Title: Creation of English and Hindi Verb Hierarchies and their Application to Hindi WordNet Building and English-Hindi MT
1Creation of English and Hindi Verb Hierarchies
and their Application to Hindi WordNet Building
andEnglish-Hindi MT
- Debasri Chakrabarti, Gajanan Krishna Rane,
Pushpak Bhattacharyya. - Computer Science and Engineering Department,
- Indian Institute of Technology, Bombay,
- Mumbai, 40076, India.
- debasri,gkrane,pb_at_cse.iitb.ac.in
2Introduction
- Verb hierarchy
- creation of the verb hierarchy for English and
Hindi - verbs.
- organized according to semantics and syntax
- semantic hierarchy - through the super-ordinate
terms and the inbuilt ontology of the UNL KB. - syntactic information- through UNL case relations
- System is based on
- English verb classes and their alternation
(Levin) - UNL System UW Manual, Knowledge base (KB)
specification - Semantic relations of English WordNet
3Levins Class of English verbs
- Classification of the English verbs
- Adopted from English Verb Classes and
Alternation of Beth Levin. - Details of Levins work
- Levins classification of the English verb is
the most significant and celebrated work. - Assumption underlying Levins work
- Syntactic behavior of a verb is semantically
determined - Levin investigated and exploited this hypothesis
for about 3200 English verbs.
4Details of Levins work
- Verb Classes
- Preliminary Investigation
- considerable correlation between some facets of
the - semantics of verbs and their syntactic behavior
- 200 semantic classes defined in Levins system
- each class share a number of alternations
- Example of verb classes
- verbs of putting , verbs of communication,
correspond - verbs etc.
5The Universal Networking Language (UNL)
- Universal Networking Language (UNL)
- electronic language for computers to express and
- exchange information.
- UNL system consists
- Universal words (UW) Vocabulary of UNL
- Relations, attributes Syntax of UNL
- UNL knowledge base (KB) Semantics of UNL
6The Universal Networking Language
- UNL represents information
- sentence-by-sentence as a hyper-graph
- concepts as nodes and relations as arcs
- Sentence is a hyper-graph
- a node in the structure can itself be a graph
- the node is called a compound word (CW)
7Graphical representation in UNL
_at_ entry _at_ present
eat (iclgtdo)
agt
ins
obj
John (iofgtperson)
spoon (iclgtartifact)
rice (iclgtfood)
John eats rice with a spoon
8Verbal Concepts in UNL
- Verbal concepts in the UNL system are organized
into three categories - (iclgtdo) for defining the concept of an event
which is - caused by something or someone
- change (iclgtdo) as in She changed the dress
- (iclgtoccur) for defining the concept of an event
that - happens of its own accord
- change (iclgtoccur) as in The weather
will change - (iclgtbe) for defining the concept of a state verb
- remember (iclgtbe) as in Do you
remember me?
9Verbal Concepts in UNL
- do(agtgtthing,golgtthing,iclgtdo,objgtthing,ptngtth
ing,srcgtthing) - do(agtgtvolitional thing,iclgtdo(agtgtthing))
- do(agtgtliving thing,iclgtdo(agtgtvolitional
thing)) - do(agtgthumangtliving thing,iclgtdo(agtgtlivi
ng thing)) - do(agtgtthing,golgtthing,iclgtdo
, objgtthing,ptngtthing,srcgtthing)
Partial hierarchical structure for do
10do in UNL KB
- Semantic hierarchy in terms of the inbuilt
ontology in KB - do(agtgtthing,golgtthing,iclgtdo,objgtthing,ptngtth
ing,srcgtthing) - do(iclgtdo(agtgtthing,golgtthing,objgtthing),golgt
abstract thing,objgtabstract thing) - do(iclgtdo(agtgtthing,golgtabstract
thing,objgtabstract
- thing),golgtcustomgtabstract
thing,ob jgtcustomgtabstract thing) -
do(golgtabstract thing)
do(golgtcustom)
do(golgtthing)
11Creation of the verb hierarchy
- First, a particular verb class is selected from
Levin. - Next the class is categorized according to the
UNL format - Parent node of a class is obtained through
English wordnet and various dictionaries
12Creation of the verb hierarchy
- put
- Put your clothes in the cupboard.
