データ分析コンテストの
勝者解答から学ぶ

4742812a011db89b01a52af6722640b8?s=47 @smly
March 03, 2018
15k

 データ分析コンテストの
勝者解答から学ぶ

ステアラボ人工知能シンポジウム 2018 講演資料
https://stair.connpass.com/event/76647/

4742812a011db89b01a52af6722640b8?s=128

@smly

March 03, 2018
Tweet

Transcript

  1. 2.

    ࣗݾ঺հ ϦΫϧʔτςΫϊϩδʔζ"5-ͷ4S4PGUXBSF&OHJOFFSͰ͢ɽ ˝,BHHMFΦϑΟεͰͷू߹ࣸਅ αϯϑϥϯγεί ɾ,BHHMFSྺ೥ (SBOENBTUFS )JHIFTUSBOLUI 
 ɾ,BHHMF 5PQpOJTIFTY

    1SJ[FY 
 ɾ5PQ$PEFS.BSBUIPO.BUDI 8*/4 
 ɾ"$.,%% ,%%$VQTUQSJ[FXJOOFS ˝ϦΫϧʔτςΫϊϩδʔζ"5- ޿ඌ ˞લճͷߨԋ͔Β̏ݸ૿͑·ͨ͠
  2. 3.

    ࠓ೔ͷ࿩ w ͜͜਺೥ͷ,BHHMFͱσʔλ෼ੳίϯςετ   w ͲͷΑ͏ʹͯ͠σʔλ෼ੳίϯςετͰউ͔ͭ   w

    ࠷ۙͷίϯςετʹ͓͚Δղ๏ͷղઆ  ࠓճ͸Ұൠ࿦ΑΓ΍΍֤࿦ʹ౿ΈࠐΜͩ࿩Λͯ͠Έ·͢ɻ ⼀一般 ⼀一般 応⽤用 ͕͜͜ຊ୊
  3. 4.

    ࠓ೔ͷ࿩ w ͜͜਺೥ͷ,BHHMFͱσʔλ෼ੳίϯςετ   w ͲͷΑ͏ʹͯ͠σʔλ෼ੳίϯςετͰউ͔ͭ   w

    ࠷ۙͷίϯςετʹ͓͚Δղ๏ͷղઆ  ࠓճ͸Ұൠ࿦ΑΓ΍΍֤࿦ʹ౿ΈࠐΜͩ࿩Λͯ͠Έ·͢ɻ
  4. 5.

    ৼΓฦΓɿ//๖ժɾπʔϧ։ൃՃ೤  w ೥͝Ζ͸4UF⒎FO3FOEMFͷMJC'. 'BDUPSJ[BUJPO .BDIJOFT<> ΍3JF+PIOTPOͷ3(' 3FHVMBSJ[FE(SFFEZ 'PSFTU<> ͳͲɺΞϧΰϦζϜͷ৽ن։ൃʹΑͬͯଞͷࠩΛ͚ͭ

    ͯউར͢Δ৔໘͕Կ౓͔͋ͬͨɻ NN 萌芽・
 ツール開発過熱 2012-2014 フランケンシュタイン
 アンサンブル 2014-2016? NN 時代・
 ⼤大規模MMデータ 2016- Kernel コンペ・
 強化学習? 2017- w ػցֶश΍σʔλϚΠχϯάΛ׆༻͢Δਓ޻͕૿Ճɺ৽͍͠ಓ۩͕ ࣍ʑʹੜ·Εͨɻ-BTBHOF DYYOFU LFSBT 9(#PPTUͳͲɻ։ൃ ऀຊਓ͕ίϯςετΛ௨ͯ͡ιϑτ΢ΣΞͷϕϯνϚʔΫΛߦͬͨ Γએ఻͢ΔͳͲ͍ͯͨ͠ɻ
  5. 6.

