Validating Automatic Concept-Based Explanations for AI-Based Digital Histopathology

Abstract

Digital histopathology poses several challenges such as label noise, class imbalance, limited availability of labelled data, and several latent biases to deep learning, negatively influencing transparency, reproducibility, and classification performance. In particular, biases are well known to cause poor generalization. Proposed tools from explainable artificial intelligence (XAI), bias detection, and bias discovery suffer from technical challenges, complexity, unintuitive usage, inherent biases, or a semantic gap. A promising XAI method, not studied in the context of digital histopathology is automated concept-based explanation (ACE). It automatically extracts visual concepts from image data. Our objective is to evaluate ACE’s technical validity following design science principals and to compare it to Guided Gradient-weighted Class Activation Mapping (Grad-CAM), a conventional pixel-wise explanation method. To that extent, we created and studied five convolutional neural networks (CNNs) in four different skin cancer settings. Our results demonstrate that ACE is a valid tool for gaining insights into the decision process of histopathological CNNs that can go beyond explanations from the control method. ACE validly visualized a class sampling ratio bias, measurement bias, sampling bias, and class-correlated bias. Furthermore, the complementary use with Guided Grad-CAM offers several benefits. Finally, we propose practical solutions for several technical challenges. In contradiction to results from the literature, we noticed lower intuitiveness in some dermatopathology scenarios as compared to concept-based explanations on real-world images.

Daniel Sauter
Daniel Sauter
Researcher in the first cohort

My research interests include Deep Learning, Computer Vision, Radiomics, and Explainable AI.

Georg C. Lodde
Georg C. Lodde
Clinician Scientist

My research interests include Dermatology, Medical Research, and Digitalization.

Felix Nensa
Felix Nensa
Speaker

My research interests include medical digitalization, computer vision and radiology.

Dirk Schadendorf
Dirk Schadendorf
Principal Investigator

My research interests include Dermatology, Medical Research, and Digitalization.

Elisabeth Livingstone
Elisabeth Livingstone
Principal Investigator

My research interests include Medical Research, Dermatology, and Digitalization.

Markus Kukuk
Markus Kukuk
Principal Investigator

My research interests include Deep Learning and Computer Vision.

Next
Previous