人工知能と機械学習 / AI and ML

Transcript

1. ஜ೾େֶ৘ใՊֶྨഅ৔ઇ೫
   CBCB!DTUTVLVCBBDKQ
   ஜ೾େֶڞ௨Պ໨ʢ৘ใʣ

   ʮσʔλαΠΤϯεʯ

   ਓ޻஌ೳͱػցֶश
   

   View Slide

2. /39
   ຊߨٛͰֶͿ͜ͱ
   w ਓ޻஌ೳ͕Ͳ͏͍͏৔໘Ͱ׆༻͞Ε͍ͯΔͷ͔
   w ͍·ͷਓ޻஌ೳʹ͸Կ͕Ͱ͖Δͷ͔
   w ػցֶशͰ͸ͲͷΑ͏ʹσʔλ͔ΒͷֶशΛߦ͏ͷ͔
   ਓ޻஌ೳʹԿ͕Ͱ͖Δͷ͔ɾػցֶश͸Ͳ͏͍͏͘͠Έ͔
   2
   

   View Slide

3. ਓ޻஌ೳ
   

   View Slide

4. /39
   ਓ޻஌ೳͱ͸
   w ਓ޻஌ೳ͸

   ʮ஌తͳػցɺಛʹɺ஌తͳίϯϐϡʔλϓϩάϥϜΛ࡞ΔՊֶͱٕज़ʯ
   w ैདྷͷίϯϐϡʔλϓϩάϥϜ͸஌తͰ͸ͳ͔ͬͨ
   w ਓ͕ؒ஌ೳΛ࢖ͬͯͰ͖Δͷͱಉ͜͡ͱΛ

   ίϯϐϡʔλϓϩάϥϜͰ΋Ͱ͖ΔΑ͏ʹ͍ͨ͠
   ਓ޻஌ೳ͸ʮ஌తͳίϯϐϡʔλϓϩάϥϜʯ
   4
   
   IUUQTXXXBJHBLLBJPSKQXIBUTBJ"*GBRIUNM
   (*)
   

   View Slide

5. /39
   ਓ޻஌ೳͱ͸
   w ஌తͰ͸ͳ͍ίϯϐϡʔλϓϩάϥϜ͸

   ༩͑ΒΕͨϧʔϧͲ͓Γͷಈ࡞͔͠Ͱ͖ͳ͍
   ⿞ίϯϐϡʔλʹѻ͑ΔܗࣜͰਓ͕ؒϧʔϧΛ࡞Δඞཁ͕͋Δ
   ⿞ෳࡶͳ໰୊Ͱ͸͢΂ͯͷঢ়گΛ໢ཏͨ͠ϧʔϧΛ࡞Δ͜ͱ͸ࠔ೉
   w ஌తͳίϯϐϡʔλϓϩάϥϜʢਓ޻஌ೳʣ͸

   ঢ়گʹԠͯ͡ͲͷΑ͏ʹಈ࡞͢Ε͹ྑ͍͔ΛྟػԠมʹ൑அ͢Δ
   ⿞ܦݧ΍ڭࡐ͔ΒϧʔϧΛֶश͢Δ
   ⿞গ਺ͷϧʔϧ͔Βਪ࿦ͯ͠൑அ͢Δ
   ਓ޻஌ೳ͸ঢ়گʹԠͯ͡ྟػԠมʹಈ࡞͢Δ
   5
   

   View Slide

6. /39
   w ஌తͰ͸ͳ͍ίϯϐϡʔλϓϩάϥϜ͸ϧʔϧʹैͬͯೝࣝ͢Δ
   w ༷ʑͳච੻ͷखॻ͖จࣈΛ໢ཏͨ͠ϧʔϧΛ࡞Δͷ͸ࠔ೉

   

   ਓ޻஌ೳͱ͸
   ྫɿखॻ͖਺ࣈจࣈΛೝࣝ͢ΔίϯϐϡʔλϓϩάϥϜ
   6
   खॻ͖਺ࣈจࣈͷग़యɿIUUQZBOOMFDVODPNFYECNOJTU
   1
   1
   ʮࠨӈࡾ෼ͷҰͷྖҬ͕ന৭ͳΒʯ
   

