Investigate your ML model's prediction 🔬(Diabetic Retinopathy diagnosis use case)
1) What is XAI?
2) Why do we need it?
3) Some XAI tools and when to use them.
4) What’s DR? How do we diagnose it?
5) Build your DR diagnosis model.
7) Understand your model’s prediction using XAI libraries like GradCAM.
8) Conclusion and future perspectives.
Please note that this presentation will focus more on the explainability part than on model creation.
Learn how to interpret your model's prediction by understanding what you're looking for and using the XAI (i.e. Explainable AI) tools to investigate the areas (for computer vision applications) your model pays attention to, to provide you with the prediction.
Sara EL-ATEIF is a Google Ph.D. Fellow/Candidate at ENSIAS, UM5R Morocco, and NVIDIA DLI University Ambassador/Instructor. She is part of the Diversity & Inclusion group of the Python Software Foundation, Lead of the TFUG Casablanca chapter as well organizer at GDG & WTM Casablanca/El Jadida. She likes to contribute to AI4Good projects using her machine learning knowledge to create a positive impact on society. Previously, as a Machine Learning Pathway Mentor at STEM-Away, she got to actively help enthusiasts get started in machine learning and draft the bio-ML and NLP virtual internship projects. She is also involved (and won) in several competitions/challenges to improve her skills with the AI Wonder Girls.