Synthetic Data Reveals Generalization Gaps in Correlated Multiple Instance Learning
Published in ML4H 2025 Symposium, Findings Track, 2025
We evaluate how well current MIL models capture inter-instance correlations by generating synthetic datasets and constructing a Bayes estimator as an optimal upper bound on model performance.
Recommended citation: Ethan Harvey, Dennis Johan Loevlie, and Michael C. Hughes. (2025). "Synthetic Data Reveals Generalization Gaps in Correlated Multiple Instance Learning." ML4H 2025 Symposium, Findings Track.
Download Paper
