Contributions to Morphology Learning using Conditional Random Fields

Teemu Ruokolainen

Contributions to Morphology Learning using Conditional Random Fields

Julkaisun otsikon käännös: Kontribuutioita morfologian oppimiseen ehdollisilla satunnaiskentillä

Tutkimustuotos: Doctoral Thesis › Collection of Articles

Abstrakti

Luonnollisen kielen käsittelyssä (LKK) tutkitaan järjestelmiä, jotka suorittavat ihmiskieleen liittyviä tehtäviä automaattisesti ilman ihmisen valvontaa. Tässä väitöskirjassa tarkastellaan LKK-järjestelmiä, jotka liittyvät morfologiseen analyysiin eli sanojen sisäisen rakenteen kuvaukseen. Morfologiset kuvaukset ovat tarpeellisia monien sovellusten, kuten hakukoneiden, kielenkääntäjien ja puheentunnistimien, kannalta, jotta kyseiset sovellukset voivat käsitellä harvinaisia ja tuntemattomia sanamuotoja. Työssä keskitytään erityisesti kahteen yleisesti käytettyyn analyysimenetelmään, morfologiseen jäsennykseen ja pilkontaan. Morfologisessa jäsennyksessä sanamuodoille annetaan luokituksia niiden morfologisten ominaisuuksien mukaan. Morfologisessa pilkonnassa sanojen sisäistä rakennetta kuvaillaan pilkkomalla sanamuodot niiden pienimpiin merkitystä sisältäviin osiin, morfeemeihin. Tässä väitöskirjassa morfologista jäsennystä ja pilkontaa lähestytään käyttäen tilastollista koneoppimismetodologiaa eli järjestelmät oppivat suorittamaan analyysin asiantuntijan muodostamien esimerkkien avulla. Erityisesti keskitytään ehdollisten satunnaiskenttien (ESK) soveltamiseen. 2000-luvun alussa julkaistua ESK-mallia on aikaisemmin sovellettu menestyksekkäästi useissa jäsennys- ja pilkontatehtävissä. Väitöskirjan ensimmäisenä pääkontribuutiona työssä tarkastellaan morfologisen pilkonnan oppimista ESK-mallin avulla. Erityisesti tarkastellaan puoliohjattua oppimisasetelmaa, jossa käytettävissä oleva data muodostuu pienestä määrästä annotoituja pilkontaesimerkkejä ja suuresta määrästä annotoimattomia, "raakoja", sanamuotoja. Kokeelliset tulokset kuudella kielellä osoittavat, että ehdotettu puoliohjattu ESK-pohjainen lähestymistapa on erittäin kilpailukykyinen menetelmä aikaisemmin julkaistuihin menetelmiin verrattuna. Erityisesti huomataan, että suljettujen luokkien ilmiöt, kuten suomen kielen suffiksaatio, voidaan oppia jo pienestä määrästä annotoituja esimerkkejä ohjatulla opetuksella. Toisaalta avoimen luokkien ilmiöt, kuten yhdyssanojen muodostaminen suomessa, voidaan oppia hyödyntämällä suurta määrää annotoimattomia sanamuotoja käyttäen puoliohjattua opetusta. Toisena pääkontribuutiona väitöskirjassa esitellään FinnPos, ensimmäinen suomen kielelle julkaistu avoimen lähdekoodin tilastollinen morfologinen jäsennin. ESK-malliin pohjautuvaa FinnPos-järjestelmää voidaan soveltaa suomen kielisen tekstin morfologiseen jäsentämiseen käyttäen Turku Dependency Treebank- ja FinnTreebank-puupankkien avulla opetettuja malleja.

