Toward Aristotelian Medical Representations: Backpropagation-Free Layer-wise Analysis for Interpretable Generalized Metric Learning on MedMNIST

TL;DR

This paper introduces A-ROM, a new interpretable medical imaging model that uses a concept dictionary and kNN classifier, avoiding backpropagation and enabling rapid, transparent concept learning on MedMNIST.

Contribution

It proposes a backpropagation-free, interpretable framework based on Vision Transformers and a concept dictionary for medical image classification.

Findings

A-ROM achieves competitive performance on MedMNIST v2.

The model provides transparent, human-readable reasoning.

It enables few-shot learning without gradient-based fine-tuning.

Abstract

While deep learning has achieved remarkable success in medical imaging, the "black-box" nature of backpropagation-based models remains a significant barrier to clinical adoption. To bridge this gap, we propose Aristotelian Rapid Object Modeling (A-ROM), a framework built upon the Platonic Representation Hypothesis (PRH). This hypothesis posits that models trained on vast, diverse datasets converge toward a universal and objective representation of reality. By leveraging the generalizable metric space of pretrained Vision Transformers (ViTs), A-ROM enables the rapid modeling of novel medical concepts without the computational burden or opacity of further gradient-based fine-tuning. We replace traditional, opaque decision layers with a human-readable concept dictionary and a k-Nearest Neighbors (kNN) classifier to ensure the model's logic remains interpretable. Experiments on the MedMNIST v2 suite demonstrate that A-ROM delivers performance competitive with standard benchmarks while providing a simple and scalable, "few-shot" solution that meets the rigorous transparency demands of modern clinical environments.