Label-Descriptive Patterns and Their Application to Characterizing Classification Errors
This work addresses the need for interpretable error analysis in machine learning, particularly for NLP classifiers, offering a way to identify systematic errors and improve model performance, though it is incremental as it builds on existing label description methods.
The authors tackled the problem of characterizing classification errors in interpretable terms by discovering feature-value patterns that correlate with correct or erroneous predictions, and they developed the Premise algorithm which effectively recovers ground truth patterns and provides actionable insights in case studies on Visual Question Answering and Named Entity Recognition.
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.