Save 37% off PRO during our Black Friday Sale! »

AAAI・WSDM 2017論文読み会

Dce743dc2547b5e2ca82c5381f504e86?s=47 KenjiH
March 27, 2017

AAAI・WSDM 2017論文読み会

A Concise Integer Linear Programming Formulation for Implicit Search Result Diversificationの紹介

Dce743dc2547b5e2ca82c5381f504e86?s=128

KenjiH

March 27, 2017
Tweet

Transcript

  1. """*ɾ84%.࿦จಡΈձ "$PODJTF*OUFHFS-JOFBS1SPHSBNNJOH'PSNVMBUJPO GPS*NQMJDJU4FBSDI3FTVMU%JWFSTJpDBUJPO Recruit Communications Co., Ltd. Kenji Hayashida

  2. 4&-'*/530%6$5*0/ ̶Recruit Communications Co., Ltd.̶ ໊લྛాݡೋ ৬ྺ֎ࢿܥ4*FS3$0ΤϯδχΞ ۀ຿಺༰ΞυςΫͷ࠷దԽɺίϯςϯπϚʔέɺσʔλύΠϓϥΠϯ झຯϓϩίϯ

  3. 84%. ̶Recruit Communications Co., Ltd.̶

  4. 84%. ̶Recruit Communications Co., Ltd.̶ ৄࡉ͸3$0ΞυςΫ-BCϒϩάͰʂ ʮXTENSDPϒϩάʯͰݕࡧ

  5. */530%6$5*0/ Ϣʔβ͕ٻΊ͍ͯΔ৘ใΛఏڙ͢Δͷ͸೉͍͠ ̶Recruit Communications Co., Ltd.̶ )BSSZ1PUUFS ຊ  өը

     ϨϏϡʔ  ΩϟϥΫλʔ  ʜ
  6. */530%6$5*0/ ̶Recruit Communications Co., Ltd.̶ 43% 4FBSDI3FTVMU%JWFSTJpDBUJPO Ϣʔβ͕ཉ͍͠৘ใΛΧόʔ͢ΔΑ͏ʹଟ༷ͳ݁ՌηοτΛฦ͢ FYQMJDJU43%ɹΫΤϦʹજࡏ͢ΔαϒτϐοΫ͕Θ͔͍ͬͯΔ৔߹ JNQMJDJU43%ɹΫΤϦʹજࡏ͢ΔαϒτϐοΫ͕Θ͔͍ͬͯͳ͍৔߹

    ଟ͘ͷ৔߹ɺજࡏαϒτϐοΫ͸Θ͔Βͳ͍ ࠓճ঺հ͢Δ࿦จ͸ɺJNQMJDJU43%ʹϑΥʔΧε
  7. */530%6$5*0/ ̶Recruit Communications Co., Ltd.̶ .BYJNBM.BSHJOBM3FMFWBODF ..3 ..3͸࠷΋୯७ͳख๏ɻ ଞʹ΋༷ʑͳख๏͕ଘࡏ͠ɺओʹҎԼͷ఺͕ҟͳΔɻ 

    ଟ༷ੑΛͲͷΑ͏ʹදݱ͢Δ͔ʁ  ద߹ੑͱଟ༷ੑΛͲͷΑ͏ʹόϥϯε͢Δ͔ʁ  ݁ՌϦετΛͲͷΑ͏ʹੜ੒͢Δ͔ʁ MMR = arg max Di 2R\S  Sim1(Di, Q) (1 ) max Dj 2S Sim2(Di, Dj) ҎԼͷج४ͰυΩϡϝϯτΛHSFFEZʹબ୒͢Δ ʢ$BSCPOFMMBOE(PMETUFJO ʣ
  8. 3&-"5&%803, ̶Recruit Communications Co., Ltd.̶ "⒏OJUZ1SPQFSHBUJPOʢ"1ʣ  ඪຊʢ୅දతͳΞΠςϜʣΛબͿ͜ͱͰΫϥελϦϯάΛߦ͏ख๏  ը૾෼ྨɺը૾ηάϝϯςʔγϣϯͳͲͰར༻͞ΕΔ

    U = Uij = ΞΠςϜJͱΞΠςϜKͷྨࣅ౓ ΞΠςϜJ͕ඪຊʹબ͹ΕΔ֬཰ (i = j) (i 6= j) \ ඇඪຊσʔλͱͦΕ͕ଐ͢ΔΫϥελͷඪຊσʔλͷྨࣅ౓ͷ࿨Λ࠷େԽ
  9. 3&-"5&%803, ̶Recruit Communications Co., Ltd.̶ %FTJSBCMF'BDJMJUZ1MBDFNFOU %'1  ࢪઃ഑ஔ໰୊ΛԠ༻ͨ͠Ϟσϧʢ;VDDPOFUBM ʣ

