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

自然科学研究の道具としての機械学習

itakigawa
September 27, 2023
20

 自然科学研究の道具としての機械学習

第6回情報科学系セミナー, 2019年12月4日, 北陸先端科学技術大学院大学 情報科学系研究棟.

itakigawa

September 27, 2023
Tweet

Transcript

  1. ࣗવՊֶݚڀͷಓ۩ͱͯ͠ͷػցֶश
    ୈ6ճ৘ใՊֶܥηϛφʔ

    2019೥12݄4೔
    ୍઒ Ұֶ (͖͕ͨΘɾ͍͕ͪ͘)
    • ཧԽֶݚڀॴ ֵ৽஌ೳ౷߹ݚڀηϯλʔ (AIP)

    iPSࡉ๔࿈ܞҩֶతϦεΫճආνʔϜ@ژ౎
    • ๺ւಓେֶ Խֶ൓Ԡ૑੒ݚڀڌ఺ (WPI-ICReDD)
    [email protected]

    View full-size slide

  2. ࣗݾ঺հɿ୍઒ Ұֶ(͖͕ͨΘɾ͍͕ͪ͘)
    2
    10೥ ๺େ
    (1995ʙ2004)
    7೥ ژେ
    (2005ʙ2011)
    7೥ ๺େ
    (2012ʙ2018)
    ౷ܭత৴߸ॲཧͱύλʔϯೝࣝ (޻ֶݚڀՊ)

    "ྼܾఆ৴߸ݯ෼཭ͷL1ϊϧϜ࠷খղͷཧ࿦෼ੳ"
    όΠΦΠϯϑΥϚςΟΫε (Խֶݚڀॴ)

    έϞΠϯϑΥϚςΟΫε (ༀֶݚڀՊ)
    σʔλۦಈՊֶɾ ཭ࢄߏ଄Λ൐͏ػցֶश

    (৘ใՊֶݚڀՊ)

    + JST͖͕͚͞: ࡐྉΠϯϑΥϚςΟΫε
    ʁ೥ ཧݚ(ژ౎)
    (2019ʙ)
    AIPηϯλʔ iPSࡉ๔࿈ܞҩֶతϦεΫճආνʔϜ

    (๺େ Խֶ൓Ԡ૑੒ݚڀڌ఺ͱΫϩΞϙ)
    ઐ໳ɿػցֶशɾσʔλϚΠχϯάͱͦͷՊֶͰͷར׆༻

    ɹɹʮσʔλ͔ΒͷֶशʯΛͲ͏໰୊ղܾʹ׆༻Ͱ͖Δͷ͔ʁ

    View full-size slide

  3. ࠷ۙͷݚڀର৅
    3
    v
    h(1)
    v
    =

    xv
    0
    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
    AAACxXichVFNS9xQFD1GW+3Y1lE3QjdDB4vdDDelYCkIgy506UdHBWOHvPgcH+aL5E1Qw+DeP+DClQWF0h/QH+DGhVsL/oTiUsFNF73JhJZWam9I3nnn3nNz3rsidFWsia56jN6+R4/7B56UBp8+ez5UHh5ZjoN25MiGE7hBtCrsWLrKlw2ttCtXw0jannDlitieyfIriYxiFfgf9G4o1z275atN5diaqWa5bgkv3eo0k4/phPm6U5mqWEK2lJ8Kz9aR2umUsoIdLrCsHBIz0t/4lW+Wq1SjPCr3gVmAKoqYD8pfYWEDARy04UHCh2bswkbMzxpMEELm1pEyFzFSeV6igxJr21wlucJmdpu/Ld6tFazP+6xnnKsd/ovLb8TKCsbpkj7TDZ3TF/pOP/7ZK817ZF52eRVdrQybQwdjS3f/VXm8amz9Vj3oWWMT73Kvir2HOZOdwunqk73Dm6X3i+PpK/pE1+z/mK7ojE/gJ7fOyYJcPHrAj2AvfGM8IPPvcdwHy29qJtXMhbfV+nQxqgG8wEtM8DwmUccc5tHg/qe4wCW+GbOGZ2gj6ZYaPYVmFH+Esf8T3nyqPg==
    h(t 1)
    v
    a(t)
    v
    h(t)
    v
    (36
    "⒏OF
    //
    //
    yv
    zv

    h(T )
    v
    xv

    *OJUJBMJ[BUJPO
    /PEF8JTF3FDVSSFOU6QEBUF
    (SBQI(BUIFSJOH
    tanh
    X
    v2V
    (yv) tanh(zv)
    !
    hG =
    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
    "EBQUJWF8FJHIUJOH

