29
Artificial Intelligence for Universal Artificial Intelligence for Universal Networking Language (UNL) Networking Language (UNL) (Perspective Bengali Language) (Perspective Bengali Language) By Deen Islam Muslim ID: 200720851 Ariful Hoque Tuhin ID: 200710698 Shohanur Rahman ID: 200720100 Under the Supervision of Under the Supervision of Md. Ahsan Arif, Sr. Lecturer Md. Ahsan Arif, Sr. Lecturer Dept. of Computer Science and Engineering Dept. of Computer Science and Engineering Asian University of Bangladesh Asian University of Bangladesh

Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

  • Upload
    varen

  • View
    52

  • Download
    0

Embed Size (px)

DESCRIPTION

Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language). By Deen Islam Muslim ID: 200720851 Ariful Hoque Tuhin ID: 200710698 Shohanur Rahman ID: 200720100. Under the Supervision of Md. Ahsan Arif, Sr. Lecturer - PowerPoint PPT Presentation

Citation preview

Page 1: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

Artificial Intelligence for Universal Networking Language Artificial Intelligence for Universal Networking Language (UNL)(UNL)

(Perspective Bengali Language)(Perspective Bengali Language)

By

Deen Islam Muslim ID: 200720851

Ariful Hoque Tuhin ID: 200710698

Shohanur Rahman ID: 200720100

Under the Supervision ofUnder the Supervision of

Md. Ahsan Arif, Sr. LecturerMd. Ahsan Arif, Sr. Lecturer

Dept. of Computer Science and EngineeringDept. of Computer Science and Engineering

Asian University of BangladeshAsian University of Bangladesh

Page 2: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

Technical TermsTechnical Terms

UNL – Universal Networking UNL – Universal Networking Language Language IAS - Institute of Advanced Studies IAS - Institute of Advanced Studies UW – Universal WordsUW – Universal Words UNLKBUNLKB – – UNL Knowledge BaseUNL Knowledge Base UNLKCIC – UNL Key Concept in UNLKCIC – UNL Key Concept in ContextContext LSLS – – Language ServersLanguage Servers

Page 3: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

ContentsContents

1.1. IntroductionIntroduction

2.2. Overview Of UNL SystemOverview Of UNL System

3.3. Mechanism of Conversion of UNL Mechanism of Conversion of UNL

4.4. System WorkingsSystem Workings

5.5. UNL ExpressionUNL Expression

6.6. UNL StructureUNL Structure

7.7. Existing ProblemsExisting Problems

8.8. Recommended SolutionRecommended Solution

9.9. ConclusionConclusion

Page 4: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

44

What is UNL?What is UNL?

Universal Networking Language (UNL) is a computer Universal Networking Language (UNL) is a computer language that enables computers to process information language that enables computers to process information and knowledge across the language barriers. and knowledge across the language barriers.

It is an artificial language that replicates the functions of It is an artificial language that replicates the functions of natural languages in human communication. natural languages in human communication.

It expresses information or knowledge in the form of It expresses information or knowledge in the form of semantic networks. semantic networks.

Unlike natural languages, UNL expressions are Unlike natural languages, UNL expressions are unambiguous. unambiguous.

Although the UNL is a language for computers, it has all the Although the UNL is a language for computers, it has all the components of a natural language.components of a natural language.

It is composed of UNL Expressions, Universal Words (UWs), It is composed of UNL Expressions, Universal Words (UWs), Relations, Attributes.Relations, Attributes.

Page 5: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

55

Overview Of UNL SystemOverview Of UNL System

It Consists OfIt Consists Of

Language Resources:Language Resources: UNLKB (Knowledge Base) – Linguistic Knowledge on concepts that UNLKB (Knowledge Base) – Linguistic Knowledge on concepts that

are common to every language.are common to every language.

Universal Word Dictionary, analysis and generation rules.Universal Word Dictionary, analysis and generation rules.

