Difference between revisions of "English and Nivkh"

From LING073
Jump to: navigation, search
 
(27 intermediate revisions by 3 users not shown)
Line 1: Line 1:
−
Resources for machine translation between [[Nivkh]] and [[English]]
+
Resources for machine translation between [[English]] and [[Nivkh]]
  
 
==External resources==
 
==External resources==
 
*[https://github.swarthmore.edu/Ling073-sp19/ling073-niv-eng language pair repo]
 
*[https://github.swarthmore.edu/Ling073-sp19/ling073-niv-eng language pair repo]
−
*[https://github.swarthmore.edu/Ling073-sp19/ling073-niv niv transducer repo]
+
*[https://github.swarthmore.edu/Ling073-sp19/ling073-niv niv and eng transducer repos]
−
*git-eng transducer repo
 
  
 
==Developed resources==
 
==Developed resources==
 
*[https://github.swarthmore.edu/Ling073-sp19/ling073-niv-eng-corpus parallel corpus]
 
*[https://github.swarthmore.edu/Ling073-sp19/ling073-niv-eng-corpus parallel corpus]
 +
*[[English and Nivkh/Lexical_selection|lexical selection]]
 +
*[[English and Nivkh/Contrastive Grammar|contrastive grammar]]
 +
*[[English and Nivkh/Structural transfer|structural transfer]]
 +
 +
== Initial Evaluation ==
 +
===niv → eng evaluation===
 +
Coverage using monolingual transducer:
 +
<pre>
 +
Number of tokenised words in the corpus: 198
 +
Coverage: 69.7%
 +
</pre>
 +
 +
Coverage using bilingual transducer:
 +
<pre>
 +
Number of tokenised words in the corpus: 198
 +
Coverage: 68.69%
 +
</pre>
 +
 +
{| class="wikitable"
 +
|-
 +
! Sentence
 +
! Translation
 +
! lexical transfer
 +
! full translation
 +
|-
 +
| иф иғдь.
 +
| he killed him.
 +
| ^иф<prn><pers><p3><sg>/he<prn><p3><mf><sg>$ ^иғ<v><tv><ind>/kill<vblex><ind>$^.<sent>/.<sent>$
 +
| #he #kill.
 +
|-
 +
| иф иғныдь.
 +
| he will kill him.
 +
| ^иф<prn><pers><p3><sg>/he<prn><p3><mf><sg>$ ^иғ<v><tv><fut><ind>/kill<vblex><fut><ind>$^.<sent>/.<sent>$
 +
| #he #kill.
 +
|-
 +
| ӿоӻор озр видь.
 +
| then, standing up, he went away.
 +
| ^ӿоӻор<adv>/then<adv>$ ^оз<v><iv><cvb_nar><p3><sg>/stand<vblex><cvb_nar><p3><sg>$ ^ви<v><iv><ind>/go<vblex><ind>$^.<sent>/.<sent>$
 +
| then #stand #go.
 +
|-
 +
| умгу ооламотть.
 +
| the woman kissed the child.
 +
| ^умгу<n>/woman<n>$ ^оола<n>/child<n>$ ^мот<v><tv><ind>/kiss<vblex><ind>$^.<sent>/.<sent>$
 +
| #woman #child #kiss.
 +
|-
 +
| умгу леп пʼоолаардь.
 +
| the woman fed her child bread.
 +
| ^умгу<n>/woman<n>$ ^леп<n>/bread<n>$ ^пʼи<det><ref.><poss>/prpers<det><pos><p3><mf><pl>$ ^оола<n>/child<n>$ ^ар<v><tv><ind>/feed<vblex><ind>$^.<sent>/.<sent>$
 +
| #woman #bread #self #child #feed.
 +
|-
 +
| нивхӄʼатумгу пудь.
 +
| the woman who scolded the man went out.
