EMNLP2014読み会 "Efficient Non-parametric Estimation of Multiple Embeddings per Word in Vector Space"
Upcoming SlideShare
Loading in...5
×

Like this? Share it with your network

Share

EMNLP2014読み会 "Efficient Non-parametric Estimation of Multiple Embeddings per Word in Vector Space"

  • 371 views
Uploaded on

EMNLP2014の Efficient Non-parametric Estimation of Multiple Embeddings per Word in Vector Space を紹介しました

EMNLP2014の Efficient Non-parametric Estimation of Multiple Embeddings per Word in Vector Space を紹介しました

More in: Technology
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
Post
    Be the first to comment
No Downloads

Views

Total Views
371
On Slideshare
334
From Embeds
37
Number of Embeds
17

Actions

Shares
Downloads
8
Comments
0
Likes
10

Embeds 37

https://twitter.com 12
https://www.facebook.com 10
http://m.facebook.com&_=1416674334818 HTTP 1
http://m.facebook.com&_=1416658864406 HTTP 1
http://m.facebook.com&_=1416650755191 HTTP 1
http://m.facebook.com&_=1416648515037 HTTP 1
http://m.facebook.com&_=1416643803232 HTTP 1
http://m.facebook.com&_=1416643762465 HTTP 1
http://m.facebook.com&_=1416642860364 HTTP 1
http://m.facebook.com&_=1416642087903 HTTP 1
http://m.facebook.com&_=1416641824768 HTTP 1
http://m.facebook.com&_=1416641221534 HTTP 1
http://m.facebook.com&_=1416641161619 HTTP 1
http://m.facebook.com&_=1416641284618 HTTP 1
http://m.facebook.com&_=1416638727348 HTTP 1
http://m.facebook.com&_=1416638667061 HTTP 1
http://m.facebook.com&_=1416706841661 HTTP 1

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. EMNLP2014 @PFI Efficient Non-parametric Estimation of Multiple Embeddings per Word in Vector Space Arvind Neelakantan, Jeevan Shankar, Alexandre Passos, Andrew McCallum ®Preferred Infrastructure š~ôø (@unnonouno)
  • 2. nìûÏ š~ôø (@unnonouno) ! n՞ǽ‹6ODKPYARbF6ÿå ! ´ò½‹5&7'+6:2" ! NLP¾q/fµt©†k2014- ! rc§‹¨Û¹IBM¢ýÊPFI 
  • 3. îÓ NLP¾q/fYANS ! YANSIbXJBZ9o ! ¾qÊÒl40ÈãØ9/¾qÊÒln Ã/ÊÒ>˜4;$4/¸2: ! s6¤jCbJR@/»œ5üç ! ƒd0mê* $
  • 4. ! YANSĀ3o ! žÇ½‹fŒecf’yg.ĀÑf> 2" ! à-;õ3f*"" 
  • 5. “þ ! word2vec>ÙÉ )1(/àÇ.í™/”× WEPa>Á; ! àÇ/—å+ÇÂĄă•¥¿>tx.¥ ! ”×WEPa/™>nŽ*4;w„5«£ ); 
  • 6. ù*=;Skip-gram [Mikolov+13] $(&#%'  éā 0 Ă * ; 
  • 7. ‰{€™*=;Skip-gram ! àÇ/WEPa v(wt) +GbODKP/WEP a v(c) /|Ð>IFAQ.<;  Ă*;
  • 8. Multi Sense Skip-gram (MSSG) model 1. ÞÚàÇ/GbODKPWEPa9›æG bODKP>±h 2. ”×âÀ/g9iËGbODKP/ßú $ ”×>ˆó 3. #/”×.v¼";”×WEPa>ˆó 4. ˆó<$”×WEPa.Ý')ÞÚGbOD KPz );+.";Skip-gram+ t! 
  • 9. ù*=;MSSG Skip-gram  $(&#%'"!   "
  • 10. 
  • 11. ‰{€™*=;MSSG ! ˆó $”×WEPa>¯% ! àÇ/WEPa0”×ï.ª:GbODKP /WEPa01(%ª; 
  • 12. @aH`LZ*=;MSSG ! ÞÚGbO DKP> ¯')”× >ˆó"; G Qu ;% 
  • 13. Non-Parametric MSSG (NP-MSSG) model ! MSSG*0”×WEPa/™04¬&%'$ ! NP-MSSG*0<>nŽ*4; ! 4w0àö.ƒ2*+GbODKPWEP a/›æƶ$9} ”×WEPa> ͜";% 
  • 14. NP-MSSG.;”×/ˆó/Ôw ! ƒ2*+ÆGbODKPh— $9}9 ”×Ëä>²:‘)8 ! k(wt)0ƒ2*wt.²:‘)$”×/™ ! vcontext0wt/›æGbODKPWEPa ! (wt, k)0kˉ/”×/GbODKPg¦ 
  • 15. q„/2+4 ! MSSG ! àÇ.v )kÅ/”×WEPa>²:‘);S_ [ Mkè !  
  • 16. "  ! ! ˆó<$”×WEPa>¯')ñ‚.Skip-gram ! NP-MSSG ! kÅ/”×WEPa>²:‘);/0t! !  
  • 17.  "#    ! ˆó<$”×WEPa*Skip-gram";/0t! 
  • 18. ŠÌ¡°Apple ! ° +J^VK<.p<; 
  • 19. ŠÌ¡°Run) ! ”×/Ąă-àÇ.0º/”×WEPa³Á<; 
  • 20. ”×/ßú…>­—*;/ŠÌ localSimiËú$”×t¹/÷ÖcŸÜ 
  • 21. ·Î ! Š01+?,t!ŠÌ>6')$(^o^) ! E_K™>nŽ*z";w„.áð )$ /*?-/*/ +ëÄ ! ÞÚàÇ*4;8:QD][bPŒ‡/P TNE*4$1nÕ-–"; ! S_F_UWEPa>¯