Indonesian English Parallel Texts for StatisticalMachineTranslation - PowerPoint PPT Presentation

Title: Indonesian English Parallel Texts for StatisticalMachineTranslation


1
Indonesian English Parallel Texts for
Statistical Machine Translation
( Hammam Riza, Adiansya Prasetya, Henky Mulyadi )
2
Background

• The Republic of Indonesia is an
• Archipelago of 13,000 islands that spread over an
  area of 1,900,000 square kilometers
• Population of 245,000,000 (July. 2006 estimated)
• 7 growth of the GDP was recorded on per year
• Indonesian economy and political conditions are
  gradually stabilizing
• Indonesia is back on the track to become an
  industrialized nation
• Bahasa Indonesia became the formal language of
  the country, uniting its citizens who speak
  different languages
• Bahasa Indonesia has become the language that
  bridges the language barrier among Indonesians
  who have different mother-tongues
• The vocabulary of bahasa Indonesia has been
  extensively influenced by outside languages,
  especially Sanskrit, Arabic, Chinese, Dutch, and
  English, as well as local languages such as
  Javanese and Batavian

3
Research Topics
Tourism in Asia
Social Service In Asia
Safety and Security in Asia
Archiving of Asian Language
Education In Asia
Business in Asia
Multi-lingual Speech and Language Transcription
and formats
Multi-lingual Speech translation
Multi-lingual Speech Translation
Multi-lingual Speech Transcription
Multi-lingual Speech and Text Archive
Parallel Corpus ( Synonymous Speech Text)
English Language
Indonesian Language
SpeechText
SpeechText
Parallel Corpus Format
Dictionary
4
Corpus Collection and Processing

• Data collection schema

Antara News Agency (oracle DB)
Corpus Indonesian- English
Selection Transformation
Selected DB (SQL 2000)
Alignment Article
Alignment Sentences
Corpus Indonesian- English
Web News
Collection Alignment Sentences
Toggle Cleaning
Conversion to Text
Indonesian Text
Clean Corpus
English Text
5
Translation of SMT System (1)

• A. Translation model
• gt SRI Language Modeling Toolkit which extracts a
  3-gram language model from the data. Besides the
  SRILM distribution, you will also need the
  following freely available tools ANSI-C/C
  compiler, gcc version 3.4.3 or higher, GNU make,
  GNU gawk, GNU gzip, Tcl, CYGWIN porting layer, to
  build SRILM on a Microsoft Windows system.
• gt Functionalities of SRILM
• Generate the n-gram count file from the corpus
• Train the language model from the n-gram count
  file
• Calculate the test data perplexity using the
  trained language model

Training corpus
ngram corpus
Count file
Lexicon
ngram count
LM
Test data
ngram
ppl
6
Translation of SMT System (2)

• B. Language Model
• bin contains GIZA which is an implementation
  based on the IBM models, and mkcls which divides
  words into probabilistically based classes.
• In order to compile GIZA you may need
• a recent version of the GNU compiler (2.95 or
  higher)
• a recent version of assembler and linker which do
  not have restrictions with respect to the length
  of symbol names
• corpus is where the data should be placed when
  training the translation model.

source

• Translation Model
• Program (SRILM)
• Compiler

Data preparations
Pharaoh SMT Generation System
Train Phrase Model
target
7
Testing System Performance (1)

• Sentence translation process use decoder Pharaoh
  
• Files used for translasi are
• pharaoh (executable)
• pharaoh.ini
• xkalimat.lm
• phrase-table
• Example
• Type the command like this
• echo Can I check in now ./pharaoh f
  ./pharaoh.ini gt OUT
• The process will yield file OUT, to see result
  type
• cat OUT
• Presented results Dapatkah saya check in
  sekarang

8
Testing System Performance (2)
Testing Performance
Bleu Score Sample 275,000 sentence bleu score
is 0.878
9
Thank you...
sunset in Kuta, Bali