Benchmarking PNW Model for MedMNIST to 100% Accuracy

TL;DR

This paper presents a benchmarking approach for the PNW model on MedMNIST datasets, achieving near-perfect accuracy on most datasets by training error-free models.

Contribution

It introduces the concept of Artificial Special Intelligence and demonstrates error-free training on multiple biomedical datasets, surpassing previous accuracy levels.

Findings

All datasets except three with double-labeling were trained to perfection.

The method achieves 100% accuracy on most MedMNIST datasets.

Error-free training indicates potential for highly reliable biomedical classification models.

Abstract

In this paper, we introduce a new concept called Artificial Special Intelligence by which Machine Learning models for the classification problem can be trained error-free, thus acquiring the capability of not making repeated mistakes. The method is applied to 18 MedMNIST biomedical datasets. Except for three datasets, which suffer from the double-labeling problem, all are trained to perfection.