     ϝλώϡʔϦεςΟοΫͳΞϓϩʔνʢIJMMDMJNCJOHʣͰܭࢉ  ຊݚڀͷCBTFMJOFϞσϧͷͭ
  10. 130104&%.&5)0% ̶Recruit Communications Co., Ltd.̶ *-1*%ʢJOUFHFSMJOFBSQSPHSBNNJOHBQQSPBDIGPSJNQMJDJU43%ʣ

  11. 130104&%.&5)0% ̶Recruit Communications Co., Ltd.̶ *-1*% YJJυΩϡϝϯτEJ͕ඪຊʹબ͹Ε͔ͨʁ YJK J㱠KυΩϡϝϯτEJ͕EKΛඪຊʹબΜ͔ͩʁ

  12. 130104&%.&5)0% ̶Recruit Communications Co., Ltd.̶ *-1*% ద߹ੑΛ࠷େԽ

  13. 130104&%.&5)0% ̶Recruit Communications Co., Ltd.̶ *-1*% ୅දੑΛ࠷େԽ

  14. 130104&%.&5)0% ̶Recruit Communications Co., Ltd.̶ *-1*% NυΩϡϝϯταΠζͱLඪຊαΠζͷେ͖͞Λௐ੔ʢNLʣ

  15. 130104&%.&5)0% ̶Recruit Communications Co., Ltd.̶ *-1*% ඪຊʹબ͹ΕΔυΩϡϝϯτ͸Lݸ

  16. 130104&%.&5)0% ̶Recruit Communications Co., Ltd.̶ *-1*% υΩϡϝϯτEJͷඪຊ͸͚ͭͩ

  17. 130104&%.&5)0% ̶Recruit Communications Co., Ltd.̶ *-1*% EJ͕EKΛඪຊʹબΜͩͳΒɺEK͸ඪຊυΩϡϝϯτͰ͋Δ

  18. 130104&%.&5)0% ̶Recruit Communications Co., Ltd.̶  %'1͸*-1*%ͷಛघέʔεͱߟ͑Δ͜ͱ͕Ͱ͖Δ  *-1*%Ͱදݱ͢Δ͜ͱͰ࠷ద஋ΛܭࢉՄೳ 

    ͦͷଞͷ*NQMJDJU43%ͷख๏ʢ..3 .51ͳͲʣ΋ಉ༷ʹ*-1*%ͰදݱՄೳ *-1*% %'1
  19. &91&3*.&/54 ̶Recruit Communications Co., Ltd.̶ 53&$8FC 5SBDL #. 43% *-1*%

     %'1   53&$8FC5SBDL  ΫΤϦ  GBDFUFERVFSZෳ਺ͷҟͳΔαϒτϐοΫΛ࣋ͭΫΤϦ  BNCJHVPVTRVFSZෳ਺ͷ·ͬͨ͘ҟͳΔղऍΛ࣋ͭΫΤϦ JOJUJBMSFUSJFWBM DIPPTF999EPDVNFOUT DIPPTFEPDVNFOUT
  20. &91&3*.&/54 ̶Recruit Communications Co., Ltd.̶  ͢΂ͯͷΫΤϦʹରͯ͠ɺ%*-1*%㱢%%'1  %'1Ͱ͸࠷ద஋͕͑ΒΕͯͳ͍͜ͱΛ࣮ূ 

    υΩϡϝϯταΠζ͕૿͑Δ΄Ͳࠩ͸ݦஶ ࠷దੑͷݕূ   ͓ࣜΑͼ  ࣜͷЕɺ  ͔ࣜΒLͷ܎਺Λআ͘ʢଟ༷ੑʹϑΥʔΧεʣ  ಉ͡ܗͷ࠷దԽ໰୊ʹͯ͠ɺ*-1*%ͱ%'1ͷ࠷ద஋Λൺֱ ˞3&'&3&/$&4<>ΑΓҾ༻
  21. &91&3*.&/54 ̶Recruit Communications Co., Ltd.̶ ֤Ϟσϧͷੑೳൺֱ  *-1*%͸ଞͷϕʔεϥΠϯख๏ͱൺ΂ͯੑೳ͕Α͍  ଟ͘ͷվળ͸౷ܭతʹ༗ҙ

