Corpus Linguistics and Ontologies

1
Corpus Linguistics and Ontologies

• Martin Volk
• Stockholm University

2
Overview of todays lecture

• What is Corpus Linguistics?
• Why do Computational Linguists work with corpora?
• What is an Ontology?
• Why do Computational Linguists work with / on
  Ontologies?
• Your task for next week!

3
Sources for linguistic information

• Introspection (own usage and judgement)
• Usage and judgement by others
• Questioning (goal-driven)
• interview
• questionaire
• Observation ('involuntary' utterances)
• spoken utterances (? corpora)
• written utterances (? corpora)

4
What is a corpus?

• a text collection
• a representative text collection
• a representative and structured text collection
• a representative, structured and annotated text
  collection
• ...

5
Example

• 'ob' is a conjunction in German.
• Is 'ob' also used as a preposition?
• Introspection
• Rothenburg ob der Tauber
• Dictionary (Wahrig. Deutsches Wörterbuch. 1996)
  Präp. mit Dativ veraltet ob dem Wasserfall
• Web Google 'ob dem'
• Sage Der Wilde Jäger ob dem Neuenburgersee
• Corpus

6
Corpus Examples

• CZ_94 ... fiel schier vom Stuhl ob der Äusserung
  eines Ozeanologen ...
• CZ_94 Bei manchem Ölgiganten kam ob der
  Ergebnisse gar Euphorie auf.
• CZ_94 ... rieben sich vergnügt die Hände ob des
  zu erwartenden Schlagabtauschs.
• ob is a preposition with genitive!!
• in CZ corpus 'ob' is tagged as preposition 21
  times (but some incorrect)

7
What is corpus linguistics?http//www.engl.polyu.
edu.hk/corpuslinguist/corpus.htm

• Corpus linguistics is simply the study of
  language through corpus-based research, but it
  differs from traditional linguistics in its
  insistence on the systematic study of authentic
  examples of language in use.
• Text linguistics vs corpus linguistics
• Illustration vs evidence
• Introspection and informant testing vs
  observation of text

8
Corpus research in Linguistics

• Lexicography (Dictionaries)
• collocations ? idioms, support verb units,
• Grammaticography (Reference grammars)
• Learner corpora Language acquisition
• Parallel corpora Translation
• Historical corpora Language development

9
Construction of Corpora

• Written text is easier to obtain than spoken
  text. Some examples
• Newspapers
• Fiction (e.g. fairy tales)
• Technical Literature (e.g. manuals, medicine)
• Personal letters Email
• Advertising (incl. political propaganda)
• Belief and Thought (e.g. bible)

10
Corpora of spoken language

• Spontaneous spoken language
• recording of dialogues (e.g. telephone
  conversation)
• Prepared spoken language
• Public speeches (e.g. in parliament)
• Radio or TV news
• Spoken utterances must be transcribed for
  linguistic research.

11
Brown Corpus (1964)

• 1 million words
• 500 texts
• out of 15 different text types
• with 2000 words each

12
British National Corpus (BNC)

• 1995, 100 Mio. words
• 90 written English, 10 spoken English
• 3209 texts
• out of 10 different text types written and
• 6 text types spoken
• with lt 40'000 words each
• ? multi-purpose corpus

13
The Stockholm-Umeå-Corpus (SUC)

• a one-million word corpus
• with 500 text samples
• from different text genres
• with Part-of-Speech information,
• morpho-syntactic information,
• and base forms,
• all manually checked.
• SUC_example_sent.htm

14
Types of corpora

• Raw texts
• Automatically annotated corpora
• Texts with Part-of-Speech tags
• Partially parsed texts
• Manually annotated corpora
• Treebank (e.g. the Penn Treebank)
• FrameNet

15
Types of Corpora

• Balanced Corpora vs. special corpora
• Spoken vs. written language
• Monolingual vs. Multilingual Corpora
• Parallel vs. comparable corpora

