Upgrade to Pro — share decks privately, control downloads, hide ads and more …

クックパッドと自然言語処理/cookpad-and-nlp

j.harashima
February 20, 2017
4
24k

 クックパッドと自然言語処理/cookpad-and-nlp

j.harashima

February 20, 2017
Tweet

More Decks by j.harashima

See All by j.harashima
日本語レシピデータセットの継続的な構築と複合的な利用/JED2022
junharashima
0
10k
自然言語処理の基礎と応用 〜 料理と医療を題材として 〜 /JADI2021
junharashima
0
4.2k
企業での研究開発の楽しさと苦労/WAP-Tech-Talk
junharashima
1
9.9k
クックパッドにおける研究開発/HCG2020
junharashima
0
3.5k
Calorie Estimation in a Real-World Recipe Service/iaai-20
junharashima
0
13k
クックパッドと機械学習（短縮版）/MLTBP
junharashima
0
8.1k
クックパッドと NLP/CV/nlpaper-challenge
junharashima
0
82
AI 部門の実戦投入/repro-tech-meetup
junharashima
0
3.5k
クックパッドと機械学習/uec-aix-seminar
junharashima
0
10k

Featured

See All Featured
Exploring the Power of Turbo Streams & Action Cable | RailsConf2023
kevinliebholz
28
4.5k
Put a Button on it: Removing Barriers to Going Fast.
kastner
60
3.6k
Documentation Writing (for coders)
carmenintech
67
4.5k
The Cult of Friendly URLs
andyhume
78
6.1k
The Pragmatic Product Professional
lauravandoore
32
6.4k
Art, The Web, and Tiny UX
lynnandtonic
298
20k
Build your cross-platform service in a week with App Engine
jlugia
229
18k
GitHub's CSS Performance
jonrohan
1030
460k
Distributed Sagas: A Protocol for Coordinating Microservices
caitiem20
330
21k
Refactoring Trust on Your Teams (GOTO; Chicago 2020)
rmw
33
2.7k
Let's Do A Bunch of Simple Stuff to Make Websites Faster
chriscoyier
507
140k
Scaling GitHub
holman
459
140k

Transcript

  1. ΫοΫύουͱࣗવݴޠॲཧ ΫοΫύουגࣜձࣾݪౡ७

  2. ೥݄ ೥݄ ژ౎େֶݴޠϝσΟΞݚڀࣨ഑ଐ ΫοΫύουೖࣾ ݪౡ७ ೥݄ ৽ઃͷݚڀ։ൃ෦໳ʹҟಈ ത࢜ʢ৘ใֶʣऔಘ ೥݄ ࣗવݴޠॲཧͷݚڀ

    ɾಉٛදݱͷ֫ಘʢ*+$/-1ʣ ɾݕࡧ݁Ռͷཁ໿ʢ/-1*9ʣ ɾݕࡧ݁ՌͷϦϥϯΩϯάʢ*+$/-1ʣ ɾ೔ຊޠจͷѹॖʢ$0-*/(ʣ αʔϏε։ൃ෦໳ ࣗવݴޠॲཧͱը૾ೝࣝͷݚڀɾ։ൃ ɾϨγϐίʔύεͷߏஙʢ-3&$ʣ ɾϨγϐͷ຋༁ʢ8"5ʣ

  3. ΫοΫύου Πϯλʔωοτ্ͰϨγϐͷ౤ߘɾݕࡧ౳͕Ͱ͖Δ ೔ຊ࠷େͷྉཧϨγϐαʔϏε

  4. ɾλΠτϧ ɾ֓ཁ ɾࡐྉ ɾखॱ ɾίπɾϙΠϯτ ɾ͜ͷϨγϐͷੜཱ͍ͪ ɾͭ͘ΕΆ ɾ ౤ߘϨγϐ

  5. ݕࡧ݁Ռ ৽ணॱ ਓؾॱ

  6. ݄࣍ར༻ऀ਺ ݄࣍ར༻ऀ਺ʢւ֎ʣ ೔࣍ར༻ऀ਺ ໿ ສਓ ໿ສਓ Ϩγϐ਺ ໿ສ඼ ໿ ສਓ

  7. ΫοΫύουͱࣗવݴޠॲཧ

  8. Ͳ͜Ͱʁ ɾϨγϐͷղੳ ɾಉٛදݱͷೝࣝ ɾϨγϐͷ෼ྨ ɾϨγϐͷ຋༁ ɾͦͷଞ

  9. Ϩγϐͷղੳ ϨγϐΛΠϯσΩγϯά͢ΔͨΊɺϨγϐதͷ֤ςΩετΛղੳ ಲͷʹΜʹ͘ে༉ম͖ ͷ ʹΜʹ͘ ে༉ ໊ࢺ ໊ࢺ ॿࢺ ໊ࢺ

    c c c c ಲ ম͖ ಈࢺ ܗଶૉղੳ

  10. ܗଶૉղੳ༻ࣙॻͷ֦ு FDIPΑͩΕܲcNFDBC0XBLBUJ ΑͩΕܲ FDIPΑͩΕܲcNFDBC0XBLBUJVDPPLQBEEJD ΑͩΕܲ ʢ˞ਓखʣ

  11. Ͳ͜Ͱʁ ɾϨγϐͷղੳ ɾಉٛදݱͷೝࣝ ɾϨγϐͷ෼ྨ ɾϨγϐͷ຋༁ ɾͦͷଞ

  12. ಉٛදݱͷೝࣝ ΠϯσΩγϯάͷࡍʹಉٛදݱΛٵऩ

  13. ಉٛදݱೝࣝ༻ࣙॻͷߏஙʢ˞ਓखʣ

  14. Ͳ͜Ͱʁ ɾϨγϐͷղੳ ɾಉٛදݱͷೝࣝ ɾϨγϐͷ෼ྨ ɾϨγϐͷ຋༁ ɾͦͷଞ

  15. Ϩγϐͷ෼ྨ Ϣʔβ͕ϒοΫϚʔΫͨ͠ϨγϐΛ47.Ͱ෼ྨ ྫ͑͹ɺʮ͓՛ࢠʯΛ ΫϦοΫ͢Δͱʜ

  16. ෼ྨثͷ࣮૷ DMBTTJpFS3FDJQF$MBTTpFSOFX lNFBUz "$$&15"#-&@13&$*4*0/  DMBTTJpFSUSBJO QPTJUJWF@SFDJQFT OFHBUJWF@SFDJQFT  DMBTT3FDJQF$MBTTJpFS

    EFGUSBJO QPTJUJWF@SFDJQFT OFHBUJWF@SFDJQFT  ʜ "$$&15"#-&@13&$*4*0/Λຬͨͤ͹TBWF NPEFMTBWF QBUI JGBDDFQUBCMF NPEFM  FOE FOE ϢʔβͷਪનϨγϐΛ ਖ਼ղσʔλͱͯ͠ར༻

  17. Ͳ͜Ͱʁ ɾϨγϐͷղੳ ɾಉٛදݱͷೝࣝ ɾϨγϐͷ෼ྨ ɾϨγϐͷ຋༁ ɾͦͷଞ

  18. ೔ຊޠͷϨγϐΛχϡʔϥϧ.5Ͱӳޠʹ຋༁ ݚ ڀ ஈ ֊ Ϩγϐͷ຋༁ ػց຋༁

  19. "UUFOUJPO.PEFM #BIEBOBVFUBM ֶशσʔλ ɾ Ϩγϐͷର༁ ࣮૷ ɾ1ZUIPOX$IBJOFS ຋༁ྫ ɾ֖·ͨ͸ΞϧϛϗΠϧͰམͱ֖͠Λͯ͠தՐͰৠ͠ম͖ʹ ɹ͠·͢ɻ

    ɹɹɹɹɹɹɹɹɹɹɹɹɹˣ ɹDPWFSXJUIBMJEPSBMVNJOVNGPJM BOETUFBNUIF ɹDIJDLFOPONFEJVNIFBU

  20. Ͳ͜Ͱʁ ɾϨγϐͷղੳ ɾಉٛදݱͷೝࣝ ɾϨγϐͷ෼ྨ ɾϨγϐͷ຋༁ ɾͦͷଞ

  21. ສ඼ͷϨγϐͱͦΕΒ͔ΒͳΔݙཱʹؔ͢Δσʔλ ɾλΠτϧ ɾ֓ཁ ɾखॱ ɾͭ͘ΕΆ ɾΧςΰϦ ɾʜ Ϩγϐίʔύεͷެ։

  22. ݚڀͷ޿͕Γ ެ։લʢʙ೥݄ʣ  େֶݚڀࣨ ެ։ޙʢ೥݄ʣ  େֶݚڀࣨ

  23. ࠓޙͷ՝୊

  24. ࠓޙͷ՝୊ ɾܗଶૉղੳͷਫ਼౓޲্ ɹɾݱঢ়͕Ͳͷఔ౓ͷਫ਼౓ͳͷ͔͕ʢ͋·Γʣ෼͔͍ͬͯͳ͍ ɹɾςετηοτΛͪΌΜͱߏஙͯ͠ɺண࣮ʹਫ਼౓Λ޲্͍ͤͨ͞ ɾಉٛදݱͷࣗಈ֫ಘ ɹɾڈ೥͔Βւ֎ల։ΛՃ଎ˠࠃ͝ͱʹࣙॻΛߏங͢Δͷ͸ͩΔ͍ʜ ɹɾΫΤϦ΍Ϩγϐ͔ΒಉٛදݱΛࣗಈͰ֫ಘ͍ͨ͠ ɾϨγϐͷཁ໿ ɹɾʮجຊతͳϨγϐʢ͓͓·͔ͳϨγϐʣ͕஌Γ͍ͨʯͱ͍͏ཁ๬͕ଟ͍ ɹɾෳ਺ͷϨγϐΛཁ໿͢Δ͜ͱͰɺجຊతͳϨγϐΛ࡞੒͍ͨ͠

    ɾର࿩ʹΑΔݕࡧʁ ɹɾ΋͏ͦ͏͍͏࣌୅ʁʢؾͷ͍ͤʁʣ

  25. 8F`SFIJSJOH ɾϦαʔναΠϯςΟετʢओʹݚڀ୲౰ʣ ɹɾαʔϏεʹಋೖ͞ΕΔػցֶशͷϩδοΫ෦෼ͷ࣮૷ ɹɾ*P5΍.3ʹ΋ͱͮ͘৽͍͠αʔϏεͷاը ɹɾ࠷৽ͷٕज़ͷΠϯϓοτͱΞ΢τϓοτ ɾϦαʔνΤϯδχΞʢओʹ։ൃ୲౰ʣ ɹɾݚڀ։ൃج൫Λޮ཰Խ͢ΔͨΊͷ։ൃج൫ͷઃܭͱߏங ɹɾػցֶशΛར༻ͨ͠εέʔϥϒϧͳαʔϏεͷ։ൃͱӡ༻ ɹɾ"OESPJEJ04ʹ͓͚ΔϓϩτλΠϐϯά ͝ڵຯ͋ΔํͷΤϯτϦʔΛ͓଴͍ͪͯ͠·͢

    IUUQTSFDSVJUDPPLQBEDPN