    4PGU"UUFOUJPO

    .FTTBHF
    6QEBUF 3FBEPVU
    ཭ࢄߏ଄
    ػցֶश
    σʔλ

    ۦಈՊֶ
    ʵߏ଄ͱ஌ࣝʵ
    ʵֶशͱ஌ೳʵ
    ʵ৘ใͱՊֶʵ
    ཭ࢄߏ଄ͷදݱͱߏ੒๏ ཭ࢄߏ଄Λೖྗɾ੍໿ͱ͢Δػցֶश
    ໦ߏ଄Ξϯαϯϒϧ ਂ૚ֶश/ܭࢉάϥϑ ֬཰తϓϩάϥϛϯά
    LightGBM (Microsoft)
    See5/C5.0 & Cubist (RuleQuest)
    CART® MARS® TreeNet®

    Random Forests® (Salford Systems)
    CatBoost (Yandex)
    TFBoost (Google)
    TenscentBoost (Tenscent)
    Sherwood decision forests

    TM
    AAAB+nicbVC7SgNBFL0bXzG+opY2g0GwCrsiaBm0sREi5gXJEmYns8mQeSwzs0JY8wm22tuJrT9j65c4SbbQ6IELh3Pu5VxOlHBmrO9/eoWV1bX1jeJmaWt7Z3evvH/QMirVhDaJ4kp3ImwoZ5I2LbOcdhJNsYg4bUfj65nffqDaMCUbdpLQUOChZDEj2DrpvtG/7ZcrftWfA/0lQU4qkKPeL3/1BoqkgkpLODamG/iJDTOsLSOcTku91NAEkzEe0q6jEgtqwmz+6hSdOGWAYqXdSIvm6s+LDAtjJiJymwLbkVn2ZuK/XiSWkm18GWZMJqmlkiyC45Qjq9CsBzRgmhLLJ45gopn7HZER1phY11bJlRIsV/CXtM6qgV8N7s4rtau8niIcwTGcQgAXUIMbqEMTCAzhCZ7hxXv0Xr03732xWvDym0P4Be/jG0AQlD4=
    AAAB+nicbVC7SgNBFL0bXzG+opY2g0GwCrsiaBm0sREi5gXJEmYns8mQeSwzs0JY8wm22tuJrT9j65c4SbbQ6IELh3Pu5VxOlHBmrO9/eoWV1bX1jeJmaWt7Z3evvH/QMirVhDaJ4kp3ImwoZ5I2LbOcdhJNsYg4bUfj65nffqDaMCUbdpLQUOChZDEj2DrpvtG/7ZcrftWfA/0lQU4qkKPeL3/1BoqkgkpLODamG/iJDTOsLSOcTku91NAEkzEe0q6jEgtqwmz+6hSdOGWAYqXdSIvm6s+LDAtjJiJymwLbkVn2ZuK/XiSWkm18GWZMJqmlkiyC45Qjq9CsBzRgmhLLJ45gopn7HZER1phY11bJlRIsV/CXtM6qgV8N7s4rtau8niIcwTGcQgAXUIMbqEMTCAzhCZ7hxXv0Xr03732xWvDym0P4Be/jG0AQlD4=
    AAAB+nicbVC7SgNBFL0bXzG+opY2g0GwCrsiaBm0sREi5gXJEmYns8mQeSwzs0JY8wm22tuJrT9j65c4SbbQ6IELh3Pu5VxOlHBmrO9/eoWV1bX1jeJmaWt7Z3evvH/QMirVhDaJ4kp3ImwoZ5I2LbOcdhJNsYg4bUfj65nffqDaMCUbdpLQUOChZDEj2DrpvtG/7ZcrftWfA/0lQU4qkKPeL3/1BoqkgkpLODamG/iJDTOsLSOcTku91NAEkzEe0q6jEgtqwmz+6hSdOGWAYqXdSIvm6s+LDAtjJiJymwLbkVn2ZuK/XiSWkm18GWZMJqmlkiyC45Qjq9CsBzRgmhLLJ45gopn7HZER1phY11bJlRIsV/CXtM6qgV8N7s4rtau8niIcwTGcQgAXUIMbqEMTCAzhCZ7hxXv0Xr03732xWvDym0P4Be/jG0AQlD4=
    AAAB+nicbVC7SgNBFL0bXzG+opY2g0GwCrsiaBm0sREi5gXJEmYns8mQeSwzs0JY8wm22tuJrT9j65c4SbbQ6IELh3Pu5VxOlHBmrO9/eoWV1bX1jeJmaWt7Z3evvH/QMirVhDaJ4kp3ImwoZ5I2LbOcdhJNsYg4bUfj65nffqDaMCUbdpLQUOChZDEj2DrpvtG/7ZcrftWfA/0lQU4qkKPeL3/1BoqkgkpLODamG/iJDTOsLSOcTku91NAEkzEe0q6jEgtqwmz+6hSdOGWAYqXdSIvm6s+LDAtjJiJymwLbkVn2ZuK/XiSWkm18GWZMJqmlkiyC45Qjq9CsBzRgmhLLJ45gopn7HZER1phY11bJlRIsV/CXtM6qgV8N7s4rtau8niIcwTGcQgAXUIMbqEMTCAzhCZ7hxXv0Xr03732xWvDym0P4Be/jG0AQlD4=
