A Language Model Approach to Keyphrase Extraction

By Takashi Tomokiyo, Matthew F. Hurst. Annual Meeting of the Association for Computational Linguistics, 2003

339 citations recorded by the source corpus.

We present a new approach to extracting keyphrases based on statistical language models. Our approach is to use pointwise KL-divergence between multiple language models for scoring both phraseness and informativeness, which can be unified into a single score to rank extracted phrases.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/92053021a4638e049743231ce4b761ae82b9cbfb

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 Computer Vision and Pattern Recognition

Citations in the corpus

Keyphrase Extraction in Scientific Publications (2007, methodology): “Tomokiyo and Hurst [3] take a language modeling approach to keyphrase generation by calculating the phraseness of a candidate, which represents the extent to which a word sequence is considered to have a phrasal quality.”
Automatic Keyphrase Extraction: A Survey of the State of the Art (2014, background): “In other words, its goal is to extract a set of phrases that are related to the main topics discussed in a given document (Tomokiyo and Hurst, 2003; Liu et al., 2009b; Ding et al., 2011; Zhao et al., 2011).”
GROBID: Combining Automatic Bibliographic Data Recognition and Term Extraction for Scholarship Publications (2009, methodology): “Term extraction in GROBID follows the approach of [3].”
The political blogosphere and the 2004 U.S. election: divided they blog (2005, background): “The algorithm for finding key bigrams combines a measure of informativeness and a measure of phraseness for a bigram into a single unified score to produce a ranked list of key bigrams [16].”
TopicRank: Graph-Based Topic Ranking for Keyphrase Extraction (2013, methodology): “…methods re-cast keyphrase extraction as a binary classiﬁcation task (Witten et al., 1999), whereas unsupervised methods apply different kinds of techniques such as language modeling (Tomokiyo and Hurst, 2003), clustering (Liu et al., 2009) or graph-based ranking (Mihalcea and Tarau, 2004).”
An exploratory study of COVID-19 misinformation on Twitter (2020, positive): “To investigate the most distinctive terms in our data, we used the pointwise Kullback Leibner divergence for Informativeness and Phraseness (KLIP) [73] as presented in [74] 9 for unigrams, bigrams and trigrams.”
Single Document Keyphrase Extraction Using Neighborhood Knowledge (2008, methodology): “Tomokiyo and Hurst (2003) use pointwise KL-divergence between multiple language models for scoring both phraseness and informativeness of phrases.”
Deep Keyphrase Generation (2017, methodology): “The ﬁrst step is to acquire a list of keyphrase candidates.”