Fast and Accurate Annotation of Short Texts with Wikipedia Pages

By Paolo Ferragina, Ugo Scaiella. 2010

284 citations recorded by the source corpus.

Several recent software systems have been designed to obtain novel annotation of cross-referencing text fragments and Wikipedia pages. Tagme is state of the art in this setting and can accurately manage short textual fragments (such as snippets of search engine results, tweets, news, or blogs) on the fly.

Subfields: Artificial Intelligence, Communication, Natural Language Processing, Information Extraction

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 Communication

In Natural Language Processing

In Information Extraction

Citations in the corpus

Word-Entity Duet Representations for Document Ranking (2017, methodology): “Entity linking systems usually contain two main steps [7]: (1) Spotting : To find surface forms in the text, for example, to identify the phrase ‘Obama’.”
Learning to Retrieve Reasoning Paths over Wikipedia Graph for Question Answering (2019, methodology): “In this experiment, we ﬁrst retrieve seed paragraphs using TF-IDF ( F = 100 ), and run an off-the-shelf entity linker (TagMe by Ferragina & Scaiella (2011)) over the paragraphs.”
Knowledge Guided Text Retrieval and Reading for Open Domain Question Answering (2019, methodology): “We use WikiExtractor 8 , TAGME (Ferragina and Scaiella, 2011) 9 and Rank-BM25 10 for Wikipedia dump parsing, entity extraction and BM25.”
KILT: a Benchmark for Knowledge Intensive Language Tasks (2020, methodology): “To find answers, the original work used distant supervision: (1) find Wikipedia entities in the question with the TAGME entity linked (Ferragina and Scaiella, 2011); (2) search for the answer (and all Wikipedia aliases) in the corresponding page; (3) if the answer is found, add the page in the evide”
Scalable Zero-shot Entity Linking with Dense Entity Retrieval (2019, methodology): “We release our code and models, as well as a system to link entity mentions to all of Wikipedia (similar to TagME (Ferragina and Scaiella, 2011)).”
EsdRank: Connecting Query and Documents through External Semi-Structured Data (2015, methodology): “Query annotation ( AnnQ ) was done by TagMe [12], one of the best systems in the Short (Query) Track at the ERD14 workshop competition [7].”
A framework for benchmarking entity-annotation systems (2013, background): “Surprisingly TagMe gives the best results, even if it was designed to annotate short texts [3, 4].”
GERBIL - Benchmarking Named Entity Recognition and Linking consistently (2017, background): “24 The new annotator includes a re-design of all TagME components, namely, the spotter, the dis-ambiguator, and the pruner.”
Bag-of-Entities Representation for Ranking (2016, methodology): “TagMe is an entity linking system [2] widely used in prior research [4, 7].”
Open the Black Box Data-Driven Explanation of Black Box Decision Systems (2018, methodology): “One example of a semantic annotation tool in the text mining domain is TagMe [13], which maps selected keywords in a text to Wikipedia concepts.”
Entity Linking meets Word Sense Disambiguation: a Unified Approach (2014, methodology): “09 (Ferragina and Scaiella, 2012) which uses the relatedness measure defined”
Design Challenges for Entity Linking (2015, background): “We are also aware of other systems such as TagMe-2 (Ferragina and Scaiella, 2012), DBpedia Spotlight (Mendes et al.”