Julkaisun otsikon käännös	Kontribuutioita morfologian oppimiseen ehdollisilla satunnaiskentillä
Alkuperäiskieli	Englanti
Pätevyys	Tohtorintutkinto
Myöntävä instituutio	Aalto-yliopisto
Valvoja/neuvonantaja	Kurimo, Mikko, Vastuuprofessori Virpioja, Sami, Ohjaaja
Kustantaja	Aalto University
Painoksen ISBN	978-952-60-6753-7
Sähköinen ISBN	978-952-60-6754-4
Tila	Julkaistu - 2016
OKM-julkaisutyyppi	G5 Tohtorinväitöskirja (artikkeli)

Tutkimusalat

luonnollinen kieli
morfologia
ehdolliset satunnaiskentät
jäsennys
pilkonta

Pääsy asiakirjaan

http://urn.fi/URN:ISBN:978-952-60-6754-4Lisenssi: Määrittelemätön

OpenUrl-saatavuus

Koko teksti

Siteeraa tätä

@phdthesis{7d88c3af93a0407a879e87aac7ad284c,

title = "Contributions to Morphology Learning using Conditional Random Fields",

abstract = "Natural language processing (NLP) refers to the study of systems performing natural language related tasks in an automatic manner, that is, without human supervision or interference. This thesis work considers NLP problems related to morphology analysis, that is, the description of internal structure of words. Acquiring knowledge of morphology is necessary in order for applications, such as search engines, machine translators, and speech recognizers, to successfully address rare and previously unseen word forms. In particular, we focus on two widely applied morphological analysis tasks, namely, morphological tagging and segmentation. In morphological tagging, the aim is to assign words in sentential contexts with word class labels describing their morphological properties. Meanwhile, morphological segmentation considers describing the inner word structure by splitting word forms into their smallest meaning-bearing units, morphemes. In the scope of this thesis, we approach the morphological tagging and segmentation problems using statistical, data-driven machine learning methodology. Using this approach, the processing systems are learned (estimated) based on training data prepared manually by a human expert. In particular, we focus on the highly influential conditional random field (CRF) model proposed for sequence tagging and segmentation in the early 2000s.As the first main contribution, the thesis discusses data-driven morphological segmentation employing the CRF model. A particular emphasis is placed on the semi-supervised learning setting, in which the available data consists of a small number of annotated segmentation examples and a large amount of unannotated raw word forms. The provided empirical evaluation on six languages shows that the proposed semi-supervised CRF-based approach is highly successful in the considered morphological segmentation task compared to earlier methods. In particular, the performed error analysis shows that closed class phenomena, such as suffixation of English and Finnish, can be learned already from a small number of annotated examples in a supervised manner. Meanwhile, open morpheme class phenomena, such as compounding of Finnish, can be learned by additionally exploiting the large unannotated word list using the semi-supervised approach. As the second main contribution, the thesis contains a presentation of FinnPos, the first open-source statistical morphological tagging and lemmatization toolkit designed specifically for Finnish. The CRF-based FinnPos system is readily applicable for tagging and lemmatization of running text with models learned from the recently published Finnish Turku Dependency Treebank and FinnTreeBank.",

keywords = "language technology, natural language, morphology, conditional random fields, tagging, segmentation, luonnollinen kieli, morfologia, ehdolliset satunnaiskent{\"a}t, j{\"a}sennys, pilkonta, language technology, natural language, morphology, conditional random fields, tagging, segmentation",

