Structured Summarization: Unified Text Segmentation and Segment Labeling as a Generation Task

TL;DR

This paper introduces a unified neural generation approach for text segmentation and labeling, achieving state-of-the-art results across document and conversation datasets without domain-specific tuning.

Contribution

It presents a novel encoder-decoder model that treats segmentation and labeling as a single generation task, simplifying and unifying previous methods.

Findings

Achieves state-of-the-art performance on multiple datasets.

Works effectively in both high- and low-resource settings.

Demonstrates the feasibility of unified generation for segmentation and labeling.

Abstract

Text segmentation aims to divide text into contiguous, semantically coherent segments, while segment labeling deals with producing labels for each segment. Past work has shown success in tackling segmentation and labeling for documents and conversations. This has been possible with a combination of task-specific pipelines, supervised and unsupervised learning objectives. In this work, we propose a single encoder-decoder neural network that can handle long documents and conversations, trained simultaneously for both segmentation and segment labeling using only standard supervision. We successfully show a way to solve the combined task as a pure generation task, which we refer to as structured summarization. We apply the same technique to both document and conversational data, and we show state of the art performance across datasets for both segmentation and labeling, under both high- and low-resource settings. Our results establish a strong case for considering text segmentation and segment labeling as a whole, and moving towards general-purpose techniques that don't depend on domain expertise or task-specific components.