Title: MACHINE BRIDGE FOR SOUTH INDIAN LANGUAGES

1

• Title MACHINE BRIDGE FOR SOUTH INDIAN LANGUAGES
• Proposers
• 1. Dr.S.Rajendran
• Head, Department of Linguistics
• Tamil University
• Thanjavur 613 005
• 2. Dr.S.Baskaran
• Head, Department of Computer Science
• Tamil University, Thanjavur 613 005
• Language Pair Tamil to Malaylam, Telugu, and
  Kannda

2

• Components that will be implemented
• Transliteration
• POS Tagging Engine
• Chunking Engine
• Morphological Analyzers Generators
• Bilingual (Multi-lingual) dictionary
• Transfer Engine
• Annotated Corpora Preparation and Validation
• Evaluation of MT system

3

• Tamil Malayalam, Kannada and Telugu
• Transliteration
• The languages show certain amount of similarity
  in their form being cognate languages.Similarities
  can be found in the case of borrowed items also.
• Interstate governmental communication and
  documentation

4

• Tamil Malayalam, Kannada and Telugu
• POS Tagging Engine
• It will be independent of the language. Hidden
  Markov models (HMMs) would be used for building
  tagging engine.
• Information given by Morphological analyzer will
  be used for POS tagging.
• 85 going up to 95
• Interstate governmental communications and
  documentation

5

• Tamil Malayalam, Kannada and Telugu
• Chunking Engine
• Same as the POS tagging engine and will be based
  on HMMs. Shallow parsing techniques will be
  implemented.
• 85 going up to 95
• Interstate governmental communications and
  documentation

6

• Tamil Malayalam, Kannada and Telugu
• Morphological Analyzers Generators
• Finite state automata will be used to do parsing
  and generation.
• 85 going upto 95.
• Interstate governmental communication and
  documents.

7

• Tamil Malayalam, Kannada and Telugu
• Bilingual (Multi-lingual) dictionary
• Tamil-Malayalam, Tamil-Kannada, and Tamil-Telugu
  translation dictionaries will be prepared.
• 85 going up to 95
• Interstate governmental communications and
  documentation

8

• Tamil Malayalam, Kannada and Telugu
• Transfer Engine
• Lexical Transfer grammar will be used.
• 85 going up to 95
• Interstate governmental communications and
  documentation

9

• Tamil Malayalam, Kannada and Telugu
• Annotated Corpora Preparation and Validation
• Automatic annotation followed by manual
  annotation. Training techniques will also be
  implemented.
• 85 going up to 95
• Interstate governmental communications and
  documentation

10

• Tamil Malayalam, Kannada and Telugu
• Evaluation of MT system
• Proper evaluation techniques will be evolved in
  course of time.

11

• THANK YOU

12

• Language pair Tamil Malayalam, Kannada, Telugu
• Name of the component
• POS Tagging Engine
• Techniques that will be used
• Estimation of the expected performance
• Preferred domain of the performance
• Other evaluation matrices

13

• Language pair Tamil Malayalam, Kannada, Telugu
• Name of the component
• POS Tagging Engine
• Techniques that will be used
• Estimation of the expected performance
• Preferred domain of the performance
• Other evaluation matrices

14

• Language pair Tamil Malayalam, Kannada, Telugu
• Name of the component
• POS Tagging Engine
• Techniques that will be used
• Estimation of the expected performance
• Preferred domain of the performance
• Other evaluation matrices

15

• Language pair Tamil Malayalam, Kannada, Telugu
• Name of the component
• POS Tagging Engine
• Techniques that will be used
• Estimation of the expected performance
• Preferred domain of the performance
• Other evaluation matrices

16

• Language pair Tamil Malayalam, Kannada, Telugu
• Name of the component
• POS Tagging Engine
• Techniques that will be used
• Estimation of the expected performance
• Preferred domain of the performance
• Other evaluation matrices