Language Servers:Language Servers: DeConverter - automatically deconverts UNL into native DeConverter - automatically deconverts UNL into native

languages. languages. EnConverters - automatically or interactively enconverts natural EnConverters - automatically or interactively enconverts natural

languages text into UNL. languages text into UNL.

Software tools: Software tools: UNL Editors - used to make UNL documents. UNL Editors - used to make UNL documents. UNL Explorers - used to view/manage UNL document by UNL Explorers - used to view/manage UNL document by

accessing UNL language servers, UNLKB & UNL Documents. accessing UNL language servers, UNLKB & UNL Documents.

Page 6: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

66

UNL Verifiers – verifies UNL expression for correctness. UNL Verifiers – verifies UNL expression for correctness. UNL Proxy servers – Provides communication with language UNL Proxy servers – Provides communication with language

servers.servers. Concept Definitions – Defines concepts in connection with Concept Definitions – Defines concepts in connection with

other concepts. other concepts. UNL Documents - the documents in which UNL expression is UNL Documents - the documents in which UNL expression is

described for each sentence of natural language.described for each sentence of natural language.

UNL SYSTEM

Page 7: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

77

Mechanism of Conversion of UNL Expressions Mechanism of Conversion of UNL Expressions ( Language Servers ) ( Language Servers )

Data Flow

Control Flow

Page 8: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

88

System WorkingsSystem Workings

People with access to the Internet can People with access to the Internet can en-converten-convert text text written in their own language into UNL expressions using written in their own language into UNL expressions using UNL editor. And likewise, any UNL expressions can be UNL editor. And likewise, any UNL expressions can be de-de-convertedconverted into a variety of native languages using the UNL into a variety of native languages using the UNL viewer (Explorer).viewer (Explorer).

Page 9: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

99

System Workings Cont..System Workings Cont..

When developing home page in Arabic, the UNL Editor When developing home page in Arabic, the UNL Editor recognizes the contents as Arabic and sends a request to recognizes the contents as Arabic and sends a request to the Arabic Language Server to “EnConvert” the text. . Once the Arabic Language Server to “EnConvert” the text. . Once the Arabic text is “EnConverted” to UNL, the Arabic the Arabic text is “EnConverted” to UNL, the Arabic Language Server sends the result back to the UNL Editor.Language Server sends the result back to the UNL Editor.

Page 10: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1010

System Workings Cont..System Workings Cont..

Home page designer can now embed UNL into their pages.Home page designer can now embed UNL into their pages.

Page 11: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1111

System Workings Cont..System Workings Cont..

When Spanish read this page, the UNL Viewer recognizes When Spanish read this page, the UNL Viewer recognizes the contents as UNL and sends a request to the Spanish the contents as UNL and sends a request to the Spanish Language Server to “DeConvert” the UNL.Language Server to “DeConvert” the UNL.

Page 12: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1212

System Workings Cont..System Workings Cont..

Once UNL is “DeConverted” to Spanish, the Spanish Once UNL is “DeConverted” to Spanish, the Spanish Language Server sends the result back to the UNL Viewer.Language Server sends the result back to the UNL Viewer.

Page 13: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1313

System Workings Cont..System Workings Cont..

If UNL is embedded in a home page, it can be read in variety If UNL is embedded in a home page, it can be read in variety ofof

languages. languages.

Page 14: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1414

UNL ExpressionUNL Expression

John is reading a novel.John is reading a novel. UNL HypergraphUNL Hypergraph

read(icl>do)

John(iof>person) novel(icl>book)

@entry.@present.@progress

agt obj

UNL ExpressionUNL Expression

[UNL][UNL]agt(read(icl>do) agt(read(icl>do) @entry.@present.@progress, , John(iof>person))John(iof>person))obj(read(icl>do) obj(read(icl>do) @entry.@present.@progress, , novel(icl>book))novel(icl>book))[/UNL]

Page 15: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1515

Universal Words (UWs)Universal Words (UWs)