   View Slide

7. /39
   ਓ޻஌ೳ͸ڭࡐʢը૾ͱ਺ࣈͷϖΞʣΛ࢖ͬͯೝࣝϧʔϧΛֶश͢Δ
   ਓ޻஌ೳͱ͸
   ྫɿखॻ͖਺ࣈจࣈΛೝࣝ͢ΔίϯϐϡʔλϓϩάϥϜ
   7
   1
   1
   9
   9
   7
   7
   ϧʔϧΛֶश
   9
   ϧʔϧΛར༻
   ڭࡐ
   ʜ
   खॻ͖਺ࣈจࣈͷग़యɿIUUQZBOOMFDVODPNFYECNOJTU
   

   View Slide

8. /39
   ೔ৗͷதͷਓ޻஌ೳ
   w εϚʔτεϐʔΧʔɿ

   ਓؒͷԻ੠ࢦࣔʹै͍༷ʑͳಈ࡞Λߦ͏
   w ϩϘοτ૟আػɿ

   ো֐෺Λආ͚ͳ͕Β෦԰શମΛ૟আ͢Δ
   w إೝূɿ

   Χϝϥલͷਓ෺͕ొ࿥͞ΕͨਓͱಉҰਓ෺͔Ͳ͏͔Λ൑ఆ
   w ίϯςϯπɾ঎඼ਪનɿ

   Ӿཡཤྺ΍ߪങཤྺʹ΋ͱ͖ͮϢʔβ͕޷ΉΞΠςϜΛఏࣔ
   զʑͷ೔ৗͷ͞·͟·ͳ৔໘Ͱਓ޻஌ೳ͕׆༻͞Ε͍ͯΔ
   8
   

   View Slide

9. /39
   ਓؒΛ௒͑Δਓ޻஌ೳ
   w *#.8BUTPOɿ

   ΞϝϦΧͷΫΠζ൪૊ʮ+FPQBSEZʯʹ௅ઓɺৗ࿈ग़ԋऀΛഁΓ༏উ
   w ౦ϩϘ͘Μɿ

   େֶೖࢼ໛ࢼͷ਺ֶɾੈք࢙Ͱภࠩ஋௒͑Λୡ੒
   w %FFQ#MVFɿ

   νΣεͷੈքνϟϯϐΦϯʹউར
   w "MQIB(Pɿ

   ғޟͷੈքτοϓع࢜ʹউར
   ΫΠζ΍ήʔϜͰਓؒΛ௒͑Δਓ޻஌ೳ͕ొ৔
   9
   

   View Slide

10. /39
    
    ݱࡏͷਓ޻஌ೳʹͰ͖Δ͜ͱɿಛԽܕWT൚༻ܕ
    w ਓ޻஌ೳ͸ʮಛԽܕʯͱʮ൚༻ܕʯͷೋछྨʹେผ͞ΕΔ
    w ಛԽܕਓ޻஌ೳɿ

    ͋Δಛఆͷঢ়گ΍໰୊ʹ͓͍ͯ஌తʹ;Δ·͏ਓ޻஌ೳ
    w ൚༻ܕਓ޻஌ೳɿ

    ਓؒͱಉ༷ʹ༷ʑͳঢ়گɾ໰୊ʹ͓͍ͯ஌తͳ;Δ·͍͕Ͱ͖Δਓ޻஌ೳ
    w ݱࡏͷਓ޻஌ೳ͸ಛԽܕਓ޻஌ೳͰ͋Γ

    ൚༻ܕਓ޻஌ೳ͸ະ࣮ͩݱ͞Ε͍ͯͳ͍

    ݱࡏͷਓ޻஌ೳ͸ಛఆͷ͜ͱ͚ͩͰ͖ΔಛԽܕਓ޻஌ೳ
    10
    

    View Slide

11. /39
    ݱࡏͷਓ޻஌ೳʹͰ͖Δ͜ͱɿը૾ॲཧ
    w ը૾ೝࣝɿ

    ࣸਅʹ͍ࣸͬͯΔ΋ͷΛೝࣝ͢Δ