author = "Teemu Ruokolainen",

year = "2016",

language = "English",

isbn = "978-952-60-6753-7",

series = "Aalto University publication series DOCTORAL DISSERTATIONS",

publisher = "Aalto University",

number = "67",

address = "Finland",

school = "Aalto University",

}

TY - THES

T1 - Contributions to Morphology Learning using Conditional Random Fields

AU - Ruokolainen, Teemu

PY - 2016

Y1 - 2016

N2 - Natural language processing (NLP) refers to the study of systems performing natural language related tasks in an automatic manner, that is, without human supervision or interference. This thesis work considers NLP problems related to morphology analysis, that is, the description of internal structure of words. Acquiring knowledge of morphology is necessary in order for applications, such as search engines, machine translators, and speech recognizers, to successfully address rare and previously unseen word forms. In particular, we focus on two widely applied morphological analysis tasks, namely, morphological tagging and segmentation. In morphological tagging, the aim is to assign words in sentential contexts with word class labels describing their morphological properties. Meanwhile, morphological segmentation considers describing the inner word structure by splitting word forms into their smallest meaning-bearing units, morphemes. In the scope of this thesis, we approach the morphological tagging and segmentation problems using statistical, data-driven machine learning methodology. Using this approach, the processing systems are learned (estimated) based on training data prepared manually by a human expert. In particular, we focus on the highly influential conditional random field (CRF) model proposed for sequence tagging and segmentation in the early 2000s.As the first main contribution, the thesis discusses data-driven morphological segmentation employing the CRF model. A particular emphasis is placed on the semi-supervised learning setting, in which the available data consists of a small number of annotated segmentation examples and a large amount of unannotated raw word forms. The provided empirical evaluation on six languages shows that the proposed semi-supervised CRF-based approach is highly successful in the considered morphological segmentation task compared to earlier methods. In particular, the performed error analysis shows that closed class phenomena, such as suffixation of English and Finnish, can be learned already from a small number of annotated examples in a supervised manner. Meanwhile, open morpheme class phenomena, such as compounding of Finnish, can be learned by additionally exploiting the large unannotated word list using the semi-supervised approach. As the second main contribution, the thesis contains a presentation of FinnPos, the first open-source statistical morphological tagging and lemmatization toolkit designed specifically for Finnish. The CRF-based FinnPos system is readily applicable for tagging and lemmatization of running text with models learned from the recently published Finnish Turku Dependency Treebank and FinnTreeBank.

AB - Natural language processing (NLP) refers to the study of systems performing natural language related tasks in an automatic manner, that is, without human supervision or interference. This thesis work considers NLP problems related to morphology analysis, that is, the description of internal structure of words. Acquiring knowledge of morphology is necessary in order for applications, such as search engines, machine translators, and speech recognizers, to successfully address rare and previously unseen word forms. In particular, we focus on two widely applied morphological analysis tasks, namely, morphological tagging and segmentation. In morphological tagging, the aim is to assign words in sentential contexts with word class labels describing their morphological properties. Meanwhile, morphological segmentation considers describing the inner word structure by splitting word forms into their smallest meaning-bearing units, morphemes. In the scope of this thesis, we approach the morphological tagging and segmentation problems using statistical, data-driven machine learning methodology. Using this approach, the processing systems are learned (estimated) based on training data prepared manually by a human expert. In particular, we focus on the highly influential conditional random field (CRF) model proposed for sequence tagging and segmentation in the early 2000s.As the first main contribution, the thesis discusses data-driven morphological segmentation employing the CRF model. A particular emphasis is placed on the semi-supervised learning setting, in which the available data consists of a small number of annotated segmentation examples and a large amount of unannotated raw word forms. The provided empirical evaluation on six languages shows that the proposed semi-supervised CRF-based approach is highly successful in the considered morphological segmentation task compared to earlier methods. In particular, the performed error analysis shows that closed class phenomena, such as suffixation of English and Finnish, can be learned already from a small number of annotated examples in a supervised manner. Meanwhile, open morpheme class phenomena, such as compounding of Finnish, can be learned by additionally exploiting the large unannotated word list using the semi-supervised approach. As the second main contribution, the thesis contains a presentation of FinnPos, the first open-source statistical morphological tagging and lemmatization toolkit designed specifically for Finnish. The CRF-based FinnPos system is readily applicable for tagging and lemmatization of running text with models learned from the recently published Finnish Turku Dependency Treebank and FinnTreeBank.

KW - language technology

KW - natural language

KW - morphology

KW - conditional random fields

KW - tagging

KW - segmentation

KW - luonnollinen kieli

KW - morfologia

KW - ehdolliset satunnaiskentät

KW - jäsennys

KW - pilkonta

KW - language technology

KW - natural language

KW - morphology

KW - conditional random fields

KW - tagging

KW - segmentation

M3 - Doctoral Thesis

SN - 978-952-60-6753-7

T3 - Aalto University publication series DOCTORAL DISSERTATIONS

PB - Aalto University

ER -

Contributions to Morphology Learning using Conditional Random Fields

Abstrakti

Tutkimusalat

Pääsy asiakirjaan

OpenUrl-saatavuus

Sormenjälki

Siteeraa tätä