    ৼΓฦΓɿ//๖ժɾπʔϧ։ൃՃ೤  w ೥ͷ*NBHF/FU *-473$ ʹ͓͚ΔϒϨʔΫεϧʔ͸େ͖ͳি ܸͰ͸͕͋ͬͨɺ·ͩ(16Λѻ͏؀ڥΛ੔උͨ͠Γ$6%"Λ࢖ͬ ͨϓϩάϥϜΛ,BHHMFͰ࢖͏Ϣʔβʔ͸গ਺೿Ͱ͋ͬͨɻ w ೥ʹ͸+PC4BMBSZ1SFEJDUJPO$IBMMFOHFͰ)JOUPOͷݚڀࣨ

    ͷ0#͕εύʔεಛ௃ྔΛೖྗͱͨ͠.-1Λ࣮૷ͯ͠ଞΛѹ౗ɻ NN 萌芽・
 ツール開発過熱 2012-2014 フランケンシュタイン
 アンサンブル 2014-2016? NN 時代・
 ⼤大規模MMデータ 2016- Kernel コンペ・
 強化学習? 2017-
  6. 7.

    ৼΓฦΓɿϑϥϯέϯγϡλΠϯ  w େྔʹϞσϧΛ࡞ΓελοΩϯά͢Δͱ͍͏ख๏͕ྲྀߦΔɻ༩͑Β ΕͨࢦඪͰগ͠Ͱ΋্ճ͍ͬͯΕ͹উར͢Δͱ͍͏ίϯςετͷ࢓ ૊Έ্ɺτοϓ૪͍͕ᷮࠩͰ͋Ε͹ϞσϧΛ૿΍͢͜ͱʹΠϯηϯ ςΟϒ͕ੜ·ΕΔɻ NN 萌芽・
 ツール開発過熱

    2012-2014 フランケンシュタイン
 アンサンブル 2014-2016? NN 時代・
 ⼤大規模MMデータ 2016- Kernel コンペ・
 強化学習? 2017- 出典:http://blog.kaggle.com/2016/04/08/homesite-quote-conversion-winners-write-up-1st-place-kazanova-faron-clobber/ w ϞσϧɺϞσϧͱΞϯα ϯϒϧʹ࢖͏Ϟσϧ਺ʹࡍݶ͕ͳ͘ ͳΓɺGSBOLFOTUFJOFOTFNCMFͳͲ ͱᎏ᎐͞ΕΔ͜ͱ΋ɻ
  7. 10.

    ৼΓฦΓɿ,FSOFMɾڧԽֶशίϯϖ  w ܭࢉࢿݯͱਫ਼౓ͷτϨʔυΦϑ͕஫໨͞Ε͸͡ΊΔɻ5XP4JHNB  .FSDBSJͳͲͷܭࢉࢿݯ΍࣮ߦ؀ڥʹ੍໿ΛՃ͑ͨίϯςετ͕ొ ৔ɻ΄͔5PQ$PEFS։࠵΋4QBDF/FU$IBMMFOHF΋ܭࢉࢿݯͷ੍ ໿ΛՃ͍͑ͯΔ w ೥ʹ͸ڧԽֶशͳͲͷίϯςετ΋ϓϥϯʹ͋Δͱද໌͞Εɺ

    ϢχʔΫͳλεΫͷίϯςετ͕૿͍͑ͯ͘ ͱࢥΘΕΔɻ ग़యɿz3FWJFXJOHBOE1SFWJFXJOHCMPHLBHHMFDPNSFWJFXJOHBOEQSFWJFXJOH NN 萌芽・
 ツール開発過熱 2012-2014 フランケンシュタイン
 アンサンブル 2014-2016? NN 時代・
 ⼤大規模MMデータ 2016- Kernel コンペ・
 強化学習? 2017-
  8. 11.

    ݴ͍͔ͨͬͨ͜ͱΛ੔ཧ͢Δͱ w Ξϯαϯϒϧͷڝ͍߹͍͸ෆໟ͕ͩɺ໰୊ઃఆʹ΋ݪҼ͕͋Δɻ w //ͷݚڀ։ൃ͕׆ൃʹͳ͓͔ͬͨ͛Ͱղ๏΋όϦΤʔγϣϯ͕๛ ෋ʹͳͬͨɻ w ܭࢉࢿݯͷ੍໿ΛೖΕͨΓɺڧԽֶशΛ୊ࡐʹ͢ΔͳͲมԽΛଓ͚ ͍ͯΔɻ NN

    萌芽・
 ツール開発過熱 2012-2014 フランケンシュタイン
 アンサンブル 2014-2016? NN 時代・
 ⼤大規模MMデータ 2016- Kernel コンペ・
 強化学習? 2017-
  9. 12.