    w ෺ମݕग़ɿ

    ࣸਅͷͲ͜ʹԿ͕͍ࣸͬͯΔͷ͔Λೝࣝ͢Δ
    w ը૾ੜ੒ɿ

    ຊ෺ͷࣸਅͷΑ͏ͳը૾Λਓ޻తʹ࡞Γग़͢
    ಛԽܕਓ޻஌ೳͰ΋༷ʑͳ͜ͱ͕࣮ݱͰ͖Δ
    11
    
    ग़యɿIUUQDTOTUBOGPSEFEVTMJEFTDTO@@MFDUVSFQEG

    
    ग़యɿIUUQTBSYJWPSHBCT
    ຊ෺ͷࣸਅ ਓ޻తͳࣸਅ
    (*)
    (*)
    

    View Slide

12. /39
    ݱࡏͷਓ޻஌ೳʹͰ͖Δ͜ͱɿݴޠॲཧ
    w จॻ෼ྨɿ

    จॻΛಛఆͷΧςΰϦʹ෼ྨ͢Δ
    w ػց຋༁ɿ

    ͋ΔݴޠͰॻ͔ΕͨจষΛผͷݴޠʹ຋༁͢Δ
    w ৘ใݕࡧɿ

    ಛఆͷ৘ใ͕ܝࡌ͞Ε͍ͯΔ΢ΣϒϖʔδΛݟ͚ͭΔ
    w ࣭໰Ԡ౴ɿ

    จষͰ༩͑ΒΕ࣭ͨ໰ʹର͢Δ౴͑Λฦ͢
    ಛԽܕਓ޻஌ೳͰ΋༷ʑͳ͜ͱ͕࣮ݱͰ͖Δ
    12
    

    View Slide

13. /39
    ݱࡏͷਓ޻஌ೳʹͰ͖Δ͜ͱɿԻ੠ॲཧ
    w Ի੠ೝࣝɿ

    ࿩͞Ε͍ͯΔ಺༰ΛจষͰॻ͖ى͜͢
    w ࿩ऀಉఆɿ

    ొ࿥͞Εͨਓ෺ͱಉ͡ਓ͕࿩͍ͯ͠Δ͔൑ఆ͢Δ
    w Ի੠߹੒ɿ

    ਓ͕ؒ࿩͍ͯ͠ΔΑ͏ͳ੠Λਓ޻తʹ࡞Γग़͢
    ಛԽܕਓ޻஌ೳͰ΋༷ʑͳ͜ͱ͕࣮ݱͰ͖Δ
    13
    

    View Slide

14. /39
    ݱࡏͷਓ޻஌ೳʹͰ͖Δ͜ͱɿϩϘοτ
    w ؀ڥೝࣝɿ

    ηϯα৘ใ౳͔ΒपғʹԿ͕͋Δ͔Λೝࣝ͢Δ
    w ܦ࿏ɾߦಈܭըɿ

    Ҡಈɾಈ࡞ࢦࣔʹै͏ͨΊʹͲͷΑ͏ͳܦ࿏Λ௨Δ͔ɺ

    Ͳͷ෦෼ΛͲͷΑ͏ʹಈ͔͔͢Λܭը͢Δ
    ಛԽܕਓ޻஌ೳͰ΋༷ʑͳ͜ͱ͕࣮ݱͰ͖Δ
    14
    

    View Slide

15. /39
    ਓ޻஌ೳͷ׆༻ྫ
    ྫᶃεϚʔτεϐʔΧʔ
    ಛԽܕਓ޻஌ೳͷ૊߹ͤͰ͞Βʹߴ౓ͳ͜ͱΛ࣮ݱͰ͖Δ
    15
    4UFQىಈΩʔϫʔυͷೝࣝ
    4UFQԻ੠ʹΑΔࢦࣔΛจষͱͯ͠ೝࣝ
    4UFQࢦࣔจ͔Β໨తΛೝࣝ
    4UFQࢦࣔΛ࣮ߦ
    

