DKPro WSD

WSD corpora

Table of WSD corpora

Date Corpus Language Style Train Format Inventory POS Lemma Sent. Notes
1998 Senseval-1 task: English lexical sample en LS no Senseval2LS HECTOR ? ? ? 1
2001 Senseval-2 task: Basque lexical sample eu LS no Senseval2LS Euskal Hiztegia subset, TEI-SGML yes yes no  
2001 Senseval-2 task: Czech all words cs AW no Senseval2AW custom, text no no no  
2001 Senseval-2 task: Dutch all words nl AW no Senseval2AW none no no yes  
2001 Senseval-2 task: English all words en AW no Senseval2AW WordNet 1.7pre no no no 2
2001 Senseval-2 task: English all words (Rada Mihalcea’s conversion) en AW no SemCor WordNet 1.7.1 through 3.0 yes yes yes 3
2001 Senseval-2 task: English group lexical sample en LS yes Senseval2LS WordNet 1.7pre yes yes no 2
2001 Senseval-2 task: English lexical sample en LS no Senseval2LS WordNet 1.7pre yes yes no 2, 3
2001 Senseval-2 task: Estonian all words et AW yes Senseval2AW Estonian EWN v37 no no no  
2001 Senseval-2 task: Italian lexical sample it LS no Senseval2LS Italian EWN no yes no  
2001 Senseval-2 task: Japanese lexical sample ja LS yes Senseval2LS custom, XML no yes no  
2001 Senseval-2 task: Korean lexical sample ko LS yes Senseval2LS custom, DOC no no no  
2001 Senseval-2 task: Spanish lexical sample es LS yes Senseval2LS custom, text no yes no  
2001 Senseval-2 task: Swedish lexical sample sv LS yes Senseval2LS custom, XML no yes no  
2002 SemCor (NLTK’s conversion to XML) en AW no SemCor WordNet 3.0 yes yes yes  
2003 TWA en LS no custom XML custom yes yes yes  
2004 Senseval-3 task 01: English all words en AW no Senseval2AW WordNet 1.7.1 no no no  
2004 Senseval-3 task 01: English all words (Rada Mihalcea’s conversion) en AW no SemCor WordNet 1.7.1 through 3.0 yes yes yes 3
2004 Senseval-3 task 02: Italian all words it AW no ? ItalWordNet ? ? ?  
2004 Senseval-3 task 03: Basque lexical sample eu LS yes custom XML ? yes yes yes  
2004 Senseval-3 task 04: Catalan lexical sample ca LS yes ? ? ? ? ?  
2004 Senseval-3 task 05: Chinese lexical sample zh LS yes ? ? ? ? ?  
2004 Senseval-3 task 06: English lexical sample en LS yes Senseval2LS WordNet 1.7.1, Wordsmyth ? ? ?  
2004 Senseval-3 task 07: Italian lexical sample it LS yes ? ? ? ? ?  
2004 Senseval-3 task 08: Romanian lexical sample ro LS yes ? ? ? ? ?  
2004 Senseval-3 task 09: Spanish lexical sample es LS yes ? ? ? ? ?  
2004 Senseval-3 task 10: Automatic subcategorization acquisition en LS no ? WordNet 1.7.1 ? ? ?  
2004 Senseval-3 task 11: Multilingual lexical sample en,hi LS yes ? ? ? ? ?  
2004 Senseval-3 task 12: WSD of WordNet glosses en AW no ? ? ? ? ?  
2004 Senseval-3 task 13: Semantic Roles en LS yes ? ? ? ? ?  
2005 Estonian WSD corpus et AW no XML, text Estonian EWN (various versions) yes yes yes  
2007 MSNBC en AW ? ? Wikipedia no no no  
2007 Semeval-1 task 01: Evaluating WSD on Cross Language Information Retrieval en,es ? ? ? ? ? ? ?  
2007 Semeval-1 task 02: Evaluating Word Sense Induction and Discrimination Systems en ? ? ? ? ? ? ?  
2007 Semeval-1 task 04: Classification of Semantic Relations between Nominals en ? ? ? ? ? ? ?  
2007 Semeval-1 task 05: Multilingual Chinese-English Lexical Sample Task en,zh LS ? ? ? ? ? ?  
2007 Semeval-1 task 06: Word-Sense Disambiguation of Prepositions en LS ? ? custom, XML ? ? ?  
2007 Semeval-1 task 07: Coarse-grained English all-words en AW no Semeval1AW WordNet 2.1 yes yes yes 3
2007 Semeval-1 task 08: Metonymy Resolution at Semeval-2007 en LS ? ? ? ? ? ?  
2007 Semeval-1 task 09: Multilevel Semantic Annotation of Catalan and Spanish ca,es ? ? ? ? ? ? ?  
2007 Semeval-1 task 10: English Lexical Substitution Task for SemEval-2007 en ? ? ? ? ? ? ?  
