Label-Descriptive Patterns and Their Application to Characterizing Classification Errors

TL;DR

This paper introduces a method to identify interpretable feature-value patterns that correlate with classifier errors, providing insights into systematic mistakes in deep learning models.

Contribution

It formulates the pattern discovery as a label description problem using the MDL principle and develops the Premise algorithm for efficient pattern set discovery.

Findings

Premise effectively recovers ground truth patterns in synthetic data.

It performs well on real-world data, even with class imbalance.

Provides actionable insights into classifier errors in NLP tasks.

Abstract

State-of-the-art deep learning methods achieve human-like performance on many tasks, but make errors nevertheless. Characterizing these errors in easily interpretable terms gives insight into whether a classifier is prone to making systematic errors, but also gives a way to act and improve the classifier. We propose to discover those feature-value combinations (i.e., patterns) that strongly correlate with correct resp. erroneous predictions to obtain a global and interpretable description for arbitrary classifiers. We show this is an instance of the more general label description problem, which we formulate in terms of the Minimum Description Length principle. To discover a good pattern set, we develop the efficient Premise algorithm. Through an extensive set of experiments we show it performs very well in practice on both synthetic and real-world data. Unlike existing solutions, it ably recovers ground truth patterns, even on highly imbalanced data over many features. Through two case studies on Visual Question Answering and Named Entity Recognition, we confirm that Premise gives clear and actionable insight into the systematic errors made by modern NLP classifiers.