    View Slide

16. /39
    ਓ޻஌ೳͷ׆༻ྫ
    ྫᶄࣗಈӡస
    ಛԽܕਓ޻஌ೳͷ૊߹ͤͰ͞Βʹߴ౓ͳ͜ͱΛ࣮ݱͰ͖Δ
    16
    ը૾ͷग़యɿIUUQTTUPSBHFHPPHMFBQJTDPNTEDQSPEWTBGFUZSFQPSUXBZNPTBGFUZSFQPSUQEG
    4UFQଞͷं྆΍าߦऀ౳ͷݕ஌
    4UFQଞͷं྆΍าߦऀ౳ͷߦಈ༧ଌ
    4UFQ໨త஍·Ͱͷܦ࿏Λܭը
    4UFQܦ࿏ʹै͏ͨΊͷಈ࡞Λܭը
    

    View Slide

17. /39
    ਓ޻஌ೳͷ·ͱΊ
    w ਓ޻஌ೳ͸஌తͳίϯϐϡʔλϓϩάϥϜͰ

    ϧʔϧΛֶशɾਪ࿦ͯ͠ྟػԠมʹಈ࡞Ͱ͖Δ
    w ͍·ͷਓ޻஌ೳ͸ಛԽܕͰ͋Δಛఆͷ໰୊͔͠ղ͚ͳ͍
    w ༷ʑͳ෼໺ͰಛԽܕਓ޻஌ೳͷݚڀ͕ਐΊΒΕ

    ͦΕΒͷ૊߹ͤͰΑΓߴ౓ͳਓ޻஌ೳ͕࣮ݱ͞Εͭͭ͋Δ
    ͍·ͷਓ޻஌ೳ͸ಛԽܕ͕༷ͩʑͳ͜ͱ͕Ͱ͖Δ
    17
    

    View Slide

18. ػցֶश
    

    View Slide

19. /39
    ػցֶशͱ͸
    w ػցֶश͸

    ίϯϐϡʔλϓϩάϥϜʹσʔλ͔ΒϧʔϧΛֶशͤ͞Δٕज़
    
    w େྔͷೖग़ྗྫͷσʔλΛڭࡐͱͯ͠༩͑

    ೖྗͱग़ྗΛରԠ͚ͮΔϧʔϧΛֶशͤ͞Δ
    ೖྗͱग़ྗΛରԠ͚ͮΔϧʔϧΛֶश͢Δ
    19
    
    ຊߨٛͰऔΓ্͛Δػցֶश͸ɺݫີʹ͸ʮڭࢣ෇͖ػցֶशʯͱݺ͹ΕΔ
    7
    ೖྗ ग़ྗ
    ը૾ ਺ࣈ
    

    View Slide

20. /39
    ػցֶशͱ͸
    ೖྗͱग़ྗΛରԠ͚ͮΔϧʔϧΛֶश͢Δ
    20
    The quick brown
    fox jumps over
    the lazy dog
    ૉૣ͍஡৭ͷޅ͸
    ͷΖ·ͳݘΛඈͼ
    ӽ͑Δ
    “Hello, world!”
    ೖྗ ग़ྗ
    ݈߁ϦεΫɿߴ
    
    ग़యɿIUUQDTOTUBOGPSEFEVTMJEFTDTO@@MFDUVSFQEG

    
    ग़యɿIUUQTNBHFOUBUFOTPSqPXPSHOTZOUIGBTUHFO
    จষ จষ
    ը૾ ෺ମͱҐஔ
    Ի੠ จষ
    ױऀ৘ใ ݈߁ϦεΫ
    (*)
    (**)
    