    FAB/HME (NEC)
    n0
    AAAB+nicbVA9SwNBFHwXv2L8ilraLAbBKtyJoGXQxjKiiYHkCHubd8mS3b1jd08IMT/BVns7sfXP2PpL3CRXaOLAg2HmPeYxUSq4sb7/5RVWVtfWN4qbpa3tnd298v5B0ySZZthgiUh0K6IGBVfYsNwKbKUaqYwEPkTD66n/8Ija8ETd21GKoaR9xWPOqHXSner63XLFr/ozkGUS5KQCOerd8nenl7BMorJMUGPagZ/acEy15UzgpNTJDKaUDWkf244qKtGE49mrE3LilB6JE+1GWTJTf1+MqTRmJCO3KakdmEVvKv7rRXIh2caX4ZirNLOo2Dw4zgSxCZn2QHpcI7Ni5AhlmrvfCRtQTZl1bZVcKcFiBcukeVYN/Gpwe16pXeX1FOEIjuEUAriAGtxAHRrAoA/P8AKv3pP35r17H/PVgpffHMIfeJ8/O4iUOw==
    AAAB+nicbVA9SwNBFHwXv2L8ilraLAbBKtyJoGXQxjKiiYHkCHubd8mS3b1jd08IMT/BVns7sfXP2PpL3CRXaOLAg2HmPeYxUSq4sb7/5RVWVtfWN4qbpa3tnd298v5B0ySZZthgiUh0K6IGBVfYsNwKbKUaqYwEPkTD66n/8Ija8ETd21GKoaR9xWPOqHXSner63XLFr/ozkGUS5KQCOerd8nenl7BMorJMUGPagZ/acEy15UzgpNTJDKaUDWkf244qKtGE49mrE3LilB6JE+1GWTJTf1+MqTRmJCO3KakdmEVvKv7rRXIh2caX4ZirNLOo2Dw4zgSxCZn2QHpcI7Ni5AhlmrvfCRtQTZl1bZVcKcFiBcukeVYN/Gpwe16pXeX1FOEIjuEUAriAGtxAHRrAoA/P8AKv3pP35r17H/PVgpffHMIfeJ8/O4iUOw==
    AAAB+nicbVA9SwNBFHwXv2L8ilraLAbBKtyJoGXQxjKiiYHkCHubd8mS3b1jd08IMT/BVns7sfXP2PpL3CRXaOLAg2HmPeYxUSq4sb7/5RVWVtfWN4qbpa3tnd298v5B0ySZZthgiUh0K6IGBVfYsNwKbKUaqYwEPkTD66n/8Ija8ETd21GKoaR9xWPOqHXSner63XLFr/ozkGUS5KQCOerd8nenl7BMorJMUGPagZ/acEy15UzgpNTJDKaUDWkf244qKtGE49mrE3LilB6JE+1GWTJTf1+MqTRmJCO3KakdmEVvKv7rRXIh2caX4ZirNLOo2Dw4zgSxCZn2QHpcI7Ni5AhlmrvfCRtQTZl1bZVcKcFiBcukeVYN/Gpwe16pXeX1FOEIjuEUAriAGtxAHRrAoA/P8AKv3pP35r17H/PVgpffHMIfeJ8/O4iUOw==
    AAAB+nicbVA9SwNBFHwXv2L8ilraLAbBKtyJoGXQxjKiiYHkCHubd8mS3b1jd08IMT/BVns7sfXP2PpL3CRXaOLAg2HmPeYxUSq4sb7/5RVWVtfWN4qbpa3tnd298v5B0ySZZthgiUh0K6IGBVfYsNwKbKUaqYwEPkTD66n/8Ija8ETd21GKoaR9xWPOqHXSner63XLFr/ozkGUS5KQCOerd8nenl7BMorJMUGPagZ/acEy15UzgpNTJDKaUDWkf244qKtGE49mrE3LilB6JE+1GWTJTf1+MqTRmJCO3KakdmEVvKv7rRXIh2caX4ZirNLOo2Dw4zgSxCZn2QHpcI7Ni5AhlmrvfCRtQTZl1bZVcKcFiBcukeVYN/Gpwe16pXeX1FOEIjuEUAriAGtxAHRrAoA/P8AKv3pP35r17H/PVgpffHMIfeJ8/O4iUOw==
    `5
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyKosegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL2LXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d16pXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AI70lYk=