A UW represents simple or compound concepts. There are A UW represents simple or compound concepts. There are two classes of UWs:two classes of UWs: Labels defined to express unit concepts, called UWs Labels defined to express unit concepts, called UWs

(Universal Words) (Universal Words) compound structures of a set of binary relations grouped compound structures of a set of binary relations grouped

together ( indicated with Compound UW-Ids)together ( indicated with Compound UW-Ids) A UW is made up of a character string (an English-language A UW is made up of a character string (an English-language

word) followed by a list of constraints.word) followed by a list of constraints. <UW>::=<Head Word>[<Constraint List>]<UW>::=<Head Word>[<Constraint List>]

Page 16: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1616

Structure of UNLStructure of UNL As a matter of example, the English sentence As a matter of example, the English sentence

‘The sky was blue?!’ can be represented in UNL ‘The sky was blue?!’ can be represented in UNL as follows: as follows:

Page 17: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1717

Structure of UNL Cont..Structure of UNL Cont..

In the example above, "sky(icl>natural world)" In the example above, "sky(icl>natural world)" and "blue(icl>color)", which represent individual and "blue(icl>color)", which represent individual concepts, are UWs; "aoj" (= attribute of an concepts, are UWs; "aoj" (= attribute of an object) is a directed binary semantic relation object) is a directed binary semantic relation linking the two UWs; and "@def", linking the two UWs; and "@def", "@interrogative", "@past", "@exclamation" and "@interrogative", "@past", "@exclamation" and "@entry" are attributes modifying UWs."@entry" are attributes modifying UWs.

Page 18: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1818

Structure of UNL Cont…Structure of UNL Cont… AttributesAttributes It represent information that cannot be conveyed by UWs and It represent information that cannot be conveyed by UWs and

relations. Normally, they represent information on tense relations. Normally, they represent information on tense (".@past", "@future", etc), reference ("@def", "@indef", etc), (".@past", "@future", etc), reference ("@def", "@indef", etc), modality ("@can", "@must", etc), focus ("@topic", "@focus", modality ("@can", "@must", etc), focus ("@topic", "@focus", etc), and so on.etc), and so on.

Relations Relations Relations are expected to represent semantic links between Relations are expected to represent semantic links between

words in every existing language. They can be ontological words in every existing language. They can be ontological (such as "icl" and "iof" referred to above), logical (such as (such as "icl" and "iof" referred to above), logical (such as "and" and "or") and thematic (such as "agt" = agent, "ins" = "and" and "or") and thematic (such as "agt" = agent, "ins" = instrument, "tim" = time, "plc" = place, etc). There are instrument, "tim" = time, "plc" = place, etc). There are currently 46 relations in the UNL Specs, and they define the currently 46 relations in the UNL Specs, and they define the syntax of UNL. syntax of UNL.

Page 19: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

1919

Existing ProblemExisting Problem

Spell CheckingSpell Checking Absence of Spell Checker Program.Absence of Spell Checker Program.

for example:for example:

A simple English sentence in a right spelling form:A simple English sentence in a right spelling form:

“ “ I live in Bangladesh “I live in Bangladesh “

UNL En-conversion:UNL En-conversion:

aoj(live(icl>inhabit>be,aoj>living_thing,plc>place).@eaoj(live(icl>inhabit>be,aoj>living_thing,plc>place).@entry.@present,i(icl>person))ntry.@present,i(icl>person))plc(live(icl>inhabit>be,aoj>living_thing,plc>place).@eplc(live(icl>inhabit>be,aoj>living_thing,plc>place).@entry.@present,bangladesh(iof>asian_country>thing))ntry.@present,bangladesh(iof>asian_country>thing))

Here “Bangladesh” is assigned to the UW of “Asian_Country”. Here “Bangladesh” is assigned to the UW of “Asian_Country”.