    View Slide

21. /39
    ػցֶशͱ͸
    w εϚʔτϑΥϯɾηϯαɾΠϯλʔωοτɾ΢ΣϒαʔϏε౳ͷීٴͰ

    σʔλ͕૿େͨ͠ͷ͕ػցֶशͷൃలͷ͖͔͚ͬ
    w ιʔγϟϧϝσΟΞ΁ͷ౤ߘ΍

    ΢ΣϒαʔϏε্Ͱͷਓʑͷߦಈϩά΋ػցֶशͷڭࡐͱͯ͠༻͍ΒΕΔ
    w ޿ࠂ഑৴ͳͲͰ΢ΣϒαʔϏεӡӦاۀ͸རӹΛ্͛ɺ

    ແྉͰ΢ΣϒαʔϏεΛఏڙ͠ͳ͕Βڭࡐ༻ͷσʔλΛऩू͍ͯ͠Δ
    ͋ΒΏΔ΋ͷ͕σʔλԽ͞ΕػցֶशͰ࢖ΘΕΔ
    21
    

    View Slide

22. /39
    ػցֶशͷ͘͠Έɿ܇࿅σʔλ
    w ྫ୊ɿʮजᙾ͕ྑੑ͔ѱੑ͔ʯ൑ఆ͢ΔͨΊͷϧʔϧΛֶशͤ͞Δ
    w ڭࡐͱͯ͠༻͍ΔσʔλΛ܇࿅σʔλͱݺͿ
    w ֶशͨ͠ϧʔϧΛ༻͍ͯ

    ܇࿅σʔλʹͳ͍जᙾ͕ྑੑ͔ѱੑ͔Λਖ਼͘͠൑ఆͰ͖ΔΑ͏ʹ͍ͨ͠

    ೖग़ྗྫΛ܇࿅σʔλͱͯ͠༻͍ͯϧʔϧΛֶश͢Δ
    22
    जᙾը૾ͷग़యɿW. N. Street et al. , Nuclear feature extraction for breast tumor diagnosis. Proc. of the International Symposium on Electronic
    Imaging, 1993.
    ྑੑ
    ѱੑ
    ྑੑ
    ѱੑ
    ܇࿅σʔλ ೖྗʢजᙾʣΛ

    ग़ྗʢྑੑPSѱੑʣʹ

    ରԠ͚ͮΔ

    ϧʔϧ
    ֶश
    

    View Slide

23. /39
    ػցֶशͷ͘͠Έɿ܇࿅σʔλ
    w ίϯϐϡʔλͰܭࢉͰ͖ΔΑ͏ʹ܇࿅σʔλ͸਺஋Ͱදݱ͞Ε͍ͯΔ
    w ྫɿʮେ͖͞ʯͱʮͰ͜΅͜౓ʯͰजᙾΛදݱ͢Δ

    ʢԿΒ͔ͷखஈͰ਺஋Խ͞Ε͍ͯΔͱ͢Δʣ
    ܇࿅σʔλ͸਺஋Ͱදݱ͞Εͳ͍ͱ͍͚ͳ͍
    23
    ܇࿅σʔλͷྫ
    ೖྗ
    ग़ྗ
    େ͖͞ Ͱ͜΅͜౓
    जᙾ 0.252 0.014 ྑੑ
    जᙾ 0.327 0.340 ྑੑ
    जᙾ 1.000 0.371 ѱੑ
    जᙾ 0.223 0.369 ѱੑ
    ʜ ʜ ʜ ʜ
    

    View Slide

24. /39
    ػցֶशͷ͘͠Έɿ܇࿅σʔλ
    ܇࿅σʔλ͸਺஋Ͱදݱ͞Εͳ͍ͱ͍͚ͳ͍
    24
    211 220 223 … 213
    217 217 220 … 210
    214 217 216 … 205
    … … … … …
    216 217 214 … 181
    The quick brown fox jumps
    over the lazy dog
    aardvark … brown … dog … zzzat
    0 … 1 … 1 … 0
    

