Difference between revisions of "Munsee and English"

From LING073
Jump to: navigation, search
(→‎umu → eng Evaluation)
(→‎Additions)
 
(48 intermediate revisions by 3 users not shown)
Line 8: Line 8:
 
== umu → eng Evaluation ==
 
== umu → eng Evaluation ==
  
−
*Coverage (number of forms analyzed) of Monolingual Transducer:
+
*Coverage (number of forms analyzed) of Monolingual Transducer: '''849/1499 (56.64%)'''
−
*Coverage of Bilingual Transducer:  
+
*Coverage of Bilingual Transducer: '''609/1296 (46.99%)'''
 
*Sentences:
 
*Sentences:
−
#Laawate weemĭ aweeniik moxa mataawsuwak.
+
#'''Laawate weemĭ aweeniik moxa mataawsuwak.'''
−
**English Translation:
+
##English Translation: "Long ago the people of the world became very wicked." 
−
**Output of <code>xyz-abc-biltrans</code> (Lexical Transfer) Mode:  
+
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^Laawate<adv>/Long ago<adv>$ ^weemĭ<det>/all<det><qnt><pl>$ ^aween<prn><pl>/person<prn><pl>$ ^moxa<adv>/very<adv>$ ^mataawsu<v><iv><aa><s_3pl>/sin<vblex><inf><s_3pl>$^.<sent>/.<sent>$
−
**Output of <code>xyz-abc</code> (Full Translation) Mode:   
+
##Output of <code>umu-eng</code> (Full Translation) Mode: #Long ago all #person very #sin
−
#Shiinzuw Noah.
+
#'''Shiinzuw Noah.'''
−
#Nal Pahtamawaas kwuphamun kpahoon.
+
##English Translation: "His name was Noah."  
−
#Kumootkeewak, waak akihkakulooneewak.
+
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^Shiinzuw<v><iv><aa><s_3sg>/To be named<v><iv><aa><s_3sg>$ ^Noah<np><ant>/Noah<np><ant><m><sg>$^.<sent>/.<sent>$
−
#Pahtamawaas wulaawsuw.
+
##Output of <code>umu-eng</code> (Full Translation) Mode: #To be named Noah
−
#Shukw Pahtamawaas weemĭ ktahwaalkoona.
+
#'''Nal Pahtamawaas kwuphamun kpahoon.'''
−
#Shiinzuw Jesus Christ.
+
##English Translation: "Then God shut the door."
−
#Shaaw ang oxkweesus manoongsuw.  
+
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^Nal<adv>/Then<adv>$ ^Pahtamawaas<np><ant>/God<np><al>$ ^*kwuphamun/*kwuphamun$ ^kpahoon<n><nn>/door<n>$^.<sent>/.<sent>$
−
#Ngáleew pambíilak.
+
##Output of <code>umu-eng</code> (Full Translation) Mode: Then #God *kwuphamun #door
−
#Tumúsheew awéeniil.
+
#'''Kumootkeewak, waak akihkakulooneewak.'''
−
 
+
##English Translation: "They steal and lie."
−
 
+
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^*Kumootkeewak/*Kumootkeewak$^,<cm>/,<cm>$ ^waak<cnjcoo>/and<cnjcoo>$ ^*akihkakulooneewak/*akihkakulooneewak$^.<sent>/.<sent>$^.<sent>/.<sent>$
−
 
