A Supervised Approach to Extractive Summarisation of Scientific Papers

By Edward Collins, Isabelle Augenstein, Sebastian Riedel. 2017

110 citations recorded by the source corpus.

Automatic summarisation is a popular approach to reduce a document to its main arguments. Recent research in the area has focused on neural approaches to summarisation, which can be very data-hungry. However, few large datasets exist and none for the traditionally popular domain of scientific publications, which opens up challenging research avenues centered on encoding large, complex documents. In this paper, we introduce a new dataset for summarisation of computer science publications by exploiting a large resource of author provided summaries and show straightforward ways of extending it further. We develop models on the dataset making use of both neural sentence encoding and traditionally used summarisation features and show that models which encode sentences as well as their local and global context perform best, significantly outperforming well-established baseline methods.

Subfields: Artificial Intelligence, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

Citations in the corpus

Recursively Summarizing Books with Human Feedback (2021, background): “…novels is sparse, there has been plenty of work on summarizing other kinds of long documents, such as scientiﬁc papers (Abu-Jbara and Radev, 2011; Collins et al., 2017; Subramanian et al., 2019; Cohan et al., 2018; Xiao and Carenini, 2019; Zhao et al., 2020; Sotudeh et al., 2020), and patents…”
Data-driven Summarization of Scientific Articles (2018, methodology): “In (Collins et al., 2017), they develop a supervised extractive summarization framework which they apply to a dataset of 10k computer science papers.”
On Generating Extended Summaries of Long Documents (2020, background): “Few prior studies have noted the importance of documents’ structure in shorter-form summary generation (Collins, Augenstein, and Riedel 2017; Cohan et al. 2018).”
A Divide-and-Conquer Approach to the Summarization of Long Documents (2020, background): “This is usually approached as a binary classiﬁcation problem, where for each sentence the model decides whether it should be included in the summary or not [6], [32], [33].”
Extractive Summarization of Long Documents by Combining Global and Local Context (2019, background): “For instance, (Teufel and Moens, 2002) (Collins et al., 2017) where, in order to determine whether a sentence should be included in the summary, they directly use the section each sentence appears in as a categorical feature with values like Highlight, Abstract, Introduction, etc..”
TLDR: Extreme Summarization of Scientific Documents (2020, methodology): “CLPubSum (Collins et al., 2017) 10.”
GUIR @ LongSumm 2020: Learning to Generate Long Summaries from Scientific Documents (2020, methodology): “Inspired by the fact that discourse information is of high importance when dealing with long documents (Conroy and Davis, 2017; Collins et al., 2017; Cohan et al., 2018) in scientific research papers, in this work, we step on utilizing such structure in summarization of scientific papers.”
A Divide-and-Conquer Approach to the Summarization of Academic Articles (2020, methodology): “Recently, summarization methods have started to focus on summarizing longer documents, such as academic articles [12, 13, 14, 15, 16].”
BillSum: A Corpus for Automatic Summarization of US Legislation (2019, methodology): “…summarization has been conducted in a variety of domains, such as news articles (Hermann et al., 2015), emails (Nenkova and Bagga, 2004), scientiﬁc papers (Teufel and Moens, 2002; Collins et al., 2017), and court pro-ceedings (Grover et al., 2004; Saravanan et al., 2008; Kim et al., 2013).”