Statistical machine translation part iii phrase based smt decoding
Download
1 / 61

Statistical Machine Translation Part III – Phrase- based SMT / Decoding - PowerPoint PPT Presentation


  • 116 Views
  • Uploaded on

Statistical Machine Translation Part III – Phrase- based SMT / Decoding. Alex Fraser Institute for Natural Language Processing University of Stuttgart 2008.07.23 EMA Summer School. Outline. Phrase- based translation Log-linear model Tuning log-linear model Decoding.

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about ' Statistical Machine Translation Part III – Phrase- based SMT / Decoding' - marek


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
Statistical machine translation part iii phrase based smt decoding

Statistical Machine TranslationPart III – Phrase-based SMT / Decoding

Alex Fraser

Institute for Natural Language Processing

University of Stuttgart

2008.07.23 EMA Summer School


Outline
Outline

  • Phrase-basedtranslation

  • Log-linear model

  • Tuning log-linear model

  • Decoding


Slide fromKoehn 2008


Slide fromKoehn 2008


Language model
Language Model

  • Usually a trigramlanguage model isusedfor p(e)

  • P(the man wenthome) = p(the | START) p(man | START the) p(went | the man) p(home | man went)

  • Language modelswork well forcomparingthegrammaticalityofstringsofthesame length

    • However, whencomparingshortstringswithlongstringstheyfavorshortstrings

    • Forthisreason, a veryimportantcomponentofthelanguage model isthelengthbonus

      • Thisis a constant > 1 multipliedforeach English word in thehypothesis


d

ModifiedfromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Outline1
Outline

  • Phrase-basedtranslation

  • Log-linear model

  • Tuning log-linear model

  • Decoding


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Outline2
Outline

  • Phrase-basedtranslation model

  • Log-linear model

  • Tuning log-linear model automatically

  • Decoding


Outline3
Outline

  • Phrase-basedtranslation model

  • Log-linear model

  • Tuning log-linear model automatically

  • Decoding

    • Basic phrase-baseddecoding

    • Dealingwithcomplexity

      • Recombination

      • Pruning

      • Future costestimation

    • Decoding output


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Slide fromKoehn 2008


Assignment 2
Assignment 2

  • Build a stateoftheartphrase-based SMT system!

    • German to English or French to English

    • Using a smallamountofdata

    • Thisis a „learningbydoing“ exercise

  • See myhomepage again


Thankyou!


ad