This Looks Like That: Interpretable Neural Networks for Medical Image Analysis

The use of deep neural networks has become increasingly popular for computer vision tasks. These models have the potential to achieve great accuracy in recognizing images, but they are often called “black boxes” because they generally suffer from a lack of “interpretability.” In fact, evidence has emerged, where some deep learning methods appeared to perform well, but based their decisions on confounding rather than truly relevant information.

This talk will focus on a neural network known as a prototypical part network (ProtoPNet). This network classifies an input image by comparing various parts of the image with learned features known as prototypes. In addition, the talk will show examples of medical image analysis using a prototypical case-based interpretable neural network model, with a focus on mammograms. Such an interpretable model will automatically detect areas of anomaly (such as spiculated mass margins) and present prototypical examples of similar anomalies found in other patients as explanations for their predictions. The talk will discuss training the neural network using a human-in-the-loop training scheme. Training is based on a small set of doctor-annotated images and is specifically designed to prevent the model from using confounding information. The explanations generated by our interpretable neural network for analyzing mammograms are useful because they augment the reports generated by radiologists, and may be able to assist doctors in diagnosing patients in the future.

Speaker(s): Chaofan Chen,

Virtual: https://events.vtools.ieee.org/m/345498