    View Slide

25. /39
    ػցֶशͷ͘͠Έɿ܇࿅σʔλ
    ܇࿅σʔλ͸਺஋Ͱදݱ͞Εͳ͍ͱ͍͚ͳ͍
    25
    ܇࿅σʔλͷྫ
    ೖྗ
    ग़ྗ
    େ͖͞ Ͱ͜΅͜౓
    जᙾ 0.252 0.014 ྑੑ
    जᙾ 0.327 0.340 ྑੑ
    जᙾ 1.000 0.371 ѱੑ
    जᙾ 0.223 0.369 ѱੑ
    ʜ ʜ ʜ ʜ
    ਤࣔ
    େ͖͞
    Ͱ͜΅͜౓
    ྑੑͷजᙾ
    ѱੑͷजᙾ
    जᙾ
    जᙾ
    

    View Slide

26. /39
    ػցֶशͷ͘͠Έɿϧʔϧ
    w ίϯϐϡʔλ͕ܭࢉͰ͖ΔΑ͏ʹϧʔϧ΋਺ࣜͰදݱ͞ΕΔ
    w ୯७ͳϧʔϧͷܗࣜɿ
    ϧʔϧ΋਺ࣜͰදݱ͞Εͳ͍ͱ͍͚ͳ͍
    26
    େ͖͞ Ͱ͜΅͜౓ 

    ͳΒྑੑɺͦΕҎ֎ͳΒѱੑ
    z = w1
    × + w2
    × + b;
    z < 0
    

    View Slide

27. /39
    ػցֶशͷ͘͠Έɿϧʔϧ
    ϧʔϧ΋਺ࣜͰදݱ͞Εͳ͍ͱ͍͚ͳ͍
    27
    େ͖͞
    Ͱ͜΅͜౓
    ྑੑͷजᙾ
    ѱੑͷजᙾ
    [ʾͷྖҬ
    [ͷྖҬ
    ͷ৔߹ɿ

    େ͖͞ Ͱ͜΅͜౓
    w1
    = 1, w2
    = − 1, b = 0
    z = −
    େ͖͞ Ͱ͜΅͜౓
    z = w1
    × + w2
    × + b;
    

    View Slide

28. /39
    ػցֶशͷ͘͠Έɿϧʔϧ
    X X Cͷ஋͕มΘΔͱϧʔϧ΋มΘΔ
    28
    େ͖͞
    Ͱ͜΅͜౓
    ྑੑͷजᙾ
    ѱੑͷजᙾ
    [ʾͷྖҬ
    [ͷྖҬ
    ͷ৔߹ɿ

    େ͖͞ Ͱ͜΅͜౓
    w1
    = 1, w2
    = 1, b = − 1
    z = + −1
    େ͖͞ Ͱ͜΅͜౓
    z = w1
    × + w2
    × + b;
    

    View Slide

29. /39
    ػցֶशͷ͘͠Έɿϧʔϧ
    ܇࿅σʔλʹ౰ͯ͸·Δྑ͍ϧʔϧΛֶश͍ͨ͠
    29
    େ͖͞
    Ͱ͜΅͜౓
    େ͖͞
    Ͱ͜΅͜౓
    
    w1
    = 1, w2
    = − 1, b = 0
    w1
    = 1, w2
    = 1, b = − 1
    ϧʔϧ͕౰ͯ͸·Βͳ͍जᙾ
    ӈͷϧʔϧͷํ͕ɺ܇࿅σʔλʹ౰ͯ͸·Δ
    

    View Slide

30. /39
    ػցֶशͷ͘͠Έɿଛࣦ
    ϧʔϧͷྑ͠ѱ͠Λ఺਺Ͱද͢
    30
    ଛࣦɿେ ଛࣦɿখ
    େ͖͞
    Ͱ͜΅͜౓
    େ͖͞
    Ͱ͜΅͜౓
    w ଛࣦɿϧʔϧͷѱ͞Λ఺਺Ͱදͨ͠΋ͷ
    w ܇࿅σʔλʹ౰ͯ͸·Βͳ͍ϧʔϧ͸ଛࣦ͕େ͖͘ͳΔ
    

