Divide et impera: hybrid multinomial classifiers from quantum binary models

TL;DR

This paper explores hybrid quantum multinomial classifiers using strategies like one-vs-one, one-vs-rest, and decision trees, highlighting their efficiency and quantum advantage in classification tasks.

Contribution

It introduces a hybrid approach to combine quantum binary models into multinomial classifiers and benchmarks their efficiency and accuracy.

Findings

Decision tree approach offers a cost-effective solution.

Achieves similar accuracy to other methods with minimal overhead.

Overhead is at most logarithmic in the number of classes.

Abstract

We investigate how to combine a collection of quantum binary models into a multinomial classifier. We employ a hybrid approach, adopting strategies like one-vs-one, one-vs-rest and a binary decision tree. We benchmark each method, by emphasizing their computational overhead and their impact on the quantum advantage. By comparison against a classical binary model (generalized using the same approach), we show that the decision tree represents a cost-effective solution, achieving similar accuracies to other methods with an overhead at most logarithmic in the total number of classes.