unsupervised approaches to sequence tagging, …nschneid/ls2lit_slides.pdf · ·...
TRANSCRIPT
![Page 1: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/1.jpg)
UnsupervisedApproachestoSequenceTagging,MorphologyInduction,andLexicalResource
AcquisitionRezaBosaghzadeh&NathanSchneider
LS2~1December2008
![Page 2: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/2.jpg)
UnsupervisedMethods– SequenceLabeling(Part‐of‐SpeechTagging)
– MorphologyInduction
– LexicalResourceAcquisition
.
She ran to the station quickly
pronoun verb preposition det noun adverb
un‐supervise‐dlearn‐ing
![Page 3: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/3.jpg)
ContrastiveEstimationSmith&Eisner(2005)
• Alreadydiscussedinclass• Keyidea:exploitsimplicitnegativeevidence
– Mutatingtrainingexamplesoftengivesungrammatical(negative)sentences
– Duringtraining,shiftprobabilitymassfromgeneratednegativeexamplestogivenpositiveexamples
• BUT:Requiresataggingdictionary,i.e.alistofpossibletagsforeachwordtype
![Page 4: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/4.jpg)
Prototype‐driventaggingHaghighi&Klein(2006)
+
PrototypesTargetLabel
UnlabeledData
PrototypeList
AnnotatedData
slidecourtesyHaghighi&Klein
![Page 5: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/5.jpg)
Prototype‐driventaggingHaghighi&Klein(2006)
Newlyremodeled2Bdrms/1Bath,spaciousupperunit,locatedinHilltopMallarea.Walkingdistancetoshopping,publictransportation,schoolsandpark.Paidwaterandgarbage.Nodogsallowed.
Newlyremodeled2Bdrms/1Bath,spaciousupperunit,locatedinHilltopMallarea.Walkingdistancetoshopping,publictransportation,schoolsandpark.Paidwaterandgarbage.Nodogsallowed.
PrototypeList
NN VBN CC JJ CD PUNC
IN NNS IN NNP RB DET
NN president IN of
VBD said NNS shares
CC and TO to
NNP Mr. PUNC .
JJ new CD million
DET the VBP are
EnglishPOS
slidecourtesyHaghighi&Klein
![Page 6: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/6.jpg)
Prototypes
Newlyremodeled2Bdrms/1Bath,spaciousupperunit,locatedinHilltopMallarea.Walkingdistancetoshopping,publictransportation,schoolsandpark.Paidwaterandgarbage.Nodogsallowed.
FEATURE kitchen, laundry
LOCATION near, close TERMS paid, utilities SIZE large, feet RESTRICT cat, smoking
InformationExtraction:ClassifiedAds
FeaturesLocationTermsRestrictSize
PrototypeList
slidecourtesyHaghighi&Klein
![Page 7: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/7.jpg)
Prototype‐driventaggingHaghighi&Klein(2006)
• Trigramtagger,samefeaturesas(Smith&Eisner2005)– Wordtype,suffixesuptolength3,contains‐hyphen,contains‐digit,initialcapitalization
• Tieeachwordtoitsmostsimilarprototype,usingcontext‐basedsimilaritytechnique(Schütze1993)– SVDdimensionalityreduction– Cosinesimilaritybetweencontextvectors
slideadaptedfromHaghighi&Klein
![Page 8: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/8.jpg)
Prototype‐driventaggingHaghighi&Klein(2006)
Pros• Doesn’trequiretaggingdictionaryCons• Stillneedatagset• Maybehardtochoosegoodprototypes
![Page 9: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/9.jpg)
UnsupervisedPOStaggingTheStateoftheArt
Bestsupervisedresult(CRF):99.5%!
![Page 10: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/10.jpg)
UnsupervisedMethods– SequenceLabeling(Part‐of‐SpeechTagging)
– MorphologyInduction
– LexicalResourceAcquisition
.
