Extreme Multi-label Classification from Aggregated Labels

TL;DR

This paper introduces a scalable algorithm for extreme multi-label classification when labels are only available at the group level, enabling effective label imputation and improved performance on large-scale MIML tasks.

Contribution

It presents a novel scalable algorithm for imputing individual labels from group labels in XMC, extending MIML frameworks, and demonstrating superior results over existing methods.

Findings

The algorithm effectively imputes individual labels from group labels.

The approach scales to large XMC and MIML datasets.

Experimental results show improved accuracy over existing methods.

Abstract

Extreme multi-label classification (XMC) is the problem of finding the relevant labels for an input, from a very large universe of possible labels. We consider XMC in the setting where labels are available only for groups of samples - but not for individual ones. Current XMC approaches are not built for such multi-instance multi-label (MIML) training data, and MIML approaches do not scale to XMC sizes. We develop a new and scalable algorithm to impute individual-sample labels from the group labels; this can be paired with any existing XMC method to solve the aggregated label problem. We characterize the statistical properties of our algorithm under mild assumptions, and provide a new end-to-end framework for MIML as an extension. Experiments on both aggregated label XMC and MIML tasks show the advantages over existing approaches.