    View Slide

31. /39
    ػցֶशͷ͘͠Έɿଛࣦ
    w ʮޡ൑ఆͷ݅਺ʯΛଛࣦͱͯ͠࢖͏ͱϧʔϧͷࡉ͔͍ҧ͍͕Θ͔Βͳ͍
    w ϧʔϧʹʮѱੑͷ֬཰ʯΛग़ྗͤ͞ɺͦΕʹ΋ͱ͖ͮଛࣦΛࢉग़
    ϧʔϧͷྑ͠ѱ͠Λ఺਺Ͱද͢
    31
    ૯࿨Λϧʔϧͷଛࣦͱ͢Δ
    ଛࣦɿ
    ܇࿅σʔλ
    [
    ѱੑͷ֬཰

    ଛࣦ
    ೖྗ
    ग़ྗ
    େ͖͞ Ͱ͜΅͜౓
    जᙾ 0.252 0.014 ྑੑ -0.542 0.336 0.410
    जᙾ 0.327 0.340 ྑੑ -0.172 0.457 0.611
    जᙾ 1.000 0.371 ѱੑ 1.208 0.770 0.261
    जᙾ 0.223 0.369 ѱੑ -0.348 0.414 0.882
    ʜ ʜ ʜ ʜ
    a = σ(z)
    

    View Slide

32. /39
    ػցֶशͷ͘͠Έɿޯ഑๏
    w ଛࣦ͕࠷খͷϧʔϧɺ

    ͭ·Γ܇࿅σʔλʹ࠷΋౰ͯ͸·ΔϧʔϧΛݟ͚ͭΔͷֶ͕श
    w ޮ཰తʹͦͷΑ͏ͳϧʔϧΛݟ͚ͭΔͨΊʹޯ഑๏͕༻͍ΒΕΔ
    w ޯ഑๏Ͱ͸ɺଛࣦ͕࠷΋খ͘͞ͳΔํ޲ʹগ͚ͩ͠ਐΉ

    ʢX X Cͷ஋Λগ͚ͩ͠มԽͤ͞Δʣ͜ͱΛ܁Γฦ͢
    ଛࣦ͕࠷খͱͳΔϧʔϧʢX X Cͷ஋ʣΛݟ͚ͭΔ
    32
    

    View Slide

33. /39
    ػցֶशͷ͘͠Έɿ൚Խೳྗ
    w ֶशͷͦ΋ͦ΋ͷ໨త͸

    ʮ܇࿅σʔλʹͳ͍ະ஌ͷೖྗʹରͯ͠ਖ਼͍͠ग़ྗΛฦ͢ʯϧʔϧͷֶश
    w ܇࿅σʔλ΁ͷ౰ͯ͸·Γ͚ͩΛߟྀ͢Δͱ

    ϧʔϧ͕܇࿅σʔλʹ౰ͯ͸·Γ͗ͯ͢ະ஌ͷೖྗͰ͸ؒҧ͑ΔڪΕ
    w X Xͷ஋͕ۃ୺ʹେ͖͍ϧʔϧ͸ա৒ద߹͍ͯ͠ΔڪΕ͕͋ΔͷͰ

    ͦͷΑ͏ͳϧʔϧʹϖφϧςΟΛ༩͑ΔΑ͏ʹଛࣦΛઃܭ͢Δ
    w ʮϧʔϧ͕ະ஌ͷೖྗʹର͠ਖ਼͍͠ग़ྗΛฦ͢ೳྗʯΛ

    ൚Խೳྗͱ͍͏
    ܇࿅σʔλʹ౰ͯ͸·Γ͗͢Δϧʔϧ͸൚Խೳྗ͕௿͍
    33
    

    View Slide

34. /39
    w ௚ઢͰදݱ͞ΕΔϧʔϧ͸ೖྗͱग़ྗͷରԠ͚͕ͮෳࡶͳ৔߹ʹෆे෼
    w ਂ૚ֶशΛ༻͍ΔͱෳࡶͳϧʔϧΛֶशͰ͖Δ
    େ͖͞
    Ͱ͜΅͜౓
    େ͖͞
    Ͱ͜΅͜౓
    ୯७ͳϧʔϧ
    ػցֶशͷ͘͠Έɿਂ૚ֶश
    ෳࡶͳϧʔϧͷֶशʹ༻͍ΒΕΔͷ͕ਂ૚ֶश
    34
    େ͖͞
    Ͱ͜΅͜౓
    ਂ૚ֶशʹΑΔϧʔϧ
    