2007 Semeval-1 task 11: English Lexical Sample Task via English-Chinese Parallel Text en LS ? ? ? ? ? ?  
2007 Semeval-1 task 12: Turkish Lexical Sample Task tr LS ? ? ? ? ? ?  
2007 Semeval-1 task 13: Web People Search en ? ? ? ? ? ? ?  
2007 Semeval-1 task 14: Affective Text en ? ? ? ? ? ? ?  
2007 Semeval-1 task 15: TempEval: A proposal for Evaluating Time-Event Temporal Relation Identification en ? ? ? ? ? ? ?  
2007 Semeval-1 task 16: Evaluation of wide coverage knowledge resources en ? ? ? ? ? ? ?  
2007 Semeval-1 task 17.1: Coarse-grained English Lexical Sample en LS yes ? OntoNotes (WordNet 1.7, 2.0, 2.1) ? ? ?  
2007 Semeval-1 task 17.2: Coarse-grained English Lexical Sample SRL en LS ? ? ? ? ? ?  
2007 Semeval-1 task 17.3: English fine-grained all-words en AW ? Senseval2AW WordNet 2.1 no no no  
2007 Semeval-1 task 18: Arabic Semantic Labeling ar ? ? ? ? ? ? ?  
2007 Semeval-1 task 19: Frame Semantic Structure Extraction en ? ? ? FrameNet 1.3 ? ? ?  
2008 AQUAINT Newswire en AW no ? Wikipedia no no no  
2009 TAC_KBP 2009 Newswire en LS yes TAC_KBP TAC_KB (October 2008 Wikipedia) no no no  
2009 TAC_KBP 2010 Newswire en LS no TAC_KBP TAC_KB (October 2008 Wikipedia) no no no  
2010 Semeval-2 task 01: Coreference Resolution in Multiple Languages mul ? ? ? ? ? ? ?  
2010 Semeval-2 task 02: Cross-Lingual Lexical Substitution mul LS no ? ? ? ? ?  
2010 Semeval-2 task 03: Cross-Lingual Word Sense Disambiguation mul ? no ? ? ? ? ?  
2010 Semeval-2 task 04: VP Ellipsis - Detection and Resolution ? ? no ? ? ? ? ?  
2010 Semeval-2 task 05: Automatic Keyphrase Extraction from Scientific Articles ? ? ? ? ? ? ? ?  
2010 Semeval-2 task 07: Argument Selection and Coercion ? ? ? ? ? ? ? ?  
2010 Semeval-2 task 08: Multi-Way Classification of Semantic Relations Between Pairs of Nominals ? ? ? ? ? ? ? ?  
2010 Semeval-2 task 09: Noun Compound Interpretation Using Paraphrasing Verbs ? ? ? ? ? ? ? ?  
2010 Semeval-2 task 10: Linking Events and their Participants in Discourse ? ? ? ? ? ? ? ?  
2010 Semeval-2 task 11: Event Detection in Chinese News Sentences zh ? ? ? ? ? ? ?  
2010 Semeval-2 task 12: Parser Training and Evaluation using Textual Entailment ? ? no ? ? ? ? ?  
2010 Semeval-2 task 13: TempEval 2 ? ? ? ? ? ? ? ?  
2010 Semeval-2 task 14: Word Sense Induction ? ? ? ? ? ? ? ?  
2010 Semeval-2 task 15: Infrequent Sense Identification for Mandarin Text to Speech Systems zh ? ? ? ? ? ? ?  
2010 Semeval-2 task 16: Japanese WSD ja ? no ? ? ? ? ?  
2010 Semeval-2 task 17: All-words Word Sense Disambiguation on a Specific Domain en AW no Semeval2AW WordNet 3.0 (?) no no no  
2010 Semeval-2 task 17: All-words Word Sense Disambiguation on a Specific Domain it AW no Semeval2AW ? no no no  
2010 Semeval-2 task 17: All-words Word Sense Disambiguation on a Specific Domain nl AW no Semeval2AW ? no no no  
2010 Semeval-2 task 17: All-words Word Sense Disambiguation on a Specific Domain zh AW no Semeval2AW ? no no no  
2010 Semeval-2 task 18: Disambiguating Sentiment Ambiguous Adjectives ? ? no ? ? ? ? ?  
2010 TAC_KBP 2010 Web data en LS yes TAC_KBP TAC_KB (October 2008 Wikipedia) no no no  
2010 TAC_KBP 2011 Newswire en LS no TAC_KBP TAC_KB (October 2008 Wikipedia) no no no  
2010 TAC_KBP 2011 Web data en LS yes TAC_KBP TAC_KB (October 2008 Wikipedia) no no no  
2011 ACE en AW ? ? Wikipedia no no no  
2011 OntoNotes 4.0 en ? no ? OntoNotes yes yes yes  
2011 WikiAmbi en LS no ? Wikipedia no no no  
2011 Yago_CoNLL en AW no AIDA Wikipedia, DBPedia no no no  
2012 MASC en LS no MASC WordNet 3.1, FrameNet yes yes yes  
2013 WebCAGe de LS no WebCAGe GermaNet yes yes yes  

