Causality in
Medical Image Computing
An official MICCAI 2020 Satellite Event - October 8
About
This MICCAI 2020 half-day tutorial introduces fundamental concepts of causality and its role in medical imaging. We will illustrate how causal reasoning provides a fresh perspective on important topics including key challenges in image-based predictive modelling such as generalization, data scarcity, confounding, robustness, reliability and responsible reporting. Theoretical concepts will be introduced and related to real-world examples from medical imaging such as image classification for computer-aided diagnosis.
The goal of the tutorial is to raise awareness of the importance of taking causal considerations into account when conducting machine learning research for medical imaging. We hope that this tutorial can provide new inputs to the community and may lay the path to exciting new research directions in medical image computing.
Videos
Topics
Introduction to causality
Causality in medical imaging and machine learning
Dataset mismatch and generalization
Learning strategies with limited data
Biases, confounding, data harmonization
Responsible reporting
Counterfactuals
KEYNOTE & PANELIST
Prof Elias Bareinboim
Department of Computer Science
Director, Causal Artificial Intelligence Lab
Columbia University
PANELIST
Prof Suchi Saria
Department of Computer Science
Department of Applied Math & Statistics
Department of Health Policy & Management
Johns Hopkins University
Learning objectives
To use the language of causality to formalize key challenges in machine learning for medical imaging
To apply causal reasoning to identify solutions towards robustness and reliability
To use causality to identify issues such as sampling biases, confounding, and domain shift
To understand the effect of causal relationships on different learning strategies
SCHEDULE
11:00 - 12:15 UTC - Session 1: Introduction to Causality
12:15 - 13:30 UTC - Session 2: Causality in Medical Imaging
12:15 UTC - Dataset shift and sampling selection bias by Daniel Castro and Ben Glocker (slides, video)
12:45 UTC - Counterfactuals in Medical Imaging by Nick Pawlowski (slides, video)
13:00 UTC - Explanation by Exaggeration by Kayhan Batmanghelich (slides, video)
13:15 UTC - Questions & Answers (video)
13:30 - 15:15 UTC - Session 3: Causality in Machine Learning
Organizers
Ben Glocker, Imperial College London
Kristin Linn, University of Pennsylvania
Daniel Coelho de Castro, Imperial College London
Nick Pawlowski, Imperial College London
Kayhan Batmanghelich, University of Pittsburgh