    ࠓ೔ͷ࿩ w ͜͜਺೥ͷ,BHHMFͱσʔλ෼ੳίϯςετ   w ͲͷΑ͏ʹͯ͠σʔλ෼ੳίϯςετͰউ͔ͭ   w

    ࠷ۙͷίϯςετʹ͓͚Δղ๏ͷղઆ  ࠓճ͸Ұൠ࿦ΑΓ΍΍֤࿦ʹ౿ΈࠐΜͩ࿩Λͯ͠Έ·͢ɻ
  10. 19.

    աֶशͷ௚ײతͳྫɿ#PPTUJOH"UUBDL ୯७Խͨ͠/ݸͷςετࣄྫͷೋ஋෼ྨ໰୊Λߟ͑Δɿ 正解 予測 1 予測 2 ࣍ϕΫτϧΛ౰ͯΔɽධՁࢦඪ͸Τϥʔ཰ͱ͢Δɽ y 2

    {0, 1}N sH(yi) yi 2 {0, 1}N ͋Δ༧ଌ ͷ1VCMJD-#είΞΛ ͱ͢Δɽ N 1 0 0 1 0 0 0 1 y = 1 1 1 0 0 1 1 1 y1 = 1 0 1 1 1 1 0 0 y2 = Public LB
 Score 1 Public LB
 Score 2 sH(y1) = 0.75 sH(y2) = 0.25 Public Private
  11. 20.

    #PPTUJOH"UUBDL<#MVN)BSEU`> Algorithm (Boosting Attack): 正解 1 0 0 1 0

    0 0 1 y = 0 0 1 0 0 0 0 1 1 1 1 0 1 0 1 1 0 1 0 1 0 1 0 0 1 0 0 1 0 0 0 1 1 0 0 1 0 1 1 1 0 0 0 0 1 1 1 0 ランダムに予測のベクトルを作成する Public Private 1 0 1 1 1 1 0 0 1 0 0 1 0 1 1 1 0 0 0 1 0 1 1 0 0 0 1 1 0 0 1 1 0 0 1 0 0 0 1 1 1 1 0 1 0 0 0 0 ϥϯμϜͳ༧ଌ͔Β1VCMJD4DPSFͷྑ͍݁ՌΛબ୒͢Δ
  12. 21.

    #PPTUJOH"UUBDL<#MVN)BSEU`> 正解 1 0 0 1 0 0 0 1

    y = 1 0 1 1 1 1 0 0 1 0 0 1 0 1 1 1 0 0 0 1 0 1 1 0 1 0 0 1 0 0 0 1 1 0 0 1 0 1 1 1 1 1 0 1 0 0 0 0 㱺 majority
 voting 1 0 0 1 0 1 1 1 Public LB
 Score sH(ˆ y) = 0.0 Algorithm (Boosting Attack): Public LB スコアの良いベクトル sH (yi) < 0.5 だけを選ぶ Public Private ϥϯμϜͳ༧ଌ͔Β1VCMJD4DPSFͷྑ͍݁ՌΛબ୒͢Δ
  13. 22.

    #PPTUJOH"UUBDL<#MVN)BSEU`> 正解 1 0 0 1 0 0 0 1

    y = 1 0 1 1 1 1 0 0 1 0 0 1 0 1 1 1 0 0 0 1 0 1 1 0 1 0 0 1 0 0 0 1 1 0 0 1 0 1 1 1 1 1 0 1 0 0 0 0 㱺 majority
 voting 1 0 0 1 0 1 1 1 Public LB
 Score sH(ˆ y) = 0.0 Private LB Score が
 良くなる保証は
 何もない !! Algorithm (Boosting Attack): Public Private Public LB スコアの良いベクトル sH (yi) < 0.5 だけを選ぶ ϥϯμϜͳ༧ଌ͔Β1VCMJD4DPSFͷྑ͍݁ՌΛબ୒͢Δ
  14. 24.

