• Tuesday 27 – Identifying Biases in Training Data for AI (4h): Part 1: 14:00–16:00, Break: 16:00–16:30, Part 2: 16:30–18:30


• Thursday 29 – Detection of Bias and Evaluation in AI Models (4h): Part 1: 14:00–16:00, Break: 16:00–16:30, Part 2: 16:30–18:30


• Friday 30 – Project Design and Implementation of Best Practices (3h): Part 1: 16:30–18:30, Break: 16:00–16:30, Part 2: 16:30–17:30

The course will be taught by María Flores, Carolina Belver, Simón Perera, Davide Cirillo, Claudia Rosas, Olivier Phillipe, Barbara Fuzi, Summer Devlin, and Salvador Capella (to be confirmed).

If you don’t want to miss it, click here to register.