AEM: Attention Entropy Maximization for Multiple Instance Learning based Whole Slide Image Classification

TL;DR

This paper introduces Attention Entropy Maximization (AEM), a regularization method for MIL-based whole slide image classification that improves model performance by encouraging attention diversity and reducing overfitting.

Contribution

The paper proposes a simple regularization technique, AEM, that enhances MIL models by increasing attention entropy, with a novel cosine weight annealing to reduce parameter sensitivity.

Findings

AEM improves performance across various models and techniques.

Attention entropy correlates positively with model accuracy.

AEM outperforms existing methods in experiments.

Abstract

Multiple Instance Learning (MIL) effectively analyzes whole slide images but faces overfitting due to attention over-concentration. While existing solutions rely on complex architectural modifications or additional processing steps, we introduce Attention Entropy Maximization (AEM), a simple yet effective regularization technique. Our investigation reveals the positive correlation between attention entropy and model performance. Building on this insight, we integrate AEM regularization into the MIL framework to penalize excessive attention concentration. To address sensitivity to the AEM weight parameter, we implement Cosine Weight Annealing, reducing parameter dependency. Extensive evaluations demonstrate AEM's superior performance across diverse feature extractors, MIL frameworks, attention mechanisms, and augmentation techniques. Here is our anonymous code: https://github.com/dazhangyu123/AEM.