Page 20: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2020

But if we if we type the “But if we if we type the “BangladeshBangladesh” word in wrong ” word in wrong spelling like “spelling like “BanladeshBanladesh” then it convert that word in ” then it convert that word in such form:such form:

aoj(live(icl>inhabit>be,aoj>living_thing,plc>place).@entaoj(live(icl>inhabit>be,aoj>living_thing,plc>place).@entry.@present,i(icl>person))ry.@present,i(icl>person))plc(live(icl>inhabit>be,aoj>living_thing,plc>place).@entplc(live(icl>inhabit>be,aoj>living_thing,plc>place).@entry.@present,banladesh)ry.@present,banladesh)

It does not define any UW for “Bangladesh”.It does not define any UW for “Bangladesh”. There for wrong conversion can be occurred. There for wrong conversion can be occurred.

Existing Problem Cont….Existing Problem Cont….

Page 21: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2121

Existing Problem Cont..Existing Problem Cont..

Maintaining Exact Grammatical PatternMaintaining Exact Grammatical Pattern

According to Russian Language Server, for a single sentence or like some According to Russian Language Server, for a single sentence or like some compound sentences it works fine. But we have found some crucial problems compound sentences it works fine. But we have found some crucial problems when en-converting and de-converting some multiple sentences.when en-converting and de-converting some multiple sentences.

For An example, using this sentence:For An example, using this sentence:

I like rice and I play football.I like rice and I play football. After performing the En-conversion and De-conversion process this will remain After performing the En-conversion and De-conversion process this will remain

same:same:

I like rice and I play football.I like rice and I play football.

But, Problem occurs when using some multiple sentences like:But, Problem occurs when using some multiple sentences like:

We play football. We try to win our every match.We play football. We try to win our every match.

Page 22: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2222

En-converting this sentence in UNL forms:En-converting this sentence in UNL forms:

agt(try(icl>attempt>do,agt>person,obj>uw).@entry.@present,we(iagt(try(icl>attempt>do,agt>person,obj>uw).@entry.@present,we(icl>group):01.@pl)cl>group):01.@pl)pos:01(match(icl>contest>thing),we(icl>group):02)pos:01(match(icl>contest>thing),we(icl>group):02)fictit(try(icl>attempt>do,agt>person,obj>uw).@entry.@present,evfictit(try(icl>attempt>do,agt>person,obj>uw).@entry.@present,every(icl>quantity,per>thing))ery(icl>quantity,per>thing))obj:01(win(icl>prize>do,agt>thing,obj>thing,scn>thing).@entry,mobj:01(win(icl>prize>do,agt>thing,obj>thing,scn>thing).@entry,match(icl>contest>thing))atch(icl>contest>thing))obj(try(icl>attempt>do,agt>person,obj>uw).@entry.@present,:01)obj(try(icl>attempt>do,agt>person,obj>uw).@entry.@present,:01)

After De-converting this UNL form back to English:After De-converting this UNL form back to English: We Play Football. We try win we match.We Play Football. We try win we match.

Existing Problems Cont…Existing Problems Cont…

Page 23: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2323

Existing Problems Cont…Existing Problems Cont…

Lack of Exact Rules and Absence of AILack of Exact Rules and Absence of AI For an example, we can use this sentence :For an example, we can use this sentence :

I don’t like play footballI don’t like play football

According to Russian Language Server, en-conversion takes this According to Russian Language Server, en-conversion takes this sentence in UNL form like this way:sentence in UNL form like this way:

aoj(like(icl>please>be,equ>enjoy,obj>uw,aoj>person).@[email protected](like(icl>please>be,equ>enjoy,obj>uw,aoj>person).@entry.@not.@present,i(icl>person))@present,i(icl>person))obj:01(play(icl>compete>do,agt>thing,obj>uw,ptn>thing).@entry,fooobj:01(play(icl>compete>do,agt>thing,obj>uw,ptn>thing).@entry,football(icl>field_game>thing))tball(icl>field_game>thing))obj(like(icl>please>be,equ>enjoy,obj>uw,aoj>person).@[email protected](like(icl>please>be,equ>enjoy,obj>uw,aoj>person).@entry.@not.@present,:01)@present,:01)

After De-convert this UNL in natural language (English):After De-convert this UNL in natural language (English): I not like play footballI not like play football

Page 24: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2424

Existing Problems Cont…Existing Problems Cont…

• There are some more example that occurs similar results :

• My name is Casper and I don't like to play football.

