Machine Translation Contd Prof. Sameer Singh CS 295: STATISTICAL - - PowerPoint PPT Presentation

▶
machine translation contd
SMART_READER_LITE
LIVE PREVIEW

Machine Translation Contd Prof. Sameer Singh CS 295: STATISTICAL - - PowerPoint PPT Presentation

Dec 31, 2022 467 likes •1.04k views

Machine Translation Contd Prof. Sameer Singh CS 295: STATISTICAL NLP WINTER 2017 March 2, 2017 Based on slides from Dan Klein, Philipp Koehn, Jacob Eisenstein, and everyone else they copied from. Omer Levy AI/ML Seminar Monday, March 6 th

slide-1
SLIDE 1

Machine Translation Contd

  • Prof. Sameer Singh

CS 295: STATISTICAL NLP WINTER 2017

March 2, 2017

Based on slides from Dan Klein, Philipp Koehn, Jacob Eisenstein, and everyone else they copied from.

slide-2
SLIDE 2

Omer Levy

CS 295: STATISTICAL NLP (WINTER 2017) 2

Understanding Word Embeddings

AI/ML Seminar Monday, March 6th 1pm-2pm DBH 4011 Meeting with Graduate Students 4:00-4:45pm Room TBA (email me)

slide-3
SLIDE 3

Upcoming…

  • Homework 4 is due on March 13
  • Write-up and data releasing soon.

Homework

  • Status report due in 1 weeks: March 7, 2017
  • Instructions coming today!
  • Almost final report, only 5 pages

Project

  • Paper summaries: March 14
  • Summary 1 graded

Summaries

CS 295: STATISTICAL NLP (WINTER 2017) 3

slide-4
SLIDE 4

Outline

EM-Algorithm for Alignments Phrase-Based MT Decoding Algorithms Syntax-Based MT

CS 295: STATISTICAL NLP (WINTER 2017) 4

slide-5
SLIDE 5

Outline

EM-Algorithm for Alignments Phrase-Based MT Decoding Algorithms Syntax-Based MT

CS 295: STATISTICAL NLP (WINTER 2017) 5

slide-6
SLIDE 6

Parameters of the IBM Models

CS 295: STATISTICAL NLP (WINTER 2017) 6

slide-7
SLIDE 7

Parameters of the IBM Models

CS 295: STATISTICAL NLP (WINTER 2017) 7

slide-8
SLIDE 8

Translation from Alignments

CS 295: STATISTICAL NLP (WINTER 2017) 8

slide-9
SLIDE 9

Alignments from Translation

CS 295: STATISTICAL NLP (WINTER 2017) 9

slide-10
SLIDE 10

Expectation Maximization

CS 295: STATISTICAL NLP (WINTER 2017) 10

Expectation Maximization

slide-11
SLIDE 11

Example

CS 295: STATISTICAL NLP (WINTER 2017) 11

slide-12
SLIDE 12

Example

CS 295: STATISTICAL NLP (WINTER 2017) 12

slide-13
SLIDE 13

Word-based MT: Problems

CS 295: STATISTICAL NLP (WINTER 2017) 13

Multi-word Alignments Non-compositionality Phrasal Translations

slide-14
SLIDE 14

Outline

EM-Algorithm for Alignments Phrase-Based MT Decoding Algorithms Syntax-Based MT

CS 295: STATISTICAL NLP (WINTER 2017) 14

slide-15
SLIDE 15

The Vauquios Triangle

CS 295: STATISTICAL NLP (WINTER 2017) 15

slide-16
SLIDE 16

Phrase-based MT

CS 295: STATISTICAL NLP (WINTER 2017) 16

Mary did not slap the green witch

slide-17
SLIDE 17

Phrase Lexicon

CS 295: STATISTICAL NLP (WINTER 2017) 17

slide-18
SLIDE 18

Learning Phrasal Alignments

CS 295: STATISTICAL NLP (WINTER 2017) 18

slide-19
SLIDE 19

Learning Phrasal Alignments

CS 295: STATISTICAL NLP (WINTER 2017) 19

slide-20
SLIDE 20

Learning Phrasal Alignments

CS 295: STATISTICAL NLP (WINTER 2017) 20

slide-21
SLIDE 21

Phrasal Alignments

CS 295: STATISTICAL NLP (WINTER 2017) 21

Should contain all the alignment points for covered words

slide-22
SLIDE 22

Learning Phrasal Alignments

CS 295: STATISTICAL NLP (WINTER 2017) 22

(Maria, Mary), (no, did not), (slap, daba una bofetada), (a la, the), (bruja, witch), (verde, green)