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyKosegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL2LXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d16pXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AI70lYk=
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyKosegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL2LXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d16pXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AI70lYk=
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyKosegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL2LXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d16pXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AI70lYk=
    `6
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyKqMegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL2LXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d16pXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AJCIlYo=
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyKqMegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL2LXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d16pXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AJCIlYo=
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyKqMegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL2LXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d16pXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AJCIlYo=
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyKqMegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL2LXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d16pXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AJCIlYo=
    n4
    AAAB+nicbVDLSgMxFL2pr1pfVZdugkVwVWakoMuiG5cV7QPaoWTSTBuaZIYkI5Sxn+BW9+7ErT/j1i8xbWehrQcuHM65l3M5YSK4sZ73hQpr6xubW8Xt0s7u3v5B+fCoZeJUU9aksYh1JySGCa5Y03IrWCfRjMhQsHY4vpn57UemDY/Vg50kLJBkqHjEKbFOulf9Wr9c8areHHiV+DmpQI5Gv/zdG8Q0lUxZKogxXd9LbJARbTkVbFrqpYYlhI7JkHUdVUQyE2TzV6f4zCkDHMXajbJ4rv6+yIg0ZiJDtymJHZllbyb+64VyKdlGV0HGVZJapugiOEoFtjGe9YAHXDNqxcQRQjV3v2M6IppQ69oquVL85QpWSeui6ntV/65WqV/n9RThBE7hHHy4hDrcQgOaQGEIz/ACr+gJvaF39LFYLaD85hj+AH3+AEHYlD8=