• This is Rosy and This is her School Bag.

• I have so many friends and among of them beauty is my best friend.

Page 25: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2525

Existing Problems Cont...Existing Problems Cont...

Verb Representation Problem:Verb Representation Problem:

Let take a look at this example: I am a good boy.

En-converted UNL form:

aoj(boy(icl>child>person,ant>girl).@entry.@in

def.@present,i(icl>person))

mod(boy(icl>child>person,ant>girl).@entry.@indef.@present,good(icl>adj,ant>bad))

De-Converted English:

I Be a good boy.

Page 26: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2626

Recommended SolutionsRecommended Solutions

Implementation of a good spell checker. Implementation of a good spell checker.

Rearrange the Sentence in Exact Grammatical Rearrange the Sentence in Exact Grammatical Pattern.Pattern.

Implementation of AI and Exact Rules.Implementation of AI and Exact Rules.

Determine the Exact Form of Verb Determine the Exact Form of Verb Representation.Representation.

Page 27: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2727

ReferencesReferences http://http://www.unu.eduwww.unu.edu//

United Nations UniversityUnited Nations University http://www.ias.unu.edu/http://www.ias.unu.edu/

United Nations University, Institute of Advanced StudiesUnited Nations University, Institute of Advanced Studies http://www.unl.ru/http://www.unl.ru/

Universal Networking Language RussianUniversal Networking Language Russian http://www.unl.ias.unu.edu/publications/gm/index.htmhttp://www.unl.ias.unu.edu/publications/gm/index.htm

Book of Universal Networking LanguageBook of Universal Networking Language http://www.iai.uni-sb.de/UNL/unl-specs.htmlhttp://www.iai.uni-sb.de/UNL/unl-specs.html

Official UNL Specification (unicode version)Official UNL Specification (unicode version) http://www.iai.uni-sb.de/UNL/unl-iai.htmlhttp://www.iai.uni-sb.de/UNL/unl-iai.html

German UNL HomepageGerman UNL Homepage http://unl.ilc.pi.cnr.it/http://unl.ilc.pi.cnr.it/

Italian UNL HomepageItalian UNL Homepage http://www.vai.dia.fi.upm.es/projects/unl_in.htmlhttp://www.vai.dia.fi.upm.es/projects/unl_in.html

Spanish UNL HomepageSpanish UNL Homepage http://www.links.nectec.or.th/unl/UnuWeb/UNLP_Homepage.htmhttp://www.links.nectec.or.th/unl/UnuWeb/UNLP_Homepage.htm

Thai UNL HomepageThai UNL Homepage http://www.unl.rss.gov.jo/http://www.unl.rss.gov.jo/

Arabic UNL HomepageArabic UNL Homepage http://www.ailab.lv/unl.htmhttp://www.ailab.lv/unl.htm

Latvian UNL HomepageLatvian UNL Homepage http://www.it.iitb.ac/ it632, 2005. http://www.it.iitb.ac/ it632, 2005.

Pushpak Bhattacharyya. Language Technology for the Web. Pushpak Bhattacharyya. Language Technology for the Web. http://www.iitb.ac.inhttp://www.iitb.ac.in Anupama Dutta. Anupama Dutta. Multilingual and meaning based search engines. Indian Multilingual and meaning based search engines. Indian

Institute of Technology, Bombay, India, November 2003.Institute of Technology, Bombay, India, November 2003.

Page 28: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2828

Thank YouThank You

Page 29: Artificial Intelligence for Universal Networking Language (UNL) (Perspective Bengali Language)

2929

Question and Answer