slide-23
SLIDE 23

Learning Phrasal Alignments

CS 295: STATISTICAL NLP (WINTER 2017) 23

(Maria, Mary), (no, did not), (slap, daba una bofetada), (a la, the), (bruja, witch), (verde, green) (Maria no, Mary did not), (no daba una bofetada, did not slap), (daba una bofetada a la, slap the), (bruja verde, green witch)

slide-24
SLIDE 24

Learning Phrasal Alignments

CS 295: STATISTICAL NLP (WINTER 2017) 24

(Maria, Mary), (no, did not), (slap, daba una bofetada), (a la, the), (bruja, witch), (verde, green) (Maria no, Mary did not), (no daba una bofetada, did not slap), (daba una bofetada a la, slap the), (bruja verde, green witch) (Maria no daba una bofetada, Mary did not slap), (no daba una bofetada a la, did not slap the), (a la bruja verde, the green witch)

slide-25
SLIDE 25

Learning Phrasal Alignments

CS 295: STATISTICAL NLP (WINTER 2017) 25

(Maria, Mary), (no, did not), (slap, daba una bofetada), (a la, the), (bruja, witch), (verde, green) (Maria no, Mary did not), (no daba una bofetada, did not slap), (daba una bofetada a la, slap the), (bruja verde, green witch) (Maria no daba una bofetada, Mary did not slap), (no daba una bofetada a la, did not slap the), (a la bruja verde, the green witch) (Maria no daba una bofetada a la, Mary did not slap the), (daba una bofetada a la bruja verde, slap the green witch)

slide-26
SLIDE 26

Phrase Translation Scores

CS 295: STATISTICAL NLP (WINTER 2017) 26

slide-27
SLIDE 27

Phrases for a Sentence

CS 295: STATISTICAL NLP (WINTER 2017) 27

wir müssen auch diese kritik ernst nehmen

(wir müssen, we must) (wir müssen auch, we must also) (ernst, seriously) ….

slide-28
SLIDE 28

Derivations for a Sentence

CS 295: STATISTICAL NLP (WINTER 2017) 28

slide-29
SLIDE 29

Distortion Limits

CS 295: STATISTICAL NLP (WINTER 2017) 29

slide-30
SLIDE 30

Distortion Scores

CS 295: STATISTICAL NLP (WINTER 2017) 30

slide-31
SLIDE 31

Scoring Derivations

CS 295: STATISTICAL NLP (WINTER 2017) 31

slide-32
SLIDE 32

The Translation Problem

CS 295: STATISTICAL NLP (WINTER 2017) 32

slide-33
SLIDE 33

A Secret of Statistical MT

CS 295: STATISTICAL NLP (WINTER 2017) 33

slide-34
SLIDE 34

Outline

EM-Algorithm for Alignments Phrase-Based MT Decoding Algorithms Syntax-Based MT

CS 295: STATISTICAL NLP (WINTER 2017) 34

slide-35
SLIDE 35

The Decoding Task

CS 295: STATISTICAL NLP (WINTER 2017) 35

slide-36
SLIDE 36

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 36

slide-37
SLIDE 37

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 37

slide-38
SLIDE 38

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 38

Mary

slide-39
SLIDE 39

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 39

Mary did not

slide-40
SLIDE 40

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 40

Mary did not give

slide-41
SLIDE 41

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 41

Mary did not give a

slide-42
SLIDE 42

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 42

Mary did not give a slap

slide-43
SLIDE 43

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 43

Mary did not give a slap to

slide-44
SLIDE 44

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 44

Mary did not give a slap to the

slide-45
SLIDE 45

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 45

Mary did not give a slap to the witch

slide-46
SLIDE 46

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 46

Mary did not give a slap to the witch green

slide-47
SLIDE 47

Monotonic Word Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 47

slide-48
SLIDE 48

Phrase Decoding: Stacks

CS 295: STATISTICAL NLP (WINTER 2017) 48

slide-49
SLIDE 49

Phrase Decoding: Stacks

CS 295: STATISTICAL NLP (WINTER 2017) 49

slide-50
SLIDE 50

Monotonic Phrase Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 50

slide-51
SLIDE 51

Monotonic Phrase Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 51

slide-52
SLIDE 52

Monotonic Phrase Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 52

(Mary)

slide-53
SLIDE 53

Monotonic Phrase Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 53

(Mary) (did not)

slide-54
SLIDE 54

Monotonic Phrase Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 54

(Mary) (did not) (slap)

slide-55
SLIDE 55

Monotonic Phrase Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 55

(Mary) (did not) (slap) (the)

slide-56
SLIDE 56

Monotonic Phrase Decoding

CS 295: STATISTICAL NLP (WINTER 2017) 56

(Mary) (did not) (slap) (the) (green witch)