    ˞3&'&3&/$&4<>ΑΓҾ༻
  22. &91&3*.&/54 ̶Recruit Communications Co., Ltd.̶ ֤Ϟσϧͷੑೳൺֱ  *-1*%͸ଞͷϕʔεϥΠϯख๏ͱൺ΂ͯੑೳ͕Α͍  ଟ͘ͷվળ͸౷ܭతʹ༗ҙ

    ˞3&'&3&/$&4<>ΑΓҾ༻
  23. &91&3*.&/54 ̶Recruit Communications Co., Ltd.̶ ܭࢉ଎౓  Ұൠʹ੔਺ઢܗܭը໰୊͸/1ࠔ೉Ͱ͋ Δ 

    ༷ʑͳ෼ࢬݶఆ๏ͷΞϧΰϦζϜ͕ఏҊ ͞Ε͍ͯΔ  ࠷৽ͷιϧόʔΛ࢖͑͹े෼ߴ଎ʹܭࢉ Մೳ ੔਺ઢܗܭըʹ͍ͭͯ  ಉ༷ͷఆࣜԽͰ͋Δ͕ɺܭࢉίετ͸ *-1*%ͷํ͕%'1ΑΓߴ͍  *-1*%Ͱ͸෼ࢬݶఆ๏Ͱ࠷ద஋ΛٻΊ ͍ͯΔ͜ͱΛߟྀ͢Ε͹ೲಘͰ͖Δ ܭࢉ଎౓ͷൺֱ ˞3&'&3&/$&4<>ΑΓҾ༻
  24. $0/$-64*0/4"/%'6563&5"4, ̶Recruit Communications Co., Ltd.̶  JNQMJDJU43%ͷ໰୊Λղͨ͘Ίͷ৽͍͠ϞσϧΛఏҊ  NݸͷυΩϡϝϯτ͔ΒLݸͷඪຊΛબͿॲཧͱͯ͠ఆࣜԽ 

    ੔਺ઢܗܭը໰୊ͱͯ͠ఆࣜԽ͢Δ͜ͱͰɺ໨తؔ਺ͷ࠷ద஋Λ ܭࢉՄೳ  JNQMJDJU43%ͱಉ༷ͷΞϓϦέʔγϣϯʢྫϨίϝϯμʔγες Ϝʣʹ΋৽͍͠ࢹ఺Λఏڙ  ύϥϝʔλLͷνϡʔχϯάʹ͍ͭͯௐ͕ࠪඞཁʢΫΤϦʹΑͬͯҟ ͳΔ͸ͣʣ  Ϋϥελ಺ͷυΩϡϝϯτͷ૬ؔɺΫϥελಉ࢜ͷ૬ؔͳͲΛߟྀ ͨ͠ϞσϦϯάͷվྑ
  25. "11&/%*953&$8FC5SBDL ̶Recruit Communications Co., Ltd.̶  53&$ʢ5FYU3&USJFWBM$POGFSFODFʣͷίϯϖ<>  8FCݕࡧٕज़ͷ୳ٻͱධՁΛߦ͏ 

    σʔληοτͷઆ໌<>  BNCJHVPVTRVFSZෳ਺ͷશ͘ผͷղऍ͕Ͱ͖ΔΫΤϦ  ྫUIFDVSSFOU  5IF$VSSFOUͱ͍͏ϥδΦ൪૊ͷϖʔδʁ  χϡʔδϟʔδʔͷ5IF$VSSFOUOFXTQBQFSʁ  αϯΞϯτχΦͷ5IF$VSSFOU.BHB[JOFʁ  GBDFUFERVFSZෳ਺ͷαϒτϐοΫΛ࣋ͭෆ໌ྎͳΫΤϦ  ྫWPMWP  7PMWPͷϗʔϜϖʔδʁ  7PMWP9$467ͷϨϏϡʔʁ  7PMWPͷσΟʔϥʔʁ
  26. "11&/%*9#. ̶Recruit Communications Co., Ltd.̶ score ( D, Q )

