A trainable document summarizer

By Julian Kupiec, Jan O. Pedersen, Francine R. Chen. 1995

●● ● ● ● To summarize is to reduce in complexity, and hence in length, while retaining some of the essential qualities of the original. This paper focusses on document extracts, a particular kind of computed document summary. Document extracts consisting of roughly 20% of the original cart be as informative as the full text of a document, which suggests that even shorter extracts may be useful indicative summmies. The trends in our results are in agreement with those of Edmundson who used a subjectively weighted combination of features as opposed to training the feature weights using a corpus.

Subfields: 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 Natural Language Processing

Citations in the corpus

Get To The Point: Summarization with Pointer-Generator Networks (2017, background): “Due to the difficulty of abstractive summarization, the great majority of past work has been extractive (Kupiec et al., 1995; Paice, 1990; Saggion and Poibeau, 2013).”
Neural Summarization by Extracting Sentences and Words (2016, methodology): “Several methods have been used in order to select the summary sentences ranging from binary classiﬁers (Kupiec et al., 1995), to hidden Markov models (Conroy and O’Leary, 2001), graph-based algorithms (Erkan and Radev, 2004; Mihalcea, 2005), and integer linear programming (Woodsend and Lapata,…”
LexRank: Graph-based Lexical Centrality as Salience in Text Summarization (2004, neutral): “Researchers have also tried to integrate machine learning into summarization as more features have been proposed and more training data have become available (Kupiec, Pedersen, & Chen, 1995; Lin, 1999; Osborne, 2002; Daum´e III & Marcu, 2004).”
Hedge Trimmer: A Parse-and-Trim Approach to Headline Generation (2003, background): “, sentence extraction (Edmundson, 1969; Johnson et al, 1993; Kupiec et al., 1995; Mann et al., 1992; Teufel and Moens, 1997; Zechner, 1995), processing of structured templates (Paice and Jones, 1993), sentence compression (Hori et al.”
Neural Document Summarization by Jointly Learning to Score and Select Sentences (2018, methodology): “Kupiec et al. (1995) use a Naive Bayes classiﬁer to learn feature combinations.”
Content Selection in Deep Learning Models of Summarization (2018, background): “Summarization of academic literature (including medical journals), has long been a research topic in NLP (Kupiec et al., 1995; Elhadad et al., 2005), but most approaches have explored facet-based summarization (Jaidka et al., 2017), which is not the focus of our work.”
Neural Latent Extractive Document Summarization (2018, background): “Early attempts mostly leverage human-engineered features (Filatova and Hatzivassiloglou, 2004) coupled with binary classiﬁers ( Kupiec et al., 1995), hidden Markov models (Conroy and O’leary, 2001), graph based methods (Mihalcea, 2005), and inte-ger linear programming (Woodsend and Lapata, 2010).”
HIBERT: Document Level Pre-training of Hierarchical Bidirectional Transformers for Document Summarization (2019, methodology): “It is usually modeled as a sentence ranking problem by using the scores from classiﬁers (Kupiec et al., 1995), sequential labeling models (Conroy and O’leary, 2001) as well as integer linear programmers (Woodsend and Lapata, 2010).”
Objective Function Learning to Match Human Judgements for Optimization-Based Summarization (2018, methodology): “First, the choice of the ob-jective function is based on ad-hoc assumptions about which quality aspects of a summary are relevant (Kupiec et al., 1995).”
Automatic Text Summarization Using a Machine Learning Approach (2002, methodology): “In this work we have explored the framework of using a ML approach to produce trainable text summarizers, in a way which was proposed a few years ago by Kupiec [7].”
Enhancing Single-Document Summarization by Combining RankNet and Third-Party Sources (2007, methodology): “The meth-ods include binary classiﬁers (Kupiec et al., 1995), Markov models (Conroy et al., 2004), Bayesian methods (Daum´e III and Marcu, 2005; Aone et al., 1998), and heuristic methods to determine feature weights (Schiffman, 2002; Lin and Hovy, 2002).”
Left-Brain/Right-Brain Multi-Document Summarization (2004, methodology): “An HMM, in contrast to a naive Bayesian approach ([5], [1]), allows the probability that sentence i is in the summary to be dependent on whether sentence i− 1 is in the summary.”
Learning to Summarize Radiology Findings (2018, methodology): “Kupiec et al. (1995) scored sentences with a feature-based statistical classiﬁer.”
Title Generation with Quasi-Synchronous Grammar (2010, background): “Much effort in automatic summarization has been devoted to sentence extraction which is often formalized as a classification task (Kupiec et al., 1995).”
A Survey on Automatic Text Summarization (2007, methodology): “2.2.1 Naive-Bayes Methods Kupiec et al. (1995) describe a method derived from Edmundson (1969) that is able to learn from data.”
Temporal summaries of new topics (2001, background): “The problems tackled by TDT are all story-based rather than sentence based.”
Learning Algorithms for Keyphrase Extraction (2000, background): “…explore the task of producing a summary of a document by extracting key sentences from the document (Luhn, 1958; Edmundson, 1969; Marsh et al., 1984; Paice, 1990; Paice and Jones, 1993; Johnson et al., 1993; Salton et al., 1994; Kupiec et al., 1995; Brandow et al., 1995; Jang and Myaeng, 1997).”