- (to put something into a certain place)
- (iclgtmove(agtgtperson,objgtconcrete
thing,golgtplace) - (loc_prepin/on/into/under/over)
- VTRANS, VOA-ACT
- hang
- He hanged the wallpaper on the wall.
- (to suspend or fasten something so that it
is held up from above and - not supported from below)
- (iclgtputgtmove(agtgtperson,objgtconcrete
thing,golgtplace) - (loc_prepfrom/on)
- VTRANS, VOA-ACT
- Partial hierarchy of the put class
13Verb Hierarchy in Hindi
- ???? rakhanaa rkHna
- put Put your things here. (to put something
into a certain place) - (iclgtact(agtgtperson,objgtconcrete thing,golgtplace)
- ???? ????? ???? ?? ???? (pna saman yha) pr
rkHo) apanaa saamaana - yahaa par rakho)
- (adv_plc (????/???? / yha) / vha) loc_postp
(?? pr) - ????, ????? rkHna, sjana rakhanaa ,
sajaanaa - arrange he arranged the books here.(to put
into a proper or systematic - manner)
- (iclgtputgtact(agtgtperson,objgtthing)
- ???? ??????? ?? ???? ?? ????? ???? usne kitabo)
ko yha) pr sjakr rkHa.) - (usne kitabo ko yahaa par sajaakar rakhaa.)
- (adv_man (?????, sjakr ???? ??, krm se))
(adv_plc (????/???? / yha) / - vha) )) loc_postp( ?? pr)
14Verb Hierarchy in Hindi
- Syntax frames specified for the put class in
English - (adv_plchere/there)
- (loc_prep)
- Sentence frames for put in Hindi
- adv_man
- adv_plc adv_man
- loc_postp adv_man
English Hindi
adv_plc (here / there) adv_man (?????, sjakr ???? ??, krm se etc )
loc_prep (in, inside, on etc) adv_plc(????/???? / yha) / vha)) loc_postp(?? pr)adv_man (?????, sjakr ???? ??, krm se etc)
loc_postp(?? ???, ke upr etc)adv_man (?????, sjakr etc)
15Verb hierarchy and the Hindi WordNet
- Application of the hierarchy in the Hindi wordnet
- will help in determining
- semantic relations like hypernymy and troponymy
- syntactic frames
- Application of the hierarchy in the Hindi wordnet
- revealed facts like
- difference in the representations for troponyms
in Hindi and English - reclassifications of the verbs in Hindi
-
16Representations of Troponyms
English sentence Hindi sentence
put put your things here. ???? rkHna ???? ????? ???? ?? ???? (pna sman yha) pr rkHo)
pile pile your books up on the shelves. ----- ???? ???? ??? ?? ?? ??? ?? ????? ????((((((usne kHane me) ek ke Upr ek saman rkHa)
cram she cram the books into the suitcase. ----- ???? ????? ?? ????? ???? ????? ?????? ????(usne bakse ke a)dr sari kitab tHu)sakr rkHI))
17Classification of Hindi Verbs
conjunct
18Classification of the Hindi Verbs
- Simple verbs
- ????(kHana) to eat
- Compound verbs
- ??? ?????(gir pêna) to fall down
- Conjunct verbs
- noun verb ???? ???? (armbH krna) to
start - adjective verb ???? ???? (Sant krna) to calm
down - adverb verb ????? ???? (utakr rkHna) to
lift
19Reclassification of the Hindi verbs
- Sentence frames of the verbs reveals
- only noun verb conjunct is a true conjunct
- Hence, a re-classification of the verbs is
needed
20Application in NLP
- The application of the verb hierarchy in NLP
- gives semantic hierarchy of a verbal concept
- enumerates syntactic details of a verb
- UNL based MT will be immensely benefited
- possible UNL relations that appear with a concept
is specified -
21Application in MT
Verb Sentence Frame UNL Relations
fight Sam and Sue fought. conj_and agtgtperson
fight Sam was fighting with Sue. prep_accompanimentwith agtgtperson, ptngtperson
fight The tribesmen fought each other. -prep_with agtgtperson, objgtperson
22Conclusion
- System statistics
- approximately 3000 English verbs
- approximately 5500 UWs
- Common English verbs are dealt with
- tested against British National Corpus
- Coverage of both English and Hindi verbs is
increasing everyday - Visualizer and an application programming
interface for the verb knowledge bases in both
the languages are under construction