She ran to the station quickly
pronoun verb preposition det noun adverb
un‐supervise‐dlearn‐ing
![Page 11: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/11.jpg)
UnsupervisedApproachestoMorphology
• Morphologyreferstotheinternalstructureofwords– Amorphemeisaminimalmeaningfullinguisticunit
– Morphemesegmentationistheprocessofdividingwordsintotheircomponentmorphemes
un‐supervise‐dlearn‐ing– Wordsegmentationistheprocessoffindingwordboundariesinastreamofspeechortextunsupervised_learning_of_natural_language
![Page 12: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/12.jpg)
ParaMor:MorphologicalparadigmsMonsonetal.(2007,2008)
• Learnsinflectionalparadigmsfromrawtext– Requiresonlyalistofwordtypesfromacorpus– Looksatwordcountsofsubstrings,andproposes(stem,suffix)pairingsbasedontypefrequency
• 3‐stagealgorithm– Stage1:Candidateparadigmsbasedonfrequencies
– Stages2‐3:Refinementofparadigmsetviamergingandfiltering
• Paradigmscanbeusedformorphemesegmentationorstemming
![Page 13: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/13.jpg)
ParaMor:MorphologicalparadigmsMonsonetal.(2007,2008)
speak dance buyhablar bailar comprarhablo bailo comprohablamos bailamos compramoshablan bailan compran… … …
• AsamplingofSpanishverbconjugations(inflections)
![Page 14: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/14.jpg)
ParaMor:MorphologicalparadigmsMonsonetal.(2007,2008)
speak dance buyhablar bailar comprarhablo bailo comprohablamos bailamos compramoshablan bailan compran… … …
• Aproposedparadigm(correct):stems{habl,bail,compr}andsuffixes{‐ar,‐o,‐amos,‐an}
![Page 15: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/15.jpg)
ParaMor:MorphologicalparadigmsMonsonetal.(2007,2008)
• Twosubsequentstages:– Filteringoutspuriousparadigms(e.g.withincorrectsegmentations)
– Mergingpartialparadigmstoovercomesparsity:smoothing
![Page 16: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/16.jpg)
ParaMor:MorphologicalparadigmsMonsonetal.(2007,2008)
speak dancehablar bailarhablo bailohablamos bailamoshablan bailan… …
• Forcertainsub‐setsofverbs,thealgorithmmayproposeparadigmswithspuriousseg‐mentations,liketheoneatleft
• Thefilteringstageofthealgorithmweedsouttheseincorrectparadigms
![Page 17: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/17.jpg)
ParaMor:MorphologicalparadigmsMonsonetal.(2007,2008)
• Whatifnotallconjugationswereinthecorpus?
speak dance buyhablar bailar comprar
bailo comprohablamos bailamos compramoshablan… … …
![Page 18: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/18.jpg)
ParaMor:MorphologicalparadigmsMonsonetal.(2007,2008)
• Anotherstageofthealgorithmmergestheseoverlappingpartialparadigmsviaclustering
speak dance buyhablar bailar comprar
bailo comprohablamos bailamos compramoshablan… … …
![Page 19: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/19.jpg)
ParaMor:MorphologicalparadigmsMonsonetal.(2007,2008)
speak dance buyhablar bailar comprarhablo bailo comprohablamos bailamos compramoshablan bailan compran… … …
• Thisamountstosmoothing,or“hallucinating”out‐of‐vocabularyitems
![Page 20: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/20.jpg)
ParaMor:MorphologicalparadigmsMonsonetal.(2007,2008)
• Heuristic‐based,deterministicalgorithmcanlearninflectionalparadigmsfromrawtext
• Currently,ParaMorassumessuffix‐basedmorphology
• Paradigmscanbeusedstraightforwardlytopredictsegmentations– CombiningtheoutputsofParaMorandMorfessor(anothersystem)wonthesegmentationtaskatMorphoChallenge2008foreverylanguage:English,Arabic,Turkish,German,andFinnish
![Page 21: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/21.jpg)
• Wordsegmentationresults–comparison
• SeeNarges&Andreas’spresentationformoreonthismodel
Goldwateretal.UnigramDP
Goldwateretal.BigramHDP
BayesianwordsegmentationGoldwateretal.(2006;insubmission)
tablefromGoldwateretal.(insubmission)
![Page 22: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/22.jpg)
MultilingualmorphemesegmentationSnyder&Barzilay(2008)
speakrs speaktuhablar parlerhablo parlehablamos parlonshablan parlent… …
• Abstractmorphemescrosslanguages:(ar,er),(o,e),(amos,ons),(an,ent),(habl,parl)
• Considersparallelphrasesandtriestofindmorphemecorrespondences
• Straymorphemesdon’tcorrespondacrosslanguages
![Page 23: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/23.jpg)
MorphologyPapers:Inputs&Outputs
• Whatdoes“unsupervised”meanforeachapproach?
![Page 24: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/24.jpg)
UnsupervisedMethods– SequenceLabeling(Part‐of‐SpeechTagging)
– MorphologyInduction
– LexicalResourceAcquisition
.
She ran to the station quickly
pronoun verb preposition det noun adverb
un‐supervise‐dlearn‐ing
![Page 25: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/25.jpg)
BilinguallexiconsfrommonolingualcorporaHaghighietal.(2008)
SourceText
TargetText
Matching
m state
world
name
SourceWords
s
nation
estado
política
TargetWords
t
mundo
nombre
diagramcourtesyHaghighietal.UsedavariantofCCA(CanonicalCorrelationAnalysis)
![Page 26: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/26.jpg)
state
Orthographic Features 1.0
1.0
1.0
#st
tat te#
5.0
20.0
10.0
Context Features
world politics society
SourceText
estado
Orthographic Features 1.0
1.0
1.0
#es
sta do#
10.0
17.0
6.0
Context Features
mundo politica sociedad
TargetText
slidecourtesyHaghighietal.