16
Parallel Corpora

• Translated texts that are aligned on the sentence
  level.
• Example from the Europarl corpus

17
Europarl examples

• Swedish-English
• the preposition med
• the verb uppmana
• the noun myndighet
• the adverb alltid
• Swedish-French
• the noun myndighet
• the adverb alltid

18
Corpora in Computational Linguistics
Corpora
annotation
Facts Rules Preferences
learning
19
The goal of Natural Language Processing in CL

• Build a system that simulates human language
  understanding
• with input a natural language utterance
• with output answers to questions as to
• who did (or will do) what to whom, when, where
  and why.
• NLP is not necessarily concerned with how people
  understand language ? Cognitive Linguistics

20
Basic problems in CL

• Knowledge is missing ( too little information)
• e.g. unknown words
• e.g. unknown relations between concepts
• Ambiguities ( too much information)
• e.g. in syntax attachment preferences
• e.g. in semantics word senses

21
Acknowledgement

• Some slides were highly influenced by or even
  copied from Anke Lüdeling's course "Introduction
  to Corpus Linguistics" at http//www.cl-ki.uni-osn
  abrueck.de/aluedeli/Corpuslinguistik.html

22
Take a breath

• Questions on Corpus Linguistics??
• Still to come
• An introduction to Ontologies

23
What is an ontology?

• An ontology holds information about what
  categories exist in the domain, what properties
  they have, and how they are related to one
  another. (Chandrasekaran et al. 1999)

24
Types of ontologies

• Top-level ontology
• Lexical ontology (e.g. WordNet)
• WordNet is a lexical database for English (with
  synonyms and hyperonyms)
• General ontology (e.g. Cyc)
• Cyc is a formalized representation of fundamental
  human knowledge facts, rules of thumb, and
  heuristics for reasoning about the objects and
  events of everyday life
• Domain ontology (e.g. UMLS)
• The Unified Medical Language System
• Task ontology (e.g. CPV)
• The Common Procurement Vocabulary

25
(No Transcript)
26
Entities in Ontologies

• Concepts
• Objects
• Events and processes
• Properties
• Attributes
• Relations (hold between entities)

27
Hierarchies in Ontologies

• Type hierarchy (type subtype isa
  hypernym-hyponym)
• Cat isa Animal.
• Part hierarchy (part whole is part of
  meronymy)
• Finger is a part of Hand
• Broader-narrower hierarchy (mixes part hierarchy
  and type hierarchy)

28
Ontology Terminology (Michael Denny)

• The following are synonym classes
• Concept, class, category, type, term, entity, set
  and thing.
• Instance, individual, resource, extension,
  description, object and entity.
• Relation, relationship, property, function, role,
  slot, attribute, association, criterion,
  constraint of, feature and predicate.

29
Why ontologies in CL?

• Give structure to knowledge.
• Improve human knowledge access.
• Model human associations.
• Improve the computers knowledge processing.
• Get closer to the meaning of an utterance.
• Get closer to the goal of NLP.

30
Ontology and its neighbors

• Lexicon
• A list of words is not an ontology.
• But cross-references (see also) or reading
  distinctions can be seen as ontological
  relations.
• Thesaurus (? is an ontology)
• a synonym (and hyperonym) lexicon.
• general vocabulary (not domain specific).
• mostly atomic concepts ( single word entries).
• Terminology database (? is an ontology Siemens
  TermDB)
• domain-specific vocabulary
• most often multi-lingual
• with definitions, context examples and source
  documentation
• single word and complex entries

31
Ontology and its neighbors

• Encyclopedia
• A collection of short descriptions about people,
  geographical entities, and popular terms.
• with cross-references.
• Knowledge base
• repository of formal knowledge representation.
• facts and rules (in Prolog).

32
Related terms to be explained later

• The semantic web
• Knowledge management
• Text mining
• Dublin Core
• RDF, OWL