    AAAB+nicbVDLSgMxFL2pr1pfVZdugkVwVWakoMuiG5cV7QPaoWTSTBuaZIYkI5Sxn+BW9+7ErT/j1i8xbWehrQcuHM65l3M5YSK4sZ73hQpr6xubW8Xt0s7u3v5B+fCoZeJUU9aksYh1JySGCa5Y03IrWCfRjMhQsHY4vpn57UemDY/Vg50kLJBkqHjEKbFOulf9Wr9c8areHHiV+DmpQI5Gv/zdG8Q0lUxZKogxXd9LbJARbTkVbFrqpYYlhI7JkHUdVUQyE2TzV6f4zCkDHMXajbJ4rv6+yIg0ZiJDtymJHZllbyb+64VyKdlGV0HGVZJapugiOEoFtjGe9YAHXDNqxcQRQjV3v2M6IppQ69oquVL85QpWSeui6ntV/65WqV/n9RThBE7hHHy4hDrcQgOaQGEIz/ACr+gJvaF39LFYLaD85hj+AH3+AEHYlD8=
    AAAB+nicbVDLSgMxFL2pr1pfVZdugkVwVWakoMuiG5cV7QPaoWTSTBuaZIYkI5Sxn+BW9+7ErT/j1i8xbWehrQcuHM65l3M5YSK4sZ73hQpr6xubW8Xt0s7u3v5B+fCoZeJUU9aksYh1JySGCa5Y03IrWCfRjMhQsHY4vpn57UemDY/Vg50kLJBkqHjEKbFOulf9Wr9c8areHHiV+DmpQI5Gv/zdG8Q0lUxZKogxXd9LbJARbTkVbFrqpYYlhI7JkHUdVUQyE2TzV6f4zCkDHMXajbJ4rv6+yIg0ZiJDtymJHZllbyb+64VyKdlGV0HGVZJapugiOEoFtjGe9YAHXDNqxcQRQjV3v2M6IppQ69oquVL85QpWSeui6ntV/65WqV/n9RThBE7hHHy4hDrcQgOaQGEIz/ACr+gJvaF39LFYLaD85hj+AH3+AEHYlD8=
    AAAB+nicbVDLSgMxFL2pr1pfVZdugkVwVWakoMuiG5cV7QPaoWTSTBuaZIYkI5Sxn+BW9+7ErT/j1i8xbWehrQcuHM65l3M5YSK4sZ73hQpr6xubW8Xt0s7u3v5B+fCoZeJUU9aksYh1JySGCa5Y03IrWCfRjMhQsHY4vpn57UemDY/Vg50kLJBkqHjEKbFOulf9Wr9c8areHHiV+DmpQI5Gv/zdG8Q0lUxZKogxXd9LbJARbTkVbFrqpYYlhI7JkHUdVUQyE2TzV6f4zCkDHMXajbJ4rv6+yIg0ZiJDtymJHZllbyb+64VyKdlGV0HGVZJapugiOEoFtjGe9YAHXDNqxcQRQjV3v2M6IppQ69oquVL85QpWSeui6ntV/65WqV/n9RThBE7hHHy4hDrcQgOaQGEIz/ACr+gJvaF39LFYLaD85hj+AH3+AEHYlD8=
    n1
    AAAB+nicbVA9SwNBFHwXv2L8ilraLAbBKtyJoGXQxjKiiYHkCHubvWTJ7t6x+04IMT/BVns7sfXP2PpL3CRXaOLAg2HmPeYxUSqFRd//8gorq2vrG8XN0tb2zu5eef+gaZPMMN5giUxMK6KWS6F5AwVK3koNpyqS/CEaXk/9h0durEj0PY5SHira1yIWjKKT7nQ36JYrftWfgSyTICcVyFHvlr87vYRlimtkklrbDvwUwzE1KJjkk1InszylbEj7vO2oporbcDx7dUJOnNIjcWLcaCQz9ffFmCprRypym4riwC56U/FfL1ILyRhfhmOh0wy5ZvPgOJMEEzLtgfSE4QzlyBHKjHC/EzaghjJ0bZVcKcFiBcukeVYN/Gpwe16pXeX1FOEIjuEUAriAGtxAHRrAoA/P8AKv3pP35r17H/PVgpffHMIfeJ8/PRyUPA==
    AAAB+nicbVA9SwNBFHwXv2L8ilraLAbBKtyJoGXQxjKiiYHkCHubvWTJ7t6x+04IMT/BVns7sfXP2PpL3CRXaOLAg2HmPeYxUSqFRd//8gorq2vrG8XN0tb2zu5eef+gaZPMMN5giUxMK6KWS6F5AwVK3koNpyqS/CEaXk/9h0durEj0PY5SHira1yIWjKKT7nQ36JYrftWfgSyTICcVyFHvlr87vYRlimtkklrbDvwUwzE1KJjkk1InszylbEj7vO2oporbcDx7dUJOnNIjcWLcaCQz9ffFmCprRypym4riwC56U/FfL1ILyRhfhmOh0wy5ZvPgOJMEEzLtgfSE4QzlyBHKjHC/EzaghjJ0bZVcKcFiBcukeVYN/Gpwe16pXeX1FOEIjuEUAriAGtxAHRrAoA/P8AKv3pP35r17H/PVgpffHMIfeJ8/PRyUPA==
