CS626-449: NLP, Speech and Web-Topics-in-AI

1
CS626-449 NLP, Speech and Web-Topics-in-AI

• Pushpak BhattacharyyaCSE Dept., IIT Bombay
• Lecture 35 Semantic Relations UNL Towards
  Dependency Parsing

2
Web at a glance

• Google indexes more than 8 billion pages
• Dominated by English
• Large part of world is deprived of this knowledge
  

3
Search Engines Today

• Keyword based
• Irrelevant results
• Meaning not taken into account
• Language specific
• No search possible across language
• No translation possible

4
Future of the World Wide Web
User
User
Translation interface
Translation interface
WWW
User
User
5
Features

• Meaning based
• More relevant results
• Multilingual
• Query in English
• Fetch document in Hindi
• Show it in English

6
Machine Translation

• Translate from one language to other
• Two approaches
• Direct
• One step
• Using Intermediate language
• Two step

7
Interlingua

• Interlingua
• Intermediate language for machine translation
• Step one
• Convert from source language text to interlingua
• Step two
• Produce target langauge text from interlingua
• UNL an interlingua in UNL system

8
Internet for the Masses
English interface
Spanish interface
Internet
Spanishviewer
English viewer
Hindi interface
Hindi viewer
9
A Semantic Graph
The student bought a new computer in June.
10
UNL representation
Representation of Knowledge
Ram is reading the newspaper
11
Knowledge Representation
UNL Graph - relations
read
agt
obj
Ram
newspaper
12
Knowledge Representation
UNL Graph - UWs
read(iclgtinterpret)
obj
agt
newspaper(iclgtprint_media)
Ram(iofgtperson)
13
Knowledge Representation
UNL graph - attributes
_at_entry _at_present _at_progress
read(iclgtinterpret)
obj
agt
_at_def
newspaper(iclgtprint_media)
Ram(iofgtperson)
Ram is reading the newspaper
14
The boy who works here went to school
Another Example
15
UNL System
16
The World-wide Universal Networking Language
(UNL) Project
Marathi
English
Russian
UNL
Spanish
Japanese
Hindi

• Language independent meaning representation.

17
The UNL System An Overview
18
Universal Networking Language

• Universal Words (UWs)
• Relations
• Attributes
• Knowledge Base

19
UNL Graph
He forwarded the mail to the minister.
20
UNL Expression

• agt (forward(iclgtsend)._at_ entry _at_ past,
  he(iclgtperson))
• obj (forward(iclgtsend)._at_ entry _at_ past,
  minister(iclgtperson))
• gol (forward(iclgtsend )._at_ entry _at_ past,
  mail(iclgtcollection). _at_def)