 +
| ^нивх<n>/person<n>$ ^ӽат<v><tv>/scold<vblex>$ ^умгу<n>/woman<n>$ ^пʼу<v><iv><ind>/go out<vblex><ind>$^.<sent>/.<sent>$
 +
| #person #scold #woman #go out.
 +
|-
 +
| умгу ӿынивхӄʼатть.
 +
| the woman scolded that man.
 +
| ^умгу<n>/woman<n>$ ^ӿы<det><dem>/that<det><dem>$ ^нивх<n>/person<n>$ ^ӽат<v><tv><ind>/scold<vblex><ind>$^.<sent>/.<sent>$
 +
| #woman #this #person #scold.
 +
|-
 +
| ытык пʼызнивхордь.
 +
| father met the man who had called him.
 +
| ^*ытык/*ытык$ ^пʼи<prn><ref.>/self<n>$ ^ыз<v><tv>/call<vblex>$ ^нивх<n>/person<n>$ ^ор<v><tv><ind>/meet<vblex><ind>$^.<sent>/.<sent>$
 +
| #father #self #call #person #meet.
 +
|-
 +
| оола пʼӈафӄзадь.
 +
| the child beat his friend.
 +
| ^оола<n>/child<n>$ ^пʼи<det><ref.><poss>/prpers<det><pos><p3><mf><pl>$ ^ӈафӄ<n>/friend<n>$ ^за<v><tv><ind>/beat<vblex><ind>$^.<sent>/.<sent>$
 +
| #child #self #friend #beat.
 +
|-
 +
| ытык оолаамӽтадь.
 +
| father praised the child.
 +
| ^ытык<n>/father<n>$ ^оола<n>/child<n>$ ^амӽта<v><tv><ind>/praise<vblex><ind>$^.<sent>/.<sent>$
 +
| #father #child #praise.
 +
|}
 +
 +
== Final Evaluation ==
 +
 +
=== Additions ===
 +
 +
*Transducer and machine translation: added 101 stems to transducer and 100 to bilingual dictionary
 +
*Structural transfer: 2 rules (pronouns and word order)
 +
*Twol: 4 rules
 +
**Set /B/ to <п> after fricatives
 +
***тыфпарк/тыф<n>+варк<part>
 +
***тыфварк/*тыфварк
 +
**Set /B/ to <б> after sonorants
 +
***^ӄанбарк/ӄан<n>+варк<part>$
 +
***канварк/*канварк$
 +
**Set /Q/ to <ӻ> after voiced stops
 +
***^ыӈгӻана/ыӈг<v><tv><pot>$
 +
**Set /Q/ to <ӽ> after voiceless stops
 +
***^ӽатӽана/ӽат<v><tv><pot>$
 +
 +
=== Monolingual Transducer ===
 +
==== Precision and Recall against annotated corpus ====
 +
Totals: 51 forms, 68 tp, 0 fp, 0 tn, 24 fn
 +
 +
Precision: 100.00000%
 +
 +
Recall: 73.91304%
 +
 +
==== Coverage over large corpus ====
 +
Number of tokenised words in the corpus: 15696
 +
 +
Coverage: 33.50%
 +
 +
Number of words in large corpus: 76713
 +
 +
Number of stems in the transducer: Unique entries: 297
 +
 +
=== MT Pair ===
 +
 +
==== WER and PER over longer corpus ====
 +
Word error rate (WER): 80.46 %
 +
 +
Position-independent word error rate (PER): 75.86 %
 +
 +
Percentage of unknown words: 26.92 %
 +
 +
==== Number of tokens ====
 +
===== Longer =====
 +
Number of words in reference: 348
 +
 +
Number of words in test: 208
 +
 +
Number of tokenised words in the corpus: 233
 +
 +
===== Large =====
 +
Number of tokenised words in the corpus: 15109
 +
 +
==== Trimmed Coverage ====
 +
===== Longer =====
 +
Coverage: 75.97%
 +
 +
===== Large =====
 +
Coverage: 30.28%
  