+
##Output of <code>umu-eng</code> (Full Translation) Mode: *Kumootkeewak, and *akihkakulooneewak.
 +
#'''Pahtamawaas wulaawsuw.'''
 +
##English Translation: "God is holy."
 +
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^Pahtamawaas<np><ant>/God<np><al>$ ^wulaawsuw<v><iv><aa><s_3sg>/to be holy<v><iv><aa><s_3sg>$^.<sent>/.<sent>$
 +
##Output of <code>umu-eng</code> (Full Translation) Mode: #God #to be holy
 +
#'''Shukw Pahtamawaas weemĭ ktahwaalkoona.'''
 +
##English Translation: "But God loves us all."
 +
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^Shukw<cnjcoo>/But<cnjcoo>$ ^Pahtamawaas<np><ant><sg>/God<np><al><sg>$ ^weemĭ<det>/all<det><qnt><pl>$ ^*ktahwaalkoona/*ktahwaalkoona$^.<sent>/.<sent>$
 +
##Output of <code>umu-eng</code> (Full Translation) Mode: But God all *ktahwaalkoona
 +
#'''Shiinzuw Jesus Christ.'''
 +
##English Translation: "His name is Jesus Christ."
 +
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^Shiinzuw<v><iv><aa><s_3sg>/To be named<v><iv><aa><s_3sg>$ ^Jesus<np><ant><sg>/Jesus<np><ant><m><sg><sg>$ ^Christ<np><ant><sg>/Christ<np><al><sg><sg>$^.<sent>/.<sent>$
 +
##Output of <code>umu-eng</code> (Full Translation) Mode: #To be named #Jesus #Christ
 +
#'''Shaaw ang oxkweesus manoongsuw.'''
 +
##English Translation: "Right away the girl would get angry."
 +
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^*Shaaw/*Shaaw$ ^ang<emph>/emphatic<adj>$ ^oxkwéesus<n><aa>/girl<n><sg>$ ^manoongsuw<v><iv><aa><s_3sg>/be angry<v><iv><aa><s_3sg>$^.<sent>/.<sent>$
 +
##Output of <code>umu-eng</code> (Full Translation) Mode: *Shaaw emphatic girl #be angry
 +
#'''Ngáleew pambíilak.'''
 +
##English Translation: "He left behind some books."
 +
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^*Ngáleew/*Ngáleew$ ^*pambíilak/*pambíilak$^.<sent>/.<sent>$^.<sent>/.<sent>$
 +
##Output of <code>umu-eng</code> (Full Translation) Mode: *Ngáleew *pambíilak.
 +
#'''Tumúsheew awéeniil.'''
 +
##English Translation: "He cut someone."
 +
##Output of <code>umu-eng-biltrans</code> (Lexical Transfer) Mode: ^*Tumúsheew/*Tumúsheew$ ^aween<prn><prn><pl><obv>/person<prn><prn><pl><obv>$^.<sent>/.<sent>$
 +
##Output of <code>umu-eng</code> (Full Translation) Mode: *Tumúsheew #person
  
 +
== Additions ==
  
 +
* We added a file umu.corpus.large.txt, which contains over 10K characters. Unfortunately, due to a lack of available resources, we are unable to reach 100K words.
 +
* '''We added 100 stems to our bilingual dictionary'''
 +
* '''We expanded our morphology (lexd) in 5 ways'''
 +
** Imperatives
 +
*** (1) VTA Imperatives (for both third and first person commands)
 +
*** (2) VAI Imperatives (for stems ending in vowels and consonants)
 +
** Negative Independent Indicative Mode for three of the verb classes
 +
*** (3) VII
 +
*** (4) VAI
 +
*** (5) VTA
 +
*** We have not implemented negatives for VTIs as that involves more complex rules regarding both the plurality of the objects and the ending of the stems.
 +
* '''We added 4 more twol rules'''
 +
** (1) Accent shifts for plurals - placed on the new penultimate vowel after a suffix is added
 +
** (2) Special case of VAI imperatives when stem ends in -xiin
 +
** (3) Realizing -oo for verb stems with consonant endings
 +
** (4) Realizing shaa as shaaw when the subsequent word begins with an "a"
  
 
[[Category:Sp25_TranslationPairs]]
 
[[Category:Sp25_TranslationPairs]]
 
[[Category:Munsee]]
 
[[Category:Munsee]]
 
[[Category:English]]
 
[[Category:English]]