    AAAB+nicbVA9SwNBFHwXv2L8ilraLAbBKtyJoGXQxjKiiYHkCHubvWTJ7t6x+04IMT/BVns7sfXP2PpL3CRXaOLAg2HmPeYxUSqFRd//8gorq2vrG8XN0tb2zu5eef+gaZPMMN5giUxMK6KWS6F5AwVK3koNpyqS/CEaXk/9h0durEj0PY5SHira1yIWjKKT7nQ36JYrftWfgSyTICcVyFHvlr87vYRlimtkklrbDvwUwzE1KJjkk1InszylbEj7vO2oporbcDx7dUJOnNIjcWLcaCQz9ffFmCprRypym4riwC56U/FfL1ILyRhfhmOh0wy5ZvPgOJMEEzLtgfSE4QzlyBHKjHC/EzaghjJ0bZVcKcFiBcukeVYN/Gpwe16pXeX1FOEIjuEUAriAGtxAHRrAoA/P8AKv3pP35r17H/PVgpffHMIfeJ8/PRyUPA==
    AAAB+nicbVA9SwNBFHwXv2L8ilraLAbBKtyJoGXQxjKiiYHkCHubvWTJ7t6x+04IMT/BVns7sfXP2PpL3CRXaOLAg2HmPeYxUSqFRd//8gorq2vrG8XN0tb2zu5eef+gaZPMMN5giUxMK6KWS6F5AwVK3koNpyqS/CEaXk/9h0durEj0PY5SHira1yIWjKKT7nQ36JYrftWfgSyTICcVyFHvlr87vYRlimtkklrbDvwUwzE1KJjkk1InszylbEj7vO2oporbcDx7dUJOnNIjcWLcaCQz9ffFmCprRypym4riwC56U/FfL1ILyRhfhmOh0wy5ZvPgOJMEEzLtgfSE4QzlyBHKjHC/EzaghjJ0bZVcKcFiBcukeVYN/Gpwe16pXeX1FOEIjuEUAriAGtxAHRrAoA/P8AKv3pP35r17H/PVgpffHMIfeJ8/PRyUPA==
    `1
    AAAB/XicbVDLSgNBEOz1GeMr6tHLYBA8hV0R9Bj04jGCeUCyhNlJJxkzM7vMzAphCf6CV717E69+i1e/xEmyB00saCiquqmmokRwY33/y1tZXVvf2CxsFbd3dvf2SweHDROnmmGdxSLWrYgaFFxh3XIrsJVopDIS2IxGN1O/+Yja8Fjd23GCoaQDxfucUeukRgeF6AbdUtmv+DOQZRLkpAw5at3Sd6cXs1SiskxQY9qBn9gwo9pyJnBS7KQGE8pGdIBtRxWVaMJs9u2EnDqlR/qxdqMsmam/LzIqjRnLyG1Kaodm0ZuK/3qRXEi2/asw4ypJLSo2D+6ngtiYTKsgPa6RWTF2hDLN3e+EDammzLrCiq6UYLGCZdI4rwR+Jbi7KFev83oKcAwncAYBXEIVbqEGdWDwAM/wAq/ek/fmvXsf89UVL785gj/wPn8AiKSVhQ==
    AAAB/XicbVDLSgNBEOz1GeMr6tHLYBA8hV0R9Bj04jGCeUCyhNlJJxkzM7vMzAphCf6CV717E69+i1e/xEmyB00saCiquqmmokRwY33/y1tZXVvf2CxsFbd3dvf2SweHDROnmmGdxSLWrYgaFFxh3XIrsJVopDIS2IxGN1O/+Yja8Fjd23GCoaQDxfucUeukRgeF6AbdUtmv+DOQZRLkpAw5at3Sd6cXs1SiskxQY9qBn9gwo9pyJnBS7KQGE8pGdIBtRxWVaMJs9u2EnDqlR/qxdqMsmam/LzIqjRnLyG1Kaodm0ZuK/3qRXEi2/asw4ypJLSo2D+6ngtiYTKsgPa6RWTF2hDLN3e+EDammzLrCiq6UYLGCZdI4rwR+Jbi7KFev83oKcAwncAYBXEIVbqEGdWDwAM/wAq/ek/fmvXsf89UVL785gj/wPn8AiKSVhQ==
    AAAB/XicbVDLSgNBEOz1GeMr6tHLYBA8hV0R9Bj04jGCeUCyhNlJJxkzM7vMzAphCf6CV717E69+i1e/xEmyB00saCiquqmmokRwY33/y1tZXVvf2CxsFbd3dvf2SweHDROnmmGdxSLWrYgaFFxh3XIrsJVopDIS2IxGN1O/+Yja8Fjd23GCoaQDxfucUeukRgeF6AbdUtmv+DOQZRLkpAw5at3Sd6cXs1SiskxQY9qBn9gwo9pyJnBS7KQGE8pGdIBtRxWVaMJs9u2EnDqlR/qxdqMsmam/LzIqjRnLyG1Kaodm0ZuK/3qRXEi2/asw4ypJLSo2D+6ngtiYTKsgPa6RWTF2hDLN3e+EDammzLrCiq6UYLGCZdI4rwR+Jbi7KFev83oKcAwncAYBXEIVbqEGdWDwAM/wAq/ek/fmvXsf89UVL785gj/wPn8AiKSVhQ==