−
==niv → eng evaluation==
 
−
This is this section.
 
  
 
[[Category:Nivkh]]
 
[[Category:Nivkh]]
 
[[Category:English]]
 
[[Category:English]]
 
[[Category:Sp19_TranslationPairs]]
 
[[Category:Sp19_TranslationPairs]]

Latest revision as of 02:21, 24 April 2019

Resources for machine translation between English and Nivkh

External resources

Developed resources

Initial Evaluation

niv → eng evaluation

Coverage using monolingual transducer: 

Number of tokenised words in the corpus: 198
Coverage: 69.7%

Coverage using bilingual transducer: 

Number of tokenised words in the corpus: 198
Coverage: 68.69%
Sentence Translation lexical transfer full translation
иф иғдь. he killed him. ^иф<prn><pers><p3><sg>/he<prn><p3><mf><sg>$ ^иғ<v><tv><ind>/kill<vblex><ind>$^.<sent>/.<sent>$ #he #kill.
иф иғныдь. he will kill him. ^иф<prn><pers><p3><sg>/he<prn><p3><mf><sg>$ ^иғ<v><tv><fut><ind>/kill<vblex><fut><ind>$^.<sent>/.<sent>$ #he #kill.
ӿоӻор озр видь. then, standing up, he went away. ^ӿоӻор<adv>/then<adv>$ ^оз<v><iv><cvb_nar><p3><sg>/stand<vblex><cvb_nar><p3><sg>$ ^ви<v><iv><ind>/go<vblex><ind>$^.<sent>/.<sent>$ then #stand #go.
умгу ооламотть. the woman kissed the child. ^умгу<n>/woman<n>$ ^оола<n>/child<n>$ ^мот<v><tv><ind>/kiss<vblex><ind>$^.<sent>/.<sent>$ #woman #child #kiss.
умгу леп пʼоолаардь. the woman fed her child bread. ^умгу<n>/woman<n>$ ^леп<n>/bread<n>$ ^пʼи<det><ref.><poss>/prpers<det><pos><p3><mf><pl>$ ^оола<n>/child<n>$ ^ар<v><tv><ind>/feed<vblex><ind>$^.<sent>/.<sent>$ #woman #bread #self #child #feed.
нивхӄʼатумгу пудь. the woman who scolded the man went out. ^нивх<n>/person<n>$ ^ӽат<v><tv>/scold<vblex>$ ^умгу<n>/woman<n>$ ^пʼу<v><iv><ind>/go out<vblex><ind>$^.<sent>/.<sent>$ #person #scold #woman #go out.
умгу ӿынивхӄʼатть. the woman scolded that man. ^умгу<n>/woman<n>$ ^ӿы<det><dem>/that<det><dem>$ ^нивх<n>/person<n>$ ^ӽат<v><tv><ind>/scold<vblex><ind>$^.<sent>/.<sent>$ #woman #this #person #scold.
ытык пʼызнивхордь. father met the man who had called him. ^*ытык/*ытык$ ^пʼи<prn><ref.>/self<n>$ ^ыз<v><tv>/call<vblex>$ ^нивх<n>/person<n>$ ^ор<v><tv><ind>/meet<vblex><ind>$^.<sent>/.<sent>$ #father #self #call #person #meet.
оола пʼӈафӄзадь. the child beat his friend. ^оола<n>/child<n>$ ^пʼи<det><ref.><poss>/prpers<det><pos><p3><mf><pl>$ ^ӈафӄ<n>/friend<n>$ ^за<v><tv><ind>/beat<vblex><ind>$^.<sent>/.<sent>$ #child #self #friend #beat.
ытык оолаамӽтадь. father praised the child. ^ытык<n>/father<n>$ ^оола<n>/child<n>$ ^амӽта<v><tv><ind>/praise<vblex><ind>$^.<sent>/.<sent>$ #father #child #praise.

Final Evaluation

Additions

  • Transducer and machine translation: added 101 stems to transducer and 100 to bilingual dictionary
  • Structural transfer: 2 rules (pronouns and word order)
  • Twol: 4 rules
    • Set /B/ to <п> after fricatives
      • тыфпарк/тыф<n>+варк<part>
      • тыфварк/*тыфварк
    • Set /B/ to <б> after sonorants
      • ^ӄанбарк/ӄан<n>+варк<part>$
      • канварк/*канварк$
    • Set /Q/ to <ӻ> after voiced stops
      • ^ыӈгӻана/ыӈг<v><tv><pot>$
    • Set /Q/ to <ӽ> after voiceless stops
      • ^ӽатӽана/ӽат<v><tv><pot>$

Monolingual Transducer

Precision and Recall against annotated corpus

Totals: 51 forms, 68 tp, 0 fp, 0 tn, 24 fn

Precision: 100.00000%

Recall: 73.91304%

Coverage over large corpus

Number of tokenised words in the corpus: 15696

Coverage: 33.50%

Number of words in large corpus: 76713

Number of stems in the transducer: Unique entries: 297

MT Pair

WER and PER over longer corpus

Word error rate (WER): 80.46 %

Position-independent word error rate (PER): 75.86 %

Percentage of unknown words: 26.92 %

Number of tokens

Longer

Number of words in reference: 348

Number of words in test: 208

Number of tokenised words in the corpus: 233

Large

Number of tokenised words in the corpus: 15109

Trimmed Coverage

Longer

Coverage: 75.97%

Large

Coverage: 30.28%

Retrieved from "http://wikis.swarthmore.edu/ling073/index.php?title=English_and_Nivkh&oldid=10490"