Latest revision as of 14:52, 6 May 2025

Resources for machine translation between Munsee and English

External Resources

umu → eng Evaluation

  • Coverage (number of forms analyzed) of Monolingual Transducer: 849/1499 (56.64%)
  • Coverage of Bilingual Transducer: 609/1296 (46.99%)
  • Sentences:
  1. Laawate weemĭ aweeniik moxa mataawsuwak.
    1. English Translation: "Long ago the people of the world became very wicked."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^Laawate<adv>/Long ago<adv>$ ^weemĭ<det>/all<det><qnt><pl>$ ^aween<prn><pl>/person<prn><pl>$ ^moxa<adv>/very<adv>$ ^mataawsu<v><iv><aa><s_3pl>/sin<vblex><inf><s_3pl>$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: #Long ago all #person very #sin
  2. Shiinzuw Noah.
    1. English Translation: "His name was Noah."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^Shiinzuw<v><iv><aa><s_3sg>/To be named<v><iv><aa><s_3sg>$ ^Noah<np><ant>/Noah<np><ant><m><sg>$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: #To be named Noah
  3. Nal Pahtamawaas kwuphamun kpahoon.
    1. English Translation: "Then God shut the door."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^Nal<adv>/Then<adv>$ ^Pahtamawaas<np><ant>/God<np><al>$ ^*kwuphamun/*kwuphamun$ ^kpahoon<n><nn>/door<n>$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: Then #God *kwuphamun #door
  4. Kumootkeewak, waak akihkakulooneewak.
    1. English Translation: "They steal and lie."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^*Kumootkeewak/*Kumootkeewak$^,<cm>/,<cm>$ ^waak<cnjcoo>/and<cnjcoo>$ ^*akihkakulooneewak/*akihkakulooneewak$^.<sent>/.<sent>$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: *Kumootkeewak, and *akihkakulooneewak.
  5. Pahtamawaas wulaawsuw.
    1. English Translation: "God is holy."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^Pahtamawaas<np><ant>/God<np><al>$ ^wulaawsuw<v><iv><aa><s_3sg>/to be holy<v><iv><aa><s_3sg>$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: #God #to be holy
  6. Shukw Pahtamawaas weemĭ ktahwaalkoona.
    1. English Translation: "But God loves us all."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^Shukw<cnjcoo>/But<cnjcoo>$ ^Pahtamawaas<np><ant><sg>/God<np><al><sg>$ ^weemĭ<det>/all<det><qnt><pl>$ ^*ktahwaalkoona/*ktahwaalkoona$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: But God all *ktahwaalkoona
  7. Shiinzuw Jesus Christ.
    1. English Translation: "His name is Jesus Christ."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^Shiinzuw<v><iv><aa><s_3sg>/To be named<v><iv><aa><s_3sg>$ ^Jesus<np><ant><sg>/Jesus<np><ant><m><sg><sg>$ ^Christ<np><ant><sg>/Christ<np><al><sg><sg>$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: #To be named #Jesus #Christ
  8. Shaaw ang oxkweesus manoongsuw.
    1. English Translation: "Right away the girl would get angry."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^*Shaaw/*Shaaw$ ^ang<emph>/emphatic<adj>$ ^oxkwéesus<n><aa>/girl<n><sg>$ ^manoongsuw<v><iv><aa><s_3sg>/be angry<v><iv><aa><s_3sg>$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: *Shaaw emphatic girl #be angry
  9. Ngáleew pambíilak.
    1. English Translation: "He left behind some books."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^*Ngáleew/*Ngáleew$ ^*pambíilak/*pambíilak$^.<sent>/.<sent>$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: *Ngáleew *pambíilak.
  10. Tumúsheew awéeniil.
    1. English Translation: "He cut someone."
    2. Output of umu-eng-biltrans (Lexical Transfer) Mode: ^*Tumúsheew/*Tumúsheew$ ^aween<prn><prn><pl><obv>/person<prn><prn><pl><obv>$^.<sent>/.<sent>$
    3. Output of umu-eng (Full Translation) Mode: *Tumúsheew #person

Additions

  • We added a file umu.corpus.large.txt, which contains over 10K characters. Unfortunately, due to a lack of available resources, we are unable to reach 100K words.
  • We added 100 stems to our bilingual dictionary
  • We expanded our morphology (lexd) in 5 ways
    • Imperatives
      • (1) VTA Imperatives (for both third and first person commands)
      • (2) VAI Imperatives (for stems ending in vowels and consonants)
    • Negative Independent Indicative Mode for three of the verb classes
      • (3) VII
      • (4) VAI
      • (5) VTA
      • We have not implemented negatives for VTIs as that involves more complex rules regarding both the plurality of the objects and the ending of the stems.
  • We added 4 more twol rules
    • (1) Accent shifts for plurals - placed on the new penultimate vowel after a suffix is added
    • (2) Special case of VAI imperatives when stem ends in -xiin
    • (3) Realizing -oo for verb stems with consonant endings
    • (4) Realizing shaa as shaaw when the subsequent word begins with an "a"