Join the next RISE-MICCAI Tutorial - March 3, 2026
Monday 23rd February 2026
Join the next RISE-MICCAI Tutorial:
Multiple Instance Learning for Histopathology: An Introduction to torchmil
Presenter: Francisco M. Castro-Macías, University of Granada, Creator and Lead Developer of torchmil
Tuesday, March 3, 2026 at 12:00 pm EST / 6:00 pm CET
Multiple Instance Learning (MIL) has established itself as a powerful framework for weakly supervised learning, particularly in scenarios where fine-grained annotations are scarce or unavailable. In the medical domain, MIL has become the de facto framework for analyzing gigapixel Whole Slide Images (WSIs) for tasks such as cancer grading and biomarker discovery.
Recently, torchmil has emerged as a comprehensive PyTorch library designed to accelerate research and lower the entry barrier for new users in this field. This tutorial is divided into two parts. The first part provides a theoretical foundation of MIL, focusing specifically on its application to computational pathology. The second part introduces the core functionalities of the torchmil library. Participants will learn how to efficiently handle pathology slides, train state-of-the-art MIL models, and perform inference on new data using torchmil.
Speaker:
Francisco M. Castro-Macías is the creator and lead developer of torchmil, a PyTorch library designed to democratize Multiple Instance Learning for histopathology. His work has appeared in leading machine learning venues, including NeurIPS, MICCAI, and Pattern Recognition, and he was awarded the Best Student Paper Award at IEEE ICIP 2024. Additionally, he has served as a reviewer for top-tier conferences and journals such as NeurIPS, ICML, and TMLR.
He received a Bachelor's degree in Mathematics, a Bachelor's degree in Computer Science, and a Master's degree in Data Science and Computer Engineering from the University of Granada, Spain. As of January 2026, he is pursuing his Ph.D. under the supervision of Prof. Rafael Molina (University of Granada), Prof. Pablo Morales-Álvarez (University of Granada), and Prof. Aggelos K. Katsaggelos (Northwestern University). His doctoral research is supported by the FPU scholarship from the Spanish Ministry of Science, Innovation, and Universities.
His research focuses on developing probabilistic models for inverse and weakly supervised problems, with a particular emphasis on medical applications. He has completed research stays at the University of Cambridge, hosted by Prof. José Miguel Hernández-Lobato, and at Northwestern University, hosted by Prof. Aggelos K. Katsaggelos.