    αʔϕΠɿ෼໺ಠಛͷΞϓϩʔνΛֶͿ ˙੺֎෼ޫ๏ʹΑΔ౔৕ௐࠪͷ؍ଌσʔλʹର͢Δલॲཧྫ
 ʢ"GSJDB4PJM1SPQFSUZ1SFEJDUJPO$IBMMFOHFʣɿ  4BWJU[LZ(PMBZpMUFS  $POUJOVVN3FNPWBM  %JTDSFUFXBWFMFU USBOTGPSNT

     'JSTU%FSJWBUJWFT  6OTVQFSWJTFE'FBUVSF4FMFDUJPO  -PH USBOTGPSNʜ ˙೴࣓ਤه࿥ͷલॲཧɾಛ௃ྔநग़ %FD.FH$IBMMFOHF ɿ PSEFS#VUUFSXPSUICBOEQBTTpMUFS 4QBUJBM'JMUFSJOH 3JFNBOOJBO.FBO  5BOHFOU4QBDF.BQQJOHʜ IUUQTXXXLBHHMFDPNDBGTJTTPJMQSPQFSUJFTEJTDVTTJPO IUUQTHJUIVCDPNBMFYBOESFCBSBDIBOU%FD.FH
 IUUQTXFCBSDIJWFPSHXFCIUUQBMFYBOESFCBSBDIBOUPSHXQDPOUFOUVQMPBETEPDVNFOUBUJPOQEG
  15. 25.

    αʔϕΠɿਂ૚ֶशͷ࣮ફతͳݚڀ੒Ռ ࢖͑ͦ͏ͳ΋ͷ͸νΣοΫ͢Δɻ̍͜͜ʙ̎೥ͷ੒Ռ͕࢖ΘΕΔ͜ͱ΋ ௝͘͠ͳ͍ɻຊεϥΠυͰ঺հ͢Δ্Ґղ๏͕ࢀߟʹͨ͠จݙͷྫɿ [Zhang+ ’17] "mixup: Beyond Empirical Risk Minimization",

    https:// arxiv.org/abs/1710.09412 [Miech+ '17] "Learnable pooling with Context Gating for video classification", https://arxiv.org/abs/1706.06905 [Tian+ '16] "Detecting Text in Natural Image with Connectionist Text Proposal Network”, In Proc. of ECCV ‘16 https://arxiv.org/abs/1609.03605 [Qi+ '17] "PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation”, In Proc. of the CVPR ‘17 https://arxiv.org/abs/1612.00593 [Ma+ '17] "Multi-View Deep Learning for Consistent Semantic Mapping with RGB-D Cameras”, In Proc. of the IROS ‘17 https://arxiv.org/abs/1703.08866
  16. 26.

    ࠓ೔ͷ࿩ w ͜͜਺೥ͷ,BHHMFͱσʔλ෼ੳίϯςετ   w ͲͷΑ͏ʹͯ͠σʔλ෼ੳίϯςετͰউ͔ͭ   w

    ࠷ۙͷίϯςετʹ͓͚Δղ๏ͷղઆ  ࠓճ͸Ұൠ࿦ΑΓ΍΍֤࿦ʹ౿ΈࠐΜͩ࿩Λͯ͠Έ·͢ɻ
  17. 37.

    4JOHMFNPEFMT #FTUTJOHMFNPEFM͸1VCMJD-#ͰҐ νʔϜத ɻ 1VCMJD-# Ґ૬౰ 1VCMJD-# Ґ૬౰ 1VCMJD-# Ґ૬౰

    1VCMJD-# Ґ૬౰ 1VCMJD-# Ґ૬౰ ResNet101-12ch ResNet101 Inception V3 DPN-92 InceptionRes V2 Expert
  18. 38.

    Ξϯαϯϒϧ ॏΈ෇͖ฏۉ ςετࣄྫ਺͕े෼ʹେ͖ͷͰɺ1VCMJD-#Λ֬ೝͭͭ͠ௐ੔ɻ      Ξϯαϯϒϧ1VCMJD-# Ґ૬౰

    1VCMJD-# Ґ૬౰ 1VCMJD-# Ґ૬౰ 1VCMJD-# Ґ૬౰ 1VCMJD-# Ґ૬౰ 1VCMJD-# Ґ૬౰ ResNet101-12ch ResNet101 Inception V3 InceptionRes V2 DPN-92 Expert
  19. 41.

