A model pretrained on synthetic bag-structured data performs in-context learning for new MIL tasks from a handful of examples and outperforms task-specific supervised baselines on twelve benchmarks.
Kömen, E
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
verdicts
UNVERDICTED 6representative citing papers
DaX is a pathology vision foundation model that extends DINOv3 with continuous magnification training and cross-scale consistency, achieving top average performance on a benchmark of 161 tasks from 44 datasets covering 28k patients.
Symb-xMIL is a post-hoc explanation framework that quantifies MIL model alignment with logical decision rules in histopathology to enable rule-based interpretability.
DECAT classifies multimodal representations into four diagnostic scenarios using null-referenced metrics and a rule-based procedure to detect shared biology versus confounders without knowing the confounder identity.
ReshapeOT improves optimal transport reliability for distribution shifts by replacing the Euclidean ground metric with a Mahalanobis distance derived from observed displacement second moments.
CFKD generates counterfactuals for human-guided correction of Clever Hans predictors in image classifiers, removing the need for confounder labels while scaling to multiple spurious correlations.
citing papers explorer
-
Symb-xMIL: Symbolic Explanations for Multiple Instance Learning in Digital Pathology
Symb-xMIL is a post-hoc explanation framework that quantifies MIL model alignment with logical decision rules in histopathology to enable rule-based interpretability.