Improving the Estimation of Word Importance for News Multi-Document Summarization

By Kai Hong, Ani Nenkova. 2014

165 citations recorded by the source corpus.

In this paper, we propose a supervised model for ranking word importance that incorporates a rich set of features. Our model is superior to prior approaches for identifying words used in human summaries. Moreover we show that an extractive summarizer which includes our estimation of word importance results in summaries comparable with the state-of-the-art by automatic evaluation. Disciplines Computer Engineering | Computer Sciences Comments University of Pennsylvania Department of Computer and Information Science Technical Report No. MSCIS-14-02. This technical report is available at ScholarlyCommons: http://repository.upenn.edu/cis_reports/989 Improving the Estimation of Word Importance for News Multi-Document Summarization Extended Technical Report Kai Hong University of Pennsylvania Philadelphia, PA, 19104 hongkai1@seas.upenn.edu Ani Nenkova University of Pennsylvania Philadelphia, PA, 19104 nenkova@seas.upenn.edu

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

MLSUM: The Multilingual Summarization Corpus (2020, methodology): “Although (Grusky et al., 2018) found indications of bias towards extractive approaches, several research efforts have used this dataset for summarization (Hong and Nenkova, 2014; Durrett et al., 2016; Paulus et al., 2017).”
Newsroom: A Dataset of 1.3 Million Summaries with Diverse Extractive Strategies (2018, methodology): “Despite this, limited work has used this dataset for summarization (Hong and Nenkova, 2014; Durrett et al., 2016; Paulus et al., 2017).”
Neural Text Summarization: A Critical Evaluation (2019, methodology): “It has been primarily used for extractive summarization (Hong and Nenkova, 2014; Li et al., 2016) and phrase-importance prediction (Yang and Nenkova, 2014; Nye and Nenkova, 2015).”
Deep Communicating Agents for Abstractive Summarization (2018, methodology): “New York Times (NYT): Although this dataset has mainly been used to train extractive summarization systems (Hong and Nenkova, 2014; Hong et al., 2015; Li et al., 2016; Durrett et al., 2016), it has recently been used for the abstractive summarization task (Paulus et al., 2018).”
Learning-Based Single-Document Summarization with Compression and Anaphoricity Constraints (2016, background): “This corpus has been explored in related contexts (Dunietz and Gillick, 2014; Hong and Nenkova, 2014), but to our knowledge it has not been directly used for singledocument summarization.”
WikiHow: A Large Scale Text Summarization Dataset (2018, methodology): “While this dataset has been mainly used for extractive systems (Hong and Nenkova, 2014; Durrett et al., 2016), Paulus et al.”
Content Selection in Deep Learning Models of Summarization (2018, background): “Non-neural network approaches often use frequency and information theoretic measures as proxies for content salience (Hong and Nenkova, 2014), but these are not explicitly used in most neural network summarization systems.”
System Combination for Multi-document Summarization (2015, methodology): “For example, we design features that estimate the intrinsic importance of words from a large corpus (Hong and Nenkova, 2014).”
Intrinsic Evaluation of Summarization Datasets (2020, background): “Conversely, NYT has been used to build extractive systems (Hong and Nenkova, 2014; Li et al., 2016).”
Recent advances in document summarization (2017, background): “Given sentences with labeled importance scores, it is straightforward to train regression models for importance prediction [53,72,209] or learning to rank models to train a model that is capable to assign high rank for the most important sentences [132,169,190].”
Graph-based Neural Multi-Document Summarization (2017, background): “, 2004a), word frequency and importance (Nenkova et al., 2006; Hong and Nenkova, 2014), among others, to rank sentence salience.”
From Extractive to Abstractive Summarization: A Journey (2019, background): “RegSum uses diverse features like parts of speech tags, named entity tags, locations and categories for supervised prediction of word importance[33].”
Objective Function Learning to Match Human Judgements for Optimization-Based Summarization (2018, background): “, 2015) or n-gram (Hong and Nenkova, 2014; Li et al., 2013) scores and then defining a composition function to get a score for the summary.”
Identiﬁcation and Characterization of Newsworthy Verbs in World News (background): “Such ideas have been explored in the past as subcomponents of extractive summarizers (Schiffman et al., 2002; Hong and Nenkova, 2014) or as features derived from small datasets for sentence compression (Woodsend and Lapata, 2012).”