    View Slide

35. /39
    w ୯७ͳϧʔϧͷ໛ࣜਤɿ
    ػցֶशͷ͘͠Έɿਂ૚ֶश
    ਂ૚ֶशͰ͸ෳ਺ͷඇઢܗؔ਺Λଟ૚ʹॏͶΔ
    35
    େ͖͞
    Ͱ͜΅͜౓
    ྑੑ·ͨ͸ѱੑ
    w1
    w2
    a=σ(z)
    େ͖͞
    Ͱ͜΅͜౓
    ྑੑ·ͨ͸ѱੑ
    w ਂ૚ֶशʹΑΔϧʔϧͷ໛ࣜਤɿ
    

    View Slide

36. /39
    ػցֶशͷ࣮૷
    w ػցֶशͷ࣮૷ʹ͸ɺϓϩάϥϛϯάݴޠ1ZUIPO͕޿͘࢖ΘΕ͍ͯΔ
    w ྫ͑͹TDJLJUMFBSOͱ͍͏1ZUIPOϥΠϒϥϦΛ࢖͏ͱɺ

    ਺ߦͷίʔυͰػցֶशΛ࣮૷͢Δ͜ͱ͕Ͱ͖Δ
    w ͍͔ͭ͘ͷαϯϓϧσʔλ΋TDJLJUMFBSOͰఏڙ͞Ε͍ͯΔ
    ༷ʑͳϥΠϒϥϦ͕ެ։͞Ε͓ͯΓ؆୯ʹ࣮૷͕Ͱ͖Δ
    36
    https://scikit-learn.org/
    ϧʔϧͷֶश
    ϧʔϧͷద༻
    

    View Slide

37. /39
    ػցֶशͷ࣮૷
    w ਂ૚ֶशͷ1ZUIPOϥΠϒϥϦ΋༷ʑͳ΋ͷ͕ఏڙ͞Ε͍ͯΔɿ

    5FOTPS'MPX ,FSBT 1Z5PSDI $IBJOFS
    w (PPHMF$PMBCPSBUPSZΛ࢖͏ͱԾ૝ϚγϯΛ࢖ͬͯ

    ਂ૚ֶशϥΠϒϥϦΛ΢Σϒϒϥ΢β͔Β؆୯ʹ࢖͏͜ͱ͕Ͱ͖Δ
    ༷ʑͳϥΠϒϥϦ͕ެ։͞Ε͓ͯΓ؆୯ʹ࣮૷͕Ͱ͖Δ
    37
    https://colab.research.google.com
    

    View Slide

38. /39
    ػցֶशͷ࣮૷
    w ,BHHMFͳͲͷػցֶशίϯϖςΟγϣϯͰ͸

    ࣮ફతͳػցֶशͷ՝୊ɾσʔλ͕ఏڙ͞Εଞऀͱڝ͏͜ͱ͕Ͱ͖Δ
    ػցֶशίϯϖςΟγϣϯͰྗࢼ͠Λ͢Δ͜ͱ͕Ͱ͖Δ
    38
    https://www.kaggle.com/c/titanic/leaderboard
    

    View Slide

39. /39
    ػցֶशͷ·ͱΊ
    w େྔͷೖग़ྗྫͷσʔλΛ܇࿅σʔλͱͯ͠࢖͍

    ೖྗͱग़ྗΛରԠ͚ͮΔϧʔϧΛֶश
    w ܇࿅σʔλʹͰ͖Δ͚ͩ౰ͯ͸·ΔΑ͏ͳϧʔϧ

    ʢଛࣦͷখ͍͞ϧʔϧʣΛޯ഑๏Λ༻͍ͯݟ͚ͭΔ
    ܇࿅σʔλΛ༻͍ͯೖྗͱग़ྗΛରԠ͚ͮΔϧʔϧΛֶश
    39
    

    View Slide