Paying down metadata debt: learning the representation of concepts using topic models

TL;DR

This paper addresses metadata debt by proposing a semisupervised topic modeling approach that learns meaningful concept representations, improves interpretability, and aids in label prediction and error detection in large datasets.

Contribution

It introduces a novel semisupervised topic model with low-rank matrix factorization and gauge transformation for explicit concept-label associations and improved interpretability.

Findings

Successfully predicts subject tags on 25,000 datasets from Kaggle.

Learns semantically meaningful features that enhance data annotation.

Enables error detection and missing feature prediction in metadata.

Abstract

We introduce a data management problem called metadata debt, to identify the mapping between data concepts and their logical representations. We describe how this mapping can be learned using semisupervised topic models based on low-rank matrix factorizations that account for missing and noisy labels, coupled with sparsity penalties to improve localization and interpretability. We introduce a gauge transformation approach that allows us to construct explicit associations between topics and concept labels, and thus assign meaning to topics. We also show how to use this topic model for semisupervised learning tasks like extrapolating from known labels, evaluating possible errors in existing labels, and predicting missing features. We show results from this topic model in predicting subject tags on over 25,000 datasets from Kaggle.com, demonstrating the ability to learn semantically meaningful features.