Entity Linking meets Word Sense Disambiguation: a Unified Approach
By Andrea Moro, Alessandro Raganato, Roberto Navigli. 2014
877 citations recorded by the source corpus.
Entity Linking (EL) and Word Sense Disambiguation (WSD) both address the lexical ambiguity of language. But while the two tasks are pretty similar, they differ in a fundamental respect: in EL the textual mention can be linked to a named entity which may or may not contain the exact mention, while in WSD there is a perfect match between the word form (better, its lemma) and a suitable word sense. In this paper we present Babelfy, a unified graph-based approach to EL and WSD based on a loose identification of candidate meanings coupled with a densest subgraph heuristic which selects high-coherence semantic interpretations. Our experiments show state-of-the-art performances on both tasks on 6 different datasets, including a multilingual setting. Babelfy is online at http://babelfy.org
Subfields: Artificial Intelligence, 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
- José Camacho-Collados
- Mohammad Taher Pilehvar
- Roberto Navigli
- Ivan Vulic
- Alessandro Raganato
- Goran Glavas
- Jie Tang
- Simone Paolo Ponzetto
In Information Extraction
- Ignacio Iacobacci
- Mohammad Taher Pilehvar
- Roberto Navigli
- José Camacho-Collados
- Philippe Cudré-Mauroux
- Alisa Smirnova
- Axel-Cyrille Ngonga Ngomo
- Ricardo Usbeck
Citations in the corpus
- Knowledge Graphs (2020, neutral): “A common trend, for example, is to combine interdependent tasks, jointly performing WSD and EL [361], or NER and EL [329, 381], or NER and RE [437, 578], etc.”
- Neural Sequence Learning Models for Word Sense Disambiguation (2017, methodology): “, 2014) (UKBgloss w2w) , and Babelfy (Moro et al., 2014).”
- Towards a Seamless Integration of Word Senses into Downstream NLP Applications (2017, background): “The lower performance of graphbased disambiguation algorithms on short texts is a known issue (Moro et al., 2014; Raganato et al., 2017), the tackling of which remains an area of exploration.”
- End-to-End Neural Entity Linking (2018, background): “…EL and ED public systems, e.g. WAT (Piccinno and Ferragina, 2014), Dbpedia Spotlight (Mendes et al., 2011), NERD-ML (van Erp et al., 2013), KEA (Steinmetz and Sack, 2013), AIDA (Hoffart et al., 2011), FREME, FOX (Speck and Ngomo, 2014), PBOH (Ganea et al., 2016) or Babelfy (Moro et al., 2014).”
- Relation Extraction Using Distant Supervision (2018, methodology): “Multiple effective methods were proposed in that context, leveraging word sense disambiguation [42], graph-based methods [23] or crowdsourcing [13].”
- From Word to Sense Embeddings: A Survey on Vector Representations of Meaning (2018, background): “Semantic networks constitute suitable knowledge resources for disambiguating large amounts of text (Agirre et al., 2014; Moro et al., 2014).”
- Word Sense Disambiguation: A Unified Evaluation Framework and Empirical Comparison (2017, methodology): “BabelNet includes Wikipedia as one of its main sources of information.”
- OAG: Toward Linking Large-scale Heterogeneous Entity Graphs (2019, background): “The problem of entity linking is related to entity matching [20, 25, 36], entity resolution [2, 3], web appearance disambiguation [1, 11], name identification [16], object distinction [37], and name disambiguation [8, 30, 40], and has been extensively studied for decades by different communities.”
- GERBIL - Benchmarking Named Entity Recognition and Linking consistently (2017, methodology): “, entity salience [9], word sense disambiguation (WSD) [41] and relation extraction [56].”
- SensEmbed: Learning Sense Embeddings for Word and Relational Similarity (2015, methodology): “As our WSD system, we opted for Babelfy4 (Moro et al., 2014), a state-of-the-art WSD and Entity Linking algorithm based on BabelNet’s semantic network.”
- From Zero to Hero: On the Limitations of Zero-Shot Language Transfer with Multilingual Transformers (2020, background): “, 2017) is impaired by the limited KB coverage and inaccurate entity linking (Moro et al., 2014; Raiman and Raiman, 2018).”
- Deep learning meets ontologies: experiments to anchor the cardiovascular disease ontology in the biomedical literature (2018, methodology): “Pilehvar and Collier [18] reported improvements by post-processing, i.e. recalculating each dimension of the resulting word vector per phenotype considering a list of weighted words obtained via Babelfy [62].”
- Embedding Words and Senses Together via Joint Knowledge-Enhanced Training (2016, methodology): “back-off strategy (Navigli, 2009; Jin et al., 2009; Zhong and Ng, 2010; Moro et al., 2014; Raganato et al., 2017) and is used in various NLP applications (Bennett et al.”