    = n X i=1 IDF( qi) f ( qi, D )( k1 + 1) f ( qi, D ) + k1(1 b + b |D| avgdl ) f(qi, D) avgdl IDF ( qi) = log N n ( qi) + 0 . 5 n ( qi) + 0 . 5 ୯ޠRJͷυΩϡϝϯτ%಺ͷUFSNGSFRVFODZ υΩϡϝϯτͷฏۉ୯ޠ਺ N n(qi) υΩϡϝϯτ਺ ୯ޠRJΛؚΉυΩϡϝϯτ਺
  27. "11&/%*9.15 ̶Recruit Communications Co., Ltd.̶  .PSEFSO1PSUGPMJP5IFSPZ  ద߹ੑͷظ଴஋ͱ෼ࢄ͓Αͼ͢Ͱʹબ୒͞ΕͨυΩϡϝϯτͱͷ૬ ؔΛߟྀͨ͠Ϟσϧ

     ҎԼͷ஋͕ߴ͍΋ͷΛHSFFEZʹબ୒͍ͯ͘͠ E(dk) w ⇢ik k υΩϡϝϯτͷظ଴ద߹ੑ υΩϡϝϯτͷద߹ੑͷඪ४ภࠩ ϥϯΫݻ༗ͷॏΈ υΩϡϝϯτEJͱELͷ૬ؔ܎਺
  28. "11&/%*9FYQFDUFEDBMM!L ̶Recruit Communications Co., Ltd.̶ ΞΠςϜTJͷTVCUPQJD෼෍ ΫΤϦͷTVCUPQJD෼෍ ద߹ੑʢద߹ɺෆద߹ʣ ҎԼͷ஋͕ߴ͍΋ͷΛHSFFEZʹબ୒͍ͯ͘͠ ˞3&'&3&/$&4<>ΑΓҾ༻

  29. "11&/%*9O&33*" ̶Recruit Communications Co., Ltd.̶ &YQFDUFE3FDJQSPDBM3BOL &33 ద߹Ϩϕϧ ήΠϯ 

         *OUFOU"XBSF&YQFDUFE3FDJQSPDBM3BOL &33*" ΫΤϦͷҙਤΛ࢖ͬͯظ଴஋Λܭࢉ SFDJQSPDBMSBOL S൪໨ͰॳΊͯҙਤJʹద߹͢ ΔυΩϡϝϯτ͕͑ΒΕͨ
  30. "11&/%*9ЋO%$( ̶Recruit Communications Co., Ltd.̶ OPSNBMJTFE%JTDPVOUFE$VNVMBUJWF(BJO O%$( g(r) g⇤(r) ϥϯΫSͷυΩϡϝϯτͷήΠϯʢద߹ϨϕϧʹԠܾͯ͡·Δʣ

    ཧ૝తʹϥϯΩϯά͞Εͨͱ͖ͷϥϯΫSͷυΩϡϝϯτͷήΠϯ ద߹Ϩϕϧ ήΠϯ       ЋO%$( ήΠϯͷ୅ΘΓʹOPWFMUZCJBTFEήΠϯΛ༻͍Δ ϥϯΫSͷυΩϡϝϯτ͸ҙਤJʹద߹͢Δ ϥϯΫSͷυΩϡϝϯτ͸ҙਤJʹద߹͠ͳ͍ ͨͩ͠ ΫΤϦ͕࣋ͭҙਤͷू߹ʹରͯ͠ద߹ੑΛܭࢉ ↵ ύϥϝʔλʢ͕Α͘࢖ΘΕΔʣ
  31. 3&'&3&/$&4 ̶Recruit Communications Co., Ltd.̶ <>l"$PODJTF*OUFHFS-JOFBS1SPHSBNNJOH'PSNVMBUJPOGPS*NQMJDJU4FBSDI 3FTVMU%JWFSTJpDBUJPOz )BJUBP:V "EBN+BUPXU 3PJ#MBODP

    )JEFP+PIP  +PFNPO+PTF -POH$IFO 'BKJF:VBO <>5FYU3&USJFWBM$POGFSFODF 53&$ )PNF1BHF <>l0WFSWJFXPGUIF53&$8FC5SBDLz $IBSMFT-"$MBSLF /JDL $SBTXFMM *BO4PCPSPG <>l)PX*OUVJUJWF"SF%JWFSTJpFE4FBSDI.FUSJDT $PODPSEBODF5FTU3FTVMUT GPSUIF%JWFSTJUZ6.FBTVSFTz 5FUTVZB4BLBJ <>%JWFSTFSFUSJFWBMWJBHSFFEZPQUJNJ[BUJPOPGFYQFDUFEDBMM!LJOBMBUFOU TVCUPQJDSFMFWBODFNPEFM 44BOOFS 4(VP 5(SBFQFM 4,IBSB[NJ 4 ,BSJNJ