    UIϞσϧͷΞϯαϯϒϧ Ћ  Ґ૬౰ ˠ Ґ૬౰ ʴΫϥε෼෍͔Βܭࢉͨ͠ϒϥοΫϦετͷώϡʔϦεςΟΫεΛద༻ (1) 訓練事例例の
 正例例の割合

    (2) Softmax (1), (2) の重み付き平均とみなす ಉҰ.%ͷ܇࿅ࣄྫΛ࢖͍ɺ$//ϞσϧͷTPGUNBYΛิਖ਼͢Δɿ
  20. 42.

    UIಛ௃நग़ͯ͠ू໿ˍߋʹֶश ϕʔεϞσϧͷCPUUMFOFDLCMPDLͷग़ྗΛ࢖ͬͯ঎඼୯Ґʹू໿ˍֶ शɻ0$3 &"45<> $3//<> ͯ͠ςΩετ৘ใΛநग़ɻ ͜ΕΒΛೖྗͱͯ͠.-1 3// /FU7MBE<>ͳͲͷτοϓϨϕϧϞ σϧΛֶशɻ࠷ޙʹϕʔεϞσϧͱτοϓϨϕϧϞσϧͰΞϯαϯϒϧɻ

    w ϕʔεϞσϧ͸*ODFQUJPO3FTOFUW 3FTOFU 4&*ODFQUJPO7 9DFQUJPO w ̐ຕͷը૾Λೖྗͯ͠ಘͨ#PUUMFOFDLGFBUVSFTΛ঎඼͝ͱʹάϧʔϓͯ͠
 τοϓϨϕϧϞσϧͷೖྗͱ͢Δ w /FU7MBEͷ1PPMJOHֶशͱ(BUJOH͸ɺաڈίϯϖl:PVUVCF.,BHHMF -BSHF4DBMF7JEFPVOEFSTUBEJOHDIBMMFOHFzͷղ๏Ͱ΋࢖ΘΕɺ-061&ͱ͠ ͯπʔϧϘοΫε͕ެ։͞Ε͍ͯΔ ग़యɿIUUQTXXXLBHHMFDPNDDEJTDPVOUJNBHFDMBTTJpDBUJPODIBMMFOHFEJTDVTTJPO
  21. 50.

    TU0$3ͯ͠NVMUJJOQVU$// $51/<>ͰςΩετྖҬͷநग़ɺ$3//<>ͰςΩετͷநग़Λߦ͏ɻ.VMUJ JOQVU$//Ͱ$3//GFBUVSFTͱ3FT/FUͷ'$GFBUVSFTΛ࿈݁͢Δɻ 
 தؒϑΝΠϧ͸VJOUʹྔࢠԽʢ ͯ͠DBTUʣͯ͠TQBSTFNBUSJYͱͯ͠
 อଘ͢Δɻ .#͕.#ҎԼʹͳΔɻ $51/ $3//

    ରԠ͢Δຕ਺ͷ
 ઐ༻3FT/FU .VMUJJOQVU $// ༧ଌ ೖྗը૾ <>%FUFDUJOH5FYUJO/BUVSBM*NBHFXJUI$POOFDUJPOJTU5FYU1SPQPTBM/FUXPSL
 <>"O&OEUP&OE5SBJOBCMF/FVSBM/FUXPSLGPS*NBHFCBTFE4FRVFODF3FDPHOJUJPOBOE*UT"QQMJDBUJPOUP4DFOF5FYU3FDPHOJUJPO
  22. 56.