    AAAB/XicbVDLSgNBEOz1GeMr6tHLYBA8hV0R9Bj04jGCeUCyhNlJJxkzM7vMzAphCf6CV717E69+i1e/xEmyB00saCiquqmmokRwY33/y1tZXVvf2CxsFbd3dvf2SweHDROnmmGdxSLWrYgaFFxh3XIrsJVopDIS2IxGN1O/+Yja8Fjd23GCoaQDxfucUeukRgeF6AbdUtmv+DOQZRLkpAw5at3Sd6cXs1SiskxQY9qBn9gwo9pyJnBS7KQGE8pGdIBtRxWVaMJs9u2EnDqlR/qxdqMsmam/LzIqjRnLyG1Kaodm0ZuK/3qRXEi2/asw4ypJLSo2D+6ngtiYTKsgPa6RWTF2hDLN3e+EDammzLrCiq6UYLGCZdI4rwR+Jbi7KFev83oKcAwncAYBXEIVbqEGdWDwAM/wAq/ek/fmvXsf89UVL785gj/wPn8AiKSVhQ==
    `2
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKewGQY9BLx4jmAckS5id9CZjZmeXmVkhLMFf8Kp3b+LVb/HqlzhJ9qCJBQ1FVTfVVJAIro3rfjmFtfWNza3idmlnd2//oHx41NJxqhg2WSxi1QmoRsElNg03AjuJQhoFAtvB+Gbmtx9RaR7LezNJ0I/oUPKQM2qs1OqhEP1av1xxq+4cZJV4OalAjka//N0bxCyNUBomqNZdz02Mn1FlOBM4LfVSjQllYzrErqWSRqj9bP7tlJxZZUDCWNmRhszV3xcZjbSeRIHdjKgZ6WVvJv7rBdFSsgmv/IzLJDUo2SI4TAUxMZlVQQZcITNiYgllitvfCRtRRZmxhZVsKd5yBaukVat6btW7u6jUr/N6inACp3AOHlxCHW6hAU1g8ADP8AKvzpPz5rw7H4vVgpPfHMMfOJ8/ijiVhg==
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKewGQY9BLx4jmAckS5id9CZjZmeXmVkhLMFf8Kp3b+LVb/HqlzhJ9qCJBQ1FVTfVVJAIro3rfjmFtfWNza3idmlnd2//oHx41NJxqhg2WSxi1QmoRsElNg03AjuJQhoFAtvB+Gbmtx9RaR7LezNJ0I/oUPKQM2qs1OqhEP1av1xxq+4cZJV4OalAjka//N0bxCyNUBomqNZdz02Mn1FlOBM4LfVSjQllYzrErqWSRqj9bP7tlJxZZUDCWNmRhszV3xcZjbSeRIHdjKgZ6WVvJv7rBdFSsgmv/IzLJDUo2SI4TAUxMZlVQQZcITNiYgllitvfCRtRRZmxhZVsKd5yBaukVat6btW7u6jUr/N6inACp3AOHlxCHW6hAU1g8ADP8AKvzpPz5rw7H4vVgpPfHMMfOJ8/ijiVhg==
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKewGQY9BLx4jmAckS5id9CZjZmeXmVkhLMFf8Kp3b+LVb/HqlzhJ9qCJBQ1FVTfVVJAIro3rfjmFtfWNza3idmlnd2//oHx41NJxqhg2WSxi1QmoRsElNg03AjuJQhoFAtvB+Gbmtx9RaR7LezNJ0I/oUPKQM2qs1OqhEP1av1xxq+4cZJV4OalAjka//N0bxCyNUBomqNZdz02Mn1FlOBM4LfVSjQllYzrErqWSRqj9bP7tlJxZZUDCWNmRhszV3xcZjbSeRIHdjKgZ6WVvJv7rBdFSsgmv/IzLJDUo2SI4TAUxMZlVQQZcITNiYgllitvfCRtRRZmxhZVsKd5yBaukVat6btW7u6jUr/N6inACp3AOHlxCHW6hAU1g8ADP8AKvzpPz5rw7H4vVgpPfHMMfOJ8/ijiVhg==
