Autoassociative Learning of Structural Representations for Modeling and Classification in Medical Imaging

TL;DR

This paper introduces neurosymbolic systems that learn structural representations by reconstructing images with visual primitives, improving accuracy and transparency in medical image classification compared to traditional deep learning.

Contribution

It presents a novel neurosymbolic approach that emphasizes structural explanations over smooth features for medical imaging tasks.

Findings

Outperforms conventional deep learning in accuracy for histological diagnosis

Provides more transparent and interpretable classification results

Demonstrates effectiveness in modeling well-defined categorical objects

Abstract

Deep learning architectures based on convolutional neural networks tend to rely on continuous, smooth features. While this characteristics provides significant robustness and proves useful in many real-world tasks, it is strikingly incompatible with the physical characteristic of the world, which, at the scale in which humans operate, comprises crisp objects, typically representing well-defined categories. This study proposes a class of neurosymbolic systems that learn by reconstructing images in terms of visual primitives and are thus forced to form high-level, structural explanations of them. When applied to the task of diagnosing abnormalities in histological imaging, the method proved superior to a conventional deep learning architecture in terms of classification accuracy, while being more transparent.