BIGPATENT: A Large-Scale Dataset for Abstractive and Coherent Summarization

TL;DR

BIGPATENT is a large-scale dataset of U.S. patent documents with human-written abstractive summaries that feature richer discourse, more even content distribution, and less extractive overlap, aiming to advance summarization research.

Contribution

The paper introduces BIGPATENT, a novel dataset with unique properties to improve understanding and generation of abstractive summaries in diverse discourse structures.

Findings

Baseline models perform poorly on the dataset.

Summarization models struggle with the complex discourse structure.

The dataset reveals new challenges for abstractive summarization.

Abstract

Most existing text summarization datasets are compiled from the news domain, where summaries have a flattened discourse structure. In such datasets, summary-worthy content often appears in the beginning of input articles. Moreover, large segments from input articles are present verbatim in their respective summaries. These issues impede the learning and evaluation of systems that can understand an article's global content structure as well as produce abstractive summaries with high compression ratio. In this work, we present a novel dataset, BIGPATENT, consisting of 1.3 million records of U.S. patent documents along with human written abstractive summaries. Compared to existing summarization datasets, BIGPATENT has the following properties: i) summaries contain a richer discourse structure with more recurring entities, ii) salient content is evenly distributed in the input, and iii) lesser and shorter extractive fragments are present in the summaries. Finally, we train and evaluate baselines and popular learning models on BIGPATENT to shed light on new challenges and motivate future directions for summarization research.