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKewGQY9BLx4jmAckS5id9CZjZmeXmVkhLMFf8Kp3b+LVb/HqlzhJ9qCJBQ1FVTfVVJAIro3rfjmFtfWNza3idmlnd2//oHx41NJxqhg2WSxi1QmoRsElNg03AjuJQhoFAtvB+Gbmtx9RaR7LezNJ0I/oUPKQM2qs1OqhEP1av1xxq+4cZJV4OalAjka//N0bxCyNUBomqNZdz02Mn1FlOBM4LfVSjQllYzrErqWSRqj9bP7tlJxZZUDCWNmRhszV3xcZjbSeRIHdjKgZ6WVvJv7rBdFSsgmv/IzLJDUo2SI4TAUxMZlVQQZcITNiYgllitvfCRtRRZmxhZVsKd5yBaukVat6btW7u6jUr/N6inACp3AOHlxCHW6hAU1g8ADP8AKvzpPz5rw7H4vVgpPfHMMfOJ8/ijiVhg==
    n2
    AAAB+nicbVDLSgMxFL2pr1pfVZdugkVwVWaKoMuiG5cV7QPaoWTSTBuaZIYkI5Sxn+BW9+7ErT/j1i8xbWehrQcuHM65l3M5YSK4sZ73hQpr6xubW8Xt0s7u3v5B+fCoZeJUU9aksYh1JySGCa5Y03IrWCfRjMhQsHY4vpn57UemDY/Vg50kLJBkqHjEKbFOulf9Wr9c8areHHiV+DmpQI5Gv/zdG8Q0lUxZKogxXd9LbJARbTkVbFrqpYYlhI7JkHUdVUQyE2TzV6f4zCkDHMXajbJ4rv6+yIg0ZiJDtymJHZllbyb+64VyKdlGV0HGVZJapugiOEoFtjGe9YAHXDNqxcQRQjV3v2M6IppQ69oquVL85QpWSatW9b2qf3dRqV/n9RThBE7hHHy4hDrcQgOaQGEIz/ACr+gJvaF39LFYLaD85hj+AH3+AD6wlD0=
    AAAB+nicbVDLSgMxFL2pr1pfVZdugkVwVWaKoMuiG5cV7QPaoWTSTBuaZIYkI5Sxn+BW9+7ErT/j1i8xbWehrQcuHM65l3M5YSK4sZ73hQpr6xubW8Xt0s7u3v5B+fCoZeJUU9aksYh1JySGCa5Y03IrWCfRjMhQsHY4vpn57UemDY/Vg50kLJBkqHjEKbFOulf9Wr9c8areHHiV+DmpQI5Gv/zdG8Q0lUxZKogxXd9LbJARbTkVbFrqpYYlhI7JkHUdVUQyE2TzV6f4zCkDHMXajbJ4rv6+yIg0ZiJDtymJHZllbyb+64VyKdlGV0HGVZJapugiOEoFtjGe9YAHXDNqxcQRQjV3v2M6IppQ69oquVL85QpWSatW9b2qf3dRqV/n9RThBE7hHHy4hDrcQgOaQGEIz/ACr+gJvaF39LFYLaD85hj+AH3+AD6wlD0=
    AAAB+nicbVDLSgMxFL2pr1pfVZdugkVwVWaKoMuiG5cV7QPaoWTSTBuaZIYkI5Sxn+BW9+7ErT/j1i8xbWehrQcuHM65l3M5YSK4sZ73hQpr6xubW8Xt0s7u3v5B+fCoZeJUU9aksYh1JySGCa5Y03IrWCfRjMhQsHY4vpn57UemDY/Vg50kLJBkqHjEKbFOulf9Wr9c8areHHiV+DmpQI5Gv/zdG8Q0lUxZKogxXd9LbJARbTkVbFrqpYYlhI7JkHUdVUQyE2TzV6f4zCkDHMXajbJ4rv6+yIg0ZiJDtymJHZllbyb+64VyKdlGV0HGVZJapugiOEoFtjGe9YAHXDNqxcQRQjV3v2M6IppQ69oquVL85QpWSatW9b2qf3dRqV/n9RThBE7hHHy4hDrcQgOaQGEIz/ACr+gJvaF39LFYLaD85hj+AH3+AD6wlD0=
    AAAB+nicbVDLSgMxFL2pr1pfVZdugkVwVWaKoMuiG5cV7QPaoWTSTBuaZIYkI5Sxn+BW9+7ErT/j1i8xbWehrQcuHM65l3M5YSK4sZ73hQpr6xubW8Xt0s7u3v5B+fCoZeJUU9aksYh1JySGCa5Y03IrWCfRjMhQsHY4vpn57UemDY/Vg50kLJBkqHjEKbFOulf9Wr9c8areHHiV+DmpQI5Gv/zdG8Q0lUxZKogxXd9LbJARbTkVbFrqpYYlhI7JkHUdVUQyE2TzV6f4zCkDHMXajbJ4rv6+yIg0ZiJDtymJHZllbyb+64VyKdlGV0HGVZJapugiOEoFtjGe9YAHXDNqxcQRQjV3v2M6IppQ69oquVL85QpