Identifying Subgroups in Biomedical Datasets Using Data Attribution

Learn how to apply data attribution methods to identify meaningful subgroups in biomedical datasets through this comprehensive tutorial presented by MIT's Djuna von Maydell. Explore the datamodels attribution method and its practical applications in life sciences, starting with basic image classification examples using CIFAR10. Master efficient strategies for training multiple lightweight classifiers while focusing on biomedical applications, particularly in single-cell and genetic datasets. Gain hands-on experience through interactive Google Colab sessions designed for all skill levels, from machine learning beginners to experienced practitioners interested in biomedical applications. Access accompanying materials including detailed slides, a GitHub repository with tutorial code, and Colab notebooks with and without outputs to enhance the learning experience and practical implementation of the techniques covered.