Legend

  • Date: The date the corpus (or the original corpus upon which this version is based) was first published
  • Corpus: The name of the corpus, possibly with a hyperlink to where it can be obtained
  • Language: The language(s) of the corpus
  • Style: AW = all-words, LS = lexical sample
  • Train: Whether the corpus is split into separate training and test corpora
  • Format: The file format for the data
  • Inventory: Which sense inventory is used to annotate the data
  • POS: Whether the corpus has POS annotations
  • Lemma: Whether the corpus has lemma annotations
  • Sent.: Whether the corpus has sentence annotations
  • Notes: Notes (see below)

Notes

  1. Incomplete mappings from HECTOR to WordNet 1.5 and WordNet 1.6 are available on the Senseval-1 web page.
  2. According to Christiane Fellbaum, the WordNet 1.7 pre-release is missing and presumed lost. A third-party copy provided to us by German Rigau was found to be incompatible with the Senseval-2 sense keys.
  3. These data sets contain errors (incorrect POS tags, mismatched ID references, invalid XML, etc.). We provide patches and conversion scripts to fix them.

Formats supported by DKPro WSD

DKPro WSD supports many (but not all) of the formats in the table above. The table below shows the DKPro WSD reader classes for supported formats:

Format Class
AIDA de.tudarmstadt.ukp.dkpro.wsd.io.reader.AidaReader
MASC de.tudarmstadt.ukp.dkpro.wsd.io.reader.MASCReader
SemCor de.tudarmstadt.ukp.dkpro.wsd.io.reader.SemCorXMLReader
Semeval1AW de.tudarmstadt.ukp.dkpro.wsd.io.reader.Semeval1AWReader
Semeval2AW de.tudarmstadt.ukp.dkpro.wsd.io.reader.Semeval2AWReader
Senseval2LS de.tudarmstadt.ukp.dkpro.wsd.io.reader.Senseval2LSReader
Senseval2AW de.tudarmstadt.ukp.dkpro.wsd.io.reader.Senseval2AWReader
TAC_KBP de.tudarmstadt.ukp.dkpro.wsd.io.reader.tacKbp.TacKbpOfficialformatReader
WebCAGe de.tudarmstadt.ukp.dkpro.wsd.io.reader.WebCAGeXMLReader