    σʔλʹର͢ΔཧղΛਂΊΔ ߹੒։ޱϨʔμʔ͸ೖࣹ֯ʢ*ODJEFODFBOHMFʣΛௐ੔ͯ͠؍ଌ͢ΔͷͰɺ
 ΄΅ಉ஋ͷάϧʔϓ͕Ͱ͖Δͱߟ͑Δ͜ͱ͕ग़དྷΔɻ 1 2 3 4 5 6 ic=44.1923

    ic=44.1925 ic=44.1919 ic=41.4321 ic=41.4315 ic=41.4317 ग़యʢࠨʣɿIUUQXXXBTDDTBHDDBFOHTBUFMMJUFTSBEBSTBUDPNQPOFOUTBTQ
 ʢӈʣ*NBHFSZ˜*#$"0 -BOETBU$PQFSOJDVT .BQEBUB˜(PPHMF ʢ˞ૉਓʹΑΔߟ࡯ͳͷͰɺؒҧ͍͕͋Δ͔΋͠Ε·ͤΜʣ
  23. 57.

    σʔληοτը૾͸͋Β͔͡ΊDSPQͨ͠΋ͷ $SPQલͷΠϝʔδਤ ༩͑ΒΕͨը૾ ic=44.1923 ic=44.1925 ic=44.1919 ic=41.4321 ic=41.4315 ic=41.4317 ߹੒։ޱϨʔμʔͷ؍ଌྖҬͷΠϝʔδਤ

    ೖࣹ֯ ͜ͷྖҬ͸͢΂ͯ ೖࣹ֯ Ծઆɿೖࣹ֯ͷগ਺දܻࣔ͢΂͕ͯ׬શҰக͢Δ৔߹ɺ ݩʑ͕ಉ͡؍ଌྖҬ˙͔Β੾ΓऔͬͨDSPQ͞Εͨը૾˙Ͱ༩͑ΒΕΔ ༩͑ΒΕͨ؍ଌσʔλ͸ɺધഫ͋Δ͍͸ණࢁ͕த৺Β͖͠΋ͷ͕ ͍ࣸͬͯΔɻ͋Β͔͡ΊDSPQ͞Εͨը૾ɻ
  24. 63.

    ೖࣹ֯͝ͱͷσʔλਖ਼نԽͰվળ ೖࣹ֯Λ࢖ͬͨσʔλਖ਼نԽΛߦͱ$7վળ %BUB"VHNFOUBUJPO SPU qJQ [PPN ΋վળʹ༗ޮͰ͋ͬͨ 前処理理 CV Public

    LB 4L CNN Scaling 0.2212 4L CNN Norm+ Scaling 0.1824 4L CNN +DA Norm + Scaling 0.1615 1VCMJD-#͸֬ೝ͍ͯ͠ͳ͍
  25. 64.

    UIೖࣹ֯ͷ৘ใΛϞσϧʹ૊ΈࠐΉ ྔతม਺ ೖࣹ֯ ͷ৘ใΛϞσϧʹ૊ΈࠐΉͨΊʹɺ$//Ϟσ ϧͷ༧ଌ݁Ռͱྔతม਺ΛTUBDLJOHͨ͠ɻ ݸͷ$//Ϟσϧ ྔతม਺YͰ-JHIU(#.ͱEFFQCJU (#%5 Ͱֶशɻ࠷ޙʹॏΈ෇͖ฏۉɻ 前処理理

    CV Public LB 4L CNN Scaling 0.2212 4L CNN Norm+Scaling 0.1824 4L CNN +DA Norm+Scaling 0.1615 stage2 (LGBM) NA 0.0895 0.0873 stage3 (Average) NA 0.0852 1VCMJD-#OE1SJWBUF-#UI
  26. 65.

     ݸͷ//ϞσϧΛ࡞੒ɻGPME$7Ͱνϡʔχϯάɻ  $7Ͱֶशͨ͠ϞσϧͷNFBOQSFEJDUJPOΛಛ௃ྔͱ͢Δ ྔతม਺   ೖࣹ֯ͰάϧʔϓԽͯ͠ूܭ NFBO NFEJBO

    DPVOU ͯ͠ಛ௃ྔ࡞੒  ೖࣹ֯ͱ  ͷNFBOQSFEJDUJPOͰ,//SFHSFTTPSʢྔతม਺ʣ  -(#.Λ    ͷಛ௃ྔͰֶशɻ߹ܭͰͭͷಛ௃ྔɻ 4. KNNreg 5. LGBM 1. CNNs 2. Mean prediction 3. Agg.
 features ೖࣹ֯ ೖྗը૾  UIΑΓগͳ͍Ϟσϧ਺ ೖࣹ֯Lۙ๣ ग़యɿIUUQTXXXLBHHMFDPNDTUBUPJMJDFCFSHDMBTTJpFSDIBMMFOHFEJTDVTTJPOɻগ͠؆ུԽͯ͠঺հ
  27. 66.

