The Automated Acquisition of Topic Signatures for Text Summarization

By Chin-Yew Lin, Eduard H. Hovy. International Conference on Computational Linguistics, 2000

554 citations recorded by the source corpus.

In order to produce a good summary, one has to identify the most relevant portions of a given text. We describe in this paper a method for automatically training topic signatures-sets of related words, with associated weights, organized around head topics and illustrate with signatures we created with 6,194 TREC collection texts over 4 selected topics. We describe the possible integration of topic signatures with outologies and its evaluaton on an automated text summarization system.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/cec852f9c844972cd2805e2196fb248136b8a32e

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

Citations in the corpus

A Survey of Text Summarization Techniques (2012, background): “Such words have been traditionally called “topic signatures” in the summarization literature [46].”
Recent automatic text summarization techniques: a survey (2016, background): “In this paper topic is represented in five different ways: • Topic signatures: Lin and Hovy (2000) suggested that a collection of terms are required to express a document’s topic.”
How to Compare Summarizers without Target Length? Pitfalls, Solutions and Re-Examination of the Neural Summarization Literature (2019, background): “Here is a list of the most cited ‘summarization’ papers of that period according to Google Scholar (Erkan and Radev, 2004; Radev et al., 2004; Gong and Liu, 2001; Conroy and O’leary, 2001; Lin and Hovy, 2000; Mihalcea, 2004; Goldstein et al., 2000).”
DBpedia - A large-scale, multilingual knowledge base extracted from Wikipedia (2015, methodology): “We use the computed weights to select the strongest related word stems for each entity and build topic signatures [27].”
Unsupervised Sentence Enhancement for Automatic Summarization (2014, methodology): “The clusters are scored according to the signature term method of Lin and Hovy (2000), which assigns an importance score to each term accord-BMFoods ing to how much more often it appears in the source text compared to some irrelevant background text using a log likelihood ratio.”
Enriching very large ontologies using the WWW (2000, background): “Lin and Hovy [9] use a likelihood ratio from maximum likelihood estimates that achieves even better performance on clean data.”
Unsupervised Learning of Narrative Event Chains (2008, background): “In summarization, topic signatures are a set of terms indicative of a topic (Lin and Hovy, 2000).”
Improving the Estimation of Word Importance for News Multi-Document Summarization (2014, methodology): “The log-likelihood ratio test (Lin and Hovy, 2000) compares the distribution of a word in the input with that in a large background corpus to identify topic words.”
Automatically Assessing Machine Summary Content Without a Gold Standard (2013, methodology): “Topic signatures are words highly descriptive of the input, as determined by the application of the log-likelihood test (Lin and Hovy 2000).”
Cross-Language Document Summarization Based on Machine Translation Quality Prediction (2010, methodology): “For single document summarization, the sentence score is usually computed by empirical combination of a number of statistical and linguistic feature values, such as term frequency, sentence position, cue words, stigma words, topic signature (Luhn 1969; Lin and Hovy, 2000).”
Recent advances in document summarization (2017, unknown): “More powerful usages include log-likelihood ratio test for identifying topic signature words that are highly descriptive of the input documents [114].”
DBpedia: A Multilingual Cross-domain Knowledge Base (2012, methodology): “We use those weights to select the strongest related terms for each entity and build topic signatures (Lin and Hovy, 2000).”
BabelNet: The automatic construction, evaluation and application of a wide-coverage multilingual semantic network (2012, methodology): “In this task, a variety of knowledge bases are assessed by first generating so-called topic signatures [65] and then using these to perform monolingual Word Sense Disambiguation (WSD) on standard datasets from previous Senseval [74] and SemEval [109] competitions.”
Abstractive Summarization of Spoken and Written Conversations Based on Phrasal Queries (2014, methodology): “We use a method described in (Lin and Hovy, 2000) in order to identify such terms and their associated weight.”
Espresso: Leveraging Generic Patterns for Automatically Harvesting Semantic Relations (2006, background): “To date, researchers have harvested, with varying success, several resources, including concept lists (Lin and Pantel 2002), topic signatures (Lin and Hovy 2000), facts (Etzioni et al. 2005), and word similarity lists (Hindle 1990).”
Left-Brain/Right-Brain Multi-Document Summarization (2004, methodology): “To identify these tokens, we used the log-likelihood statistic suggested by [4] and used first in summarization by Lin and Hovy ([6]).”
Extractive Summarization Using Supervised and Semi-Supervised Learning (2008, background): “Lin and Hovy (2000) identified signature terms that were strongly associated with documents based on statistics measures.”