Parity Partition Coding for Sharp Multi-Label Classification

TL;DR

This paper introduces parity partition coding, a novel error correcting output code technique that improves multi-label classification efficiency and accuracy by reducing the number of outputs and increasing sharpness in image classifiers.

Contribution

It proposes parity partition coding, a new coding method that enhances sharpness and reduces complexity in multi-label classification tasks.

Findings

Outperforms baseline models on MultiMNIST and CelebA datasets.

Requires fewer parameters while achieving higher accuracy.

Increases sharpness by effectively partitioning categories.

Abstract

The problem of efficiently training and evaluating image classifiers that can distinguish between a large number of object categories is considered. A novel metric, sharpness, is proposed which is defined as the fraction of object categories that are above a threshold accuracy. To estimate sharpness (along with a confidence value), a technique called fraction-accurate estimation is introduced which samples categories and samples instances from these categories. In addition, a technique called parity partition coding, a special type of error correcting output code, is introduced, increasing sharpness, while reducing the multi-class problem to a multi-label one with exponentially fewer outputs. We demonstrate that this approach outperforms the baseline model for both MultiMNIST and CelebA, while requiring fewer parameters and exceeding state of the art accuracy on individual labels.