A simple squared-error reformulation for ordinal classification

TL;DR

This paper introduces a straightforward squared-error reformulation for ordinal classification in deep neural networks, leveraging a softmax layer to improve sensitivity to class order and enable probabilistic outputs, demonstrating superior performance on a medical dataset.

Contribution

It proposes a novel squared-error based loss function with a softmax layer for ordinal classification, enhancing class order sensitivity and probabilistic modeling.

Findings

Outperforms baseline methods on diabetic retinopathy dataset

Demonstrates effectiveness of the squared-error reformulation in ordinal tasks

Validates approach with empirical results on high-resolution medical data

Abstract

In this paper, we explore ordinal classification (in the context of deep neural networks) through a simple modification of the squared error loss which not only allows it to not only be sensitive to class ordering, but also allows the possibility of having a discrete probability distribution over the classes. Our formulation is based on the use of a softmax hidden layer, which has received relatively little attention in the literature. We empirically evaluate its performance on the Kaggle diabetic retinopathy dataset, an ordinal and high-resolution dataset and show that it outperforms all of the baselines employed.