GCE-MIL: Faithful and Recoverable Evidence for Multiple Instance Learning in Whole-Slide Imaging

In three linesGCE-MIL improves multiple instance learning for whole-slide image analysis by directly optimizing evidence quality (sufficiency, necessity, recoverability) instead of relying on attention weights. Across 81 configurations (9 backbones, 9 datasets), Macro-F1 gains +0.024 and C-index +0.014, with 5× faster inference.

Summary generated by Claude — human-verified