124.11.lec17

124.11.lec17 - Dan Jurafsky Lecture 17: Machine Transla8on:...

Info iconThis preview shows pages 1–11. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Dan Jurafsky Lecture 17: Machine Transla8on: Sta8s8cal MT CS 124/LINGUIST 180: From Languages to Information Slides from Ray Mooney Statistical MT Surprising: Intui8on comes from the impossibility of transla8on Consider Hebrew adonai roi (“the lord is my shepherd”) for a culture without sheep or shepherds! Something fluent and understandable, but not faithful: “The Lord will look aKer me” Something faithful, but not fluent and nautral “The Lord is for me like somebody who looks aKer animals with co¡on-like hair” What makes a good translation Translators oKen talk about two factors we want to maximize: Faithfulness or ¡delity How close is the meaning of the transla8on to the meaning of the original (Even beLer: does the transla8on cause the reader to draw the same inferences as the original would have) Fluency or naturalness How natural the transla8on is, just considering its fluency in the target language Statistical MT: Faithfulness and Fluency formalized! Best-transla8on of a source sentence S: Developed by researchers who were originally in speech recogni8on at IBM Called the IBM model ˆ T = argmax T fluency( T )faithfulness( T , S ) The IBM model Hmm, those two factors might look familiar… Yup, it’s Bayes rule: ˆ T = argmax T fluency( T )faithfulness( T , S ) ˆ T = argmax T P ( T ) P ( S | T ) More formally Assume we are transla8ng from a foreign language sentence F to an English sentence E: F = f 1 , f 2 , f 3 ,…, f m We want to ¡nd the best English sentence E-hat = e 1 , e 2 , e 3 ,…, e n E-hat = argmax E P(E|F) = argmax E P(F|E)P(E)/P(F) = argmax E P(F|E)P(E) Translation Model Language Model The noisy channel model for MT Fluency: P(T) How to measure that this sentence That car was almost crash onto me is less fluent than this one: That car almost hit me. Answer: language models (N-grams!) For example P(hit|almost) > P(was|almost) But can use any other more sophis8cated model of grammar Advantage: this is monolingual knowledge! Faithfulness: P(S|T) French: ça me plait [that me pleases] English: that pleases me- most fluent I like it I’ll take that one How to quan8fy this? Intui8on: degree to which words in one sentence are plausible transla8ons of words in other sentence Product of probabili8es that each word in target sentence would generate each word in source sentence. Faithfulness P(S|T) Need to know, for every target language word, probability of it mapping to every source language word. How do we learn these probabili8es? Parallel texts!...
View Full Document

This document was uploaded on 06/01/2011.

Page1 / 100

124.11.lec17 - Dan Jurafsky Lecture 17: Machine Transla8on:...

This preview shows document pages 1 - 11. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online