BilingualLexiconsfromMonolingualCorporaHaghighietal.(2008)
DataRepresentation
![Page 27: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/27.jpg)
FeatureExperiments
61.1
80.1 80.289.0
0
25
50
75
100
EditDist Ortho Context MCCA
Precision
• MCCA:Orthographicandcontextfeatures
4kEN‐ESWikipediaArticlesslidecourtesyHaghighietal.
![Page 28: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/28.jpg)
NarrativeeventsChambers&Jurafsky(2008)
• Givenacorpus,identifiesrelatedeventsthatconstitutea“narrative”and(whenpossible)predicttheirtypicaltemporalordering– E.g.:NOPQPRSTUOVWXNYZPVRnarrative,withverbs:arrest,accuse,plead,testify,acquit/convict
• Keyinsight:relatedeventstendtoshareaparticipantinadocument– Thecommonparticipantmayfilldifferentsyntactic/semanticroleswithrespecttoverbs:arrest.V\]XNZ,accuse.V\]XNZ,plead.WY\]XNZ
![Page 29: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/29.jpg)
NarrativeeventsChambers&Jurafsky(2008)
• Atemporalclassifiercanreconstructpairwisecanonicaleventorderings,producingadirectedgraphforeachnarrative
![Page 30: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/30.jpg)
StatisticalverblexiconGrenager&Manning(2006)
• Fromdependencyparses,agenerativemodelpredictsforeachverb:– PropBank‐stylesemanticroles:wux0,wux1,etc.(donotnecessarilycorrespondacrossverbs)
– Theroles’syntacticrealizations,e.g.:
• Usedforsemanticrolelabeling
He gave me a cookie
subj ARG0
verb give
np#1 ARG2
np#2 ARG1
He gave a cookie to me
subj ARG0
verb give
np#2 ARG1
pp_to ARG2
![Page 31: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/31.jpg)
“Semanticity”:Ourproposedscaleofsemanticrichness
• text<POS<syntax/morphology/alignments<coreference/semanticroles/temporalordering<translations/narrativeeventsequences
• Wescoreeachmodel’sinputsandoutputsonthisscale,andcalltheinput‐to‐outputincrease“semanticgain”– Haghighietal.’sbilinguallexiconinductionwinsinthisrespect,goingfromrawtexttolexicaltranslations
![Page 32: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/32.jpg)
SemanticGain:ComparisonofMethods
![Page 33: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/33.jpg)
Robustnesstolanguagevariation• AbouthalfofthepapersweexaminedhadEnglish‐onlyevaluations
• Weconsideredwhichtechniquesweremostadaptabletoother(esp.resource‐poor)languages.Twomainfactors:– Relianceonexistingtools/resourcesforpreprocessing(parsers,coreferenceresolvers,…)
– Anylinguisticspecificityinthemodel(e.g.suffix‐basedmorphology)
![Page 34: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/34.jpg)
SummaryWeexaminedthreeareasofunsupervisedNLP:
1. Sequencetagging:HowcanwepredictPOS(ortopic)tagsforwordsinsequence?
2. Morphology:Howarewordsputtogetherfrommorphemes(andhowcanwebreakthemapart)?
3. Lexicalresources:Howcanweidentifylexicaltranslations,semanticrolesandargumentframes,ornarrativeeventsequencesfromtext?
Ineightrecentpaperswefoundavarietyofapproaches,includingheuristicalgorithms,Bayesianmethods,andEM‐styletechniques.
![Page 35: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/35.jpg)
ThankstoNoahandKevinfortheirfeedbackonthepaper;AndreasandNargesfortheir
collaborationonthepresentations;andallofyouforgivingusyourattention!
Questions?
un‐supervise‐dlearn‐ing
hablar bailar
hablo bailo
hablamos bailamos
hablan bailan
subj=give.wux0verb=givenp#1=give.wux2np#2=give.wux1
PrototypesTargetLabel
![Page 36: Unsupervised Approaches to Sequence Tagging, …nschneid/ls2lit_slides.pdf · · 2010-04-16Unsupervised Approaches to Sequence Tagging, Morphology Induction, and Lexical Resource](https://reader034.vdocuments.site/reader034/viewer/2022051601/5abcf0e27f8b9a567c8e5ea6/html5/thumbnails/36.jpg)
ImprovementIdeas
• POSTagging:Learnthetagset• Morphology:Non‐agglomerativeMorphology,Alsoparses
• LexicalResources:Trywordclasses
• All:Languagevariability