     ݸͷ//ϞσϧΛ࡞੒ɻGPME$7Ͱνϡʔχϯάɻ  $7Ͱֶशͨ͠ϞσϧͷNFBOQSFEJDUJPOΛಛ௃ྔͱ͢Δ ྔతม਺   ೖࣹ֯ͰάϧʔϓԽͯ͠ूܭ NFBO NFEJBO

    DPVOU ͯ͠ಛ௃ྔ࡞੒  ೖࣹ֯ͱ  ͷNFBOQSFEJDUJPOͰ,//SFHSFTTPSʢྔతม਺ʣ  -(#.Λ    ͷಛ௃ྔͰֶशɻ߹ܭͰͭͷಛ௃ྔɻ 4. KNNreg 5. LGBM 1. CNNs 2. Mean prediction 3. Agg.
 features ೖࣹ֯ ೖྗը૾  UIΑΓগͳ͍Ϟσϧ਺ ೖࣹ֯Lۙ๣ ग़యɿIUUQTXXXLBHHMFDPNDTUBUPJMJDFCFSHDMBTTJpFSDIBMMFOHFEJTDVTTJPOɻগ͠؆ུԽͯ͠঺հ ˞֤ϝλಛ௃ྔ͸
 $7Ͱ෼ׂͯ͠ܭࢉ
  28. 70.

    ෇࿥ɿࢀߟจݙ  1. [Blum & Hardt '15] "The Ladder: A

    Reliable Leaderboard for Machine Learning Competitions", In Proc. of the ICML '15. https://arxiv.org/abs/ 1502.04585 2. [Hardt '17] "Climbing a shaky ladder: Better adaptive risk estimation", https://arxiv.org/abs/1706.02733 3. [Rendle '10] "Factorization machines", In Proc. of the ICDM ’10. 4. [Johnson & Zhang '11] "Learning Nonlinear Functions Using Regularized Greedy Forest", https://arxiv.org/abs/1109.0887 5. [Hu+ ’17] “Squeeze-and-Excitation Networks”, In Proc. of the CVPR ’17. https://arxiv.org/abs/1709.01507 6. [He+ ’09] “Single Image Haze Removal”, In Proc. of the CVPR ’09. http://kaiminghe.com/cvpr09/
  29. 71.

    ෇࿥ɿࢀߟจݙ  7. [Zhang+ ’17] "mixup: Beyond Empirical Risk Minimization",

    https:// arxiv.org/abs/1710.09412 8. [Miech+ '17] "Learnable pooling with Context Gating for video classification", https://arxiv.org/abs/1706.06905 9. [Tian+ '16] "Detecting Text in Natural Image with Connectionist Text Proposal Network”, In Proc. of ECCV ‘16 https://arxiv.org/abs/1609.03605 10. [Qi+ '17] "PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation”, In Proc. of the CVPR ‘17 https://arxiv.org/abs/ 1612.00593 11. [Ma+ '17] "Multi-View Deep Learning for Consistent Semantic Mapping with RGB-D Cameras”, In Proc. of the IROS ‘17 https://arxiv.org/abs/ 1703.08866 12. [Shi+ '15] "An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition”, https://arxiv.org/abs/1507.05717
  30. 72.

    ෇࿥ɿࢀߟจݙ  13. [Arandjelović+ ’16] "NetVLAD: CNN architecture for weakly

    supervised place recognition”, In Proc. of the CVPR ’16 https://arxiv.org/abs/ 1511.07247 14. [Zhou+ '17] "EAST: An Efficient and Accurate Scene Text Detector", In Proc. of the CVPR’17 https://arxiv.org/abs/1704.03155v2