$30 off During Our Annual Pro Sale. View Details »

Spacyでお手軽NLP / NLP with spacy

himkt
June 13, 2018
Programming
0
830

Spacyでお手軽NLP / NLP with spacy

2018/06/13のレトリバセミナーのスライドです

himkt

June 13, 2018
Tweet

More Decks by himkt

See All by himkt
Linformer: paper reading
himkt
0
100
RoBERTa: paper reading
himkt
1
170
NLP SoTA 勉強会 / ner_2019
himkt
2
1.1k
自然言語処理 @ クックパッド / nlp at cookpad
himkt
1
390
Interpretable Machine Learning 6.3 - Prototypes and Criticisms
himkt
2
100
ニューラル固有表現抽出 / Neural Named Entity Recognition
himkt
3
450
ニューラル固有表現抽出器を実装してみる / PyNER
himkt
6
1.8k
Deep Learning Book 10その2 / deep learning book 10 vol2
himkt
2
120
ふわふわ系列ラベリング / ner 2018
himkt
5
820

Other Decks in Programming

See All in Programming
Composing an API with Kotlin vol 2 (Advanced Kotlin Dev Day 2022)
zsmb
1
200
OAI3を使った Django REST frameworkの ドキュメント生成とカスタマイズ / DjangoCongress JP 2022
shihono
0
700
KubeCon + CNCon NA 2022 Overview & Towards Something Better Than CRDs In a Post-Operator World / k8sjp54-kubecon-recap
masayaaoyama
0
470
GISとDjango GEOS APIの話
mierune
0
300
redux使うのやめました
hirasa
1
140
LintオタクによるLint解説 / Introduction to Lints
orgachem
20
5.8k
ぱっと見犬猫っぽいAWSアーキテクチャ図を Amazon Rekognitionで判定しよう
kzkmaeda
0
260
Ship Faster With Feature Flags
davidodari
0
160
RuboCop Under a Microscope
koic
8
1.2k
WebAssembly for the Backend
adriancole
0
160
AWS発デザインシステム「Cloudscape」を使ってUI爆速開発
j_3woo86
1
120
SvelteKitを本番投入してみて
kubotak
0
230

Featured

See All Featured
GraphQLとの向き合い方2022年版
quramy
19
9.3k
Code Reviewing Like a Champion
maltzj
508
38k
A better future with KSS
kneath
228
16k
GraphQLの誤解/rethinking-graphql
sonatard
36
7.6k
The Mythical Team-Month
searls
210
40k
10 Git Anti Patterns You Should be Aware of
lemiorhan
643
54k
The Invisible Side of Design
smashingmag
292
48k
Pencils Down: Stop Designing & Start Developing
hursman
113
10k
Imperfection Machines: The Place of Print at Facebook
scottboms
254
12k
Teambox: Starting and Learning
jrom
124
7.8k
How to Ace a Technical Interview
jacobian
269
21k
Learning to Love Humans: Emotional Interface Design
aarron
262
38k

Transcript

  1. Ͱ͓खܰ/-1 )JSBNBUTV!ϨτϦόηϛφʔ  ը૾IUUQTHJUIVCDPNFYQMPTJPOTQB$ZCMPCNBTUFSXFCTJUFBTTFUTJNHMPHPTWH

  2.   Tsukuba, M2, NLP himkt

  3. 5-%3 w 1ZUIPOͷࣗવݴޠॲཧϥΠϒϥϦͰ͋Δ4QB$Zͷ঺հ w Wͷ͓΋͠Ζػೳʹ͍ͭͯ w 4QB$ZͰ೔ຊޠςΩετΛॲཧ͢Δ  

  4. "CPVU4QB$Z   ը૾IUUQTTQBDZJP https://spacy.io

  5. IUUQTTQBDZJP "CPVU4QB$Z “Industrial-Strength NLP” w /POEFTUSVDUJWFUPLFOJ[BUJPO w /BNFEFOUJUZSFDPHOJUJPO w 4VQQPSUGPS

    MBOHVBHFT w TUBUJTUJDBMNPEFMTGPSMBOHVBHFT w ʜFUD   IUUQTHJUIVCDPNFYQMPTJPOTQB$Z

  6. #BTJDVTBHFPG4QB$Z  

  7. .BOZ'FBUVSFT   ը૾IUUQTTQBDZJPVTBHFGBDUTpHVSFT

  8. 'BTUFTUJOUIFXPSME w %FQFOEFODZQBSTFSͷύϑΥʔϚϯεൺֱ w ࿦จIUUQTBDMXFCPSHBOUIPMPHZ111QEG w ࿦จʹ͋Δͷ͸41&&%ͷදͰɼ"DDVSBDZ͸ࣗલͰ࡞ΒΕͨ΋ͷʁ w $IPJͷϕϯνϚʔΫ࣌ʹ͸TQB$ZW͸ະϦϦʔεͳͷͰOB 

     ը૾IUUQTTQBDZJPVTBHFGBDUTpHVSFT

  9. 4QFFEDPNQBSJTPOXJUIPUIFSMJCSBSJFT w ଞͷࣗવݴޠॲཧϥΠϒϥϦͱͷ଎౓ൺֱ w ࿦จͰ͸ͳ͘࡞ऀ͕ௐࠪͨ͠΋ͷ DPOEVDUFEJO  w ϦϙδτϦIUUQTHJUIVCDPNFYQMPTJPOTQBDZCFODINBSLT 

     ը૾IUUQTTQBDZJPVTBHFGBDUTpHVSFT

  10. .PEFMDPNQBSJTPO w ݴޠʹΑͬͯ͸ෳ਺ͷαΠζͷϞσϧ͕͋Δ FO GS FT  w 104UBHHFS /&3UBHHFS

    %FQFOEFODZQBSTFS w 8JUIPVUBOZQSFQSPDFTTJOH EBUBTFU   ը૾IUUQTTQBDZJPVTBHFGBDUTpHVSFT

  11. /-5,BOE4QB$Z w /-5,P⒎FSTTPNFPGUIFTBNFGVODUJPOBMJUZBTTQB$Z w *ODPNQBSJTPOUPTQB$Z /-5,UBLFTBNVDINPSF
 CSPBEDIVSDIBQQSPBDI w TQB$ZJTBMTPNVDINPSFQFSGPSNBODFGPDVTTFEUIBO/-5, XIFSFUIFUXPMJCSBSJFTQSPWJEFUIFTBNFGVODUJPOBMJUZ

     TQB$ZTJNQMFNFOUBUJPOXJMMVTVBMMZCFGBTUFSBOENPSF BDDVSBUF   Ҿ༻IUUQTTQBDZJPVTBHFGBDUTpHVSFT

  12. 0UIFSMJCSBSJFTBOETQB$Z   1ZUPSDIIUUQTHJUIVCDPNQZUPSDIQZUPSDICMPCNBTUFSEPDTTPVSDF@TUBUJDJNHQZUPSDIMPHPEBSLTWH "MMFO/-1IUUQTHJUIVCDPNBMMFOBJBMMFOOMQCMPCNBTUFSEPDTUBUJDBMMFOOMQMPHPEBSLQOH (FOTJNIUUQTHJUIVCDPN3B3F5FDIOPMPHJFTHFOTJNCMPCEFWFMPQEPDTTSDSFBENF@JNBHFTSBSFQOH $V1ZIUUQTHJUIVCDPNDVQZDVQZCMPCNBTUFSEPDTJNBHFDVQZ@MPHP@QYQOH JOUPSDIUFYU GPS(16BDDFMFSBUJPO XPSEWFDUPS

    QJQFMJOF

  13. 4QB$ZWͷݸਓతʹ޷͖ͳػೳ w EJTQMB$ZͰ͌͢΀Εʔ͠ʔʁ w 4QB$ZͷՄࢹԽϞδϡʔϧ w ͖Ε͍Ͱ͍͍ײ͡ͳը૾Λ࡞ͬͯ͘ΕΔ w ݻ༗දݱநग़ͱ܎Γड͚ղੳͷ݁ՌΛՄࢹԽͯ͘͠ΕΔ w

    4QB$ZͰղੳͨ͠ΦϒδΣΫτΛͦͷ··࢖͑Δ w 47(ܗࣜͷը૾͕ग़ྗ͞ΕΔ  

  14. 8FCαΠτ্ͷ/&3ͷσϞ   ը૾IUUQTFYQMPTJPOBJEFNPTEJTQMBDZFOU

  15. 4QB$ZY+VQZUFS/PUFCPPL  

  16. 4QB$ZBOE೔ຊޠ w WͰͷ೔ຊޠରԠ 13  w ೔ຊޠܗଶૉղੳث+BOPNFΛϥοϓ͢ΔܗͰ࣮૷ w WͰͷܗଶૉղੳثҠߦ *TTVF

    13  w ೔ຊޠ6OJWFSTBM%FQFOEFODZσʔλ͸6OJ%JDͰ෼ׂ͞Ε͍ͯΔ w +BOPNF͸ݱࡏͷͱ͜Ζ6OJ%JDʹະରԠ w .F$BCʹҠߦ  

  17. 4QB$ZBOE೔ຊޠ  

  18. 4QB$ZBOE೔ຊޠ   "OTXFSVTF6OJ%JD

  19. 4QB$ZBOE೔ຊޠ  

  20. ೔ຊޠ/&3%FQFOEFODZQBSTJOHXJUI4QB$Z w ݁࿦ݱࡏ͸Ͱ͖ͳ͍ w 4QB$Zʹ͸ࣗ෼Ͱ5BHHFS΍1BSTFSΛֶशͰ͖Δ
 ػߏ͕૊Έࠐ·Ε͍ͯΔ w ೔ຊޠ6OJWFSTBM%FQFOEFODJFT͸ެ։͞Ε͍ͯΔ
 IUUQTHJUIVCDPN6OJWFSTBM%FQFOEFODJFT6%@+BQBOFTF(4%ͳͲ 

     Ϟσϧࣗ࡞Ͱ͖ΔͷͰ͸ʂʁ

  21. "EEJOH-BOHVBHFTVQQPSU   4QB$ZͷݴޠϞδϡʔϧͷ
 ίϯϙʔωϯτ ը૾IUUQTTQBDZJPVTBHFBEEJOHMBOHVBHFTTFDUJPOUSBJOJOH

  22. "EEJOH-BOHVBHFTVQQPSU   ը૾IUUQTTQBDZJPVTBHFBEEJOHMBOHVBHFTTFDUJPOUSBJOJOH 1PXFSFECZ.F$BC
 4QB$ZͷֶशϞδϡʔϧ͕࢖͑ͳ͍ʁ

  23. ೔ຊޠTQB$Zͷݱࡏͷ໰୊ʁ w ࣙॻ͕6OJ%JDͰ͋Δ͜ͱ͕૝ఆ͞Ε͍ͯΔ JTTVF  w ʢ͓ͦΒ͘ʣଟ͘ͷ؀ڥͰ͸.F$BCͷσϑΥϧτࣙॻ͸*1"EJD w <50%0>+BQBOFTF5PLFOJ[FSͰ5BHHFSΛ࡞͍ͬͯΔ͕ɼ
 ͜ͷίϯετϥΫλͷҾ਺͸֎෦͔Β৮Εͳͦ͏

    w ൃԻ͕ະొ࿥ͷ୯ޠΛղੳ͢ΔͱΤϥʔ *1"EJD  13  w ࣙॻ͝ͱͷग़ྗͷࠩҟΛٵऩ͢ΔΠϯλʔϑΣʔε͕ඞཁʁ w /&3΍%FQFOEFODZ1BSTJOHͷϞσϧ͸·ͩଘࡏ͠ͳ͍ w ݱࡏ͸A"MQIBUPLFOJ[BUJPOTVQQPSUA w ෼͔ͪॻ͖͕ඞཁͳݴޠͷରԠ͸Ͳ͏͢Ε͹ʜ w தࠃޠ΋ࣅͨঢ়گͰࢭ·͍ͬͯΔ 13  

  24. ͜ΕΛ΍Ε͹ͱΓ͋͑ͣςετ͸Ͱ͖ͦ͏  USBWJTͰ.F$BCΛΠϯετʔϧ͢ΔΑ͏ʹ͢Δ
  TQBDZNPEFMTʹ6OJ%JDΛొ࿥͢Δ
  QZUIPONTQBDZEPXOMPBEKBΛ੔උ͢Δ   /&3ͱ%FQFOEFODZ1BSTJOH͸͋ͱͰௐ΂Δʜ

  25. ·ͱΊ w ࣗવݴޠॲཧϥΠϒϥϦTQB$Zͷ঺հ w *OEVTUSZൃͷࣗવݴޠॲཧϥΠϒϥϦ w WͰೖͬͨՄࢹԽϞδϡʔϧ͕͍͍ײ͡ w EJTQMB$ZΛ࢖͍͍ͬͨײ͡ͳՄࢹԽ w

    TQB$ZͰͷ೔ຊޠςΩετॲཧ͸·ͩෆ׬શ w ݱঢ়ܗଶૉղੳͷΠϯλʔϑΣʔε   QJQJOTUBMMTQBDZ 3VO5IJTDPNNBOE