Extractive Text Summarization Using Generalized Additive Models with Interactions for Sentence Selection

TL;DR

This paper explores the use of interpretable Generalized Additive Models with interactions, specifically Explainable Boosting Machine and GAMI-Net, for extractive text summarization based on linguistic features, aiming to improve interpretability over dense models.

Contribution

It introduces the application of modern interpretable GAMs with interactions to extractive text summarization, addressing the interpretability challenge in machine learning-based ATS.

Findings

Demonstrates the feasibility of using GAMs for sentence selection in summarization.

Shows that GAMs can achieve competitive performance with better interpretability.

Provides insights into feature importance and interactions in summarization tasks.

Abstract

Automatic Text Summarization (ATS) is becoming relevant with the growth of textual data; however, with the popularization of public large-scale datasets, some recent machine learning approaches have focused on dense models and architectures that, despite producing notable results, usually turn out in models difficult to interpret. Given the challenge behind interpretable learning-based text summarization and the importance it may have for evolving the current state of the ATS field, this work studies the application of two modern Generalized Additive Models with interactions, namely Explainable Boosting Machine and GAMI-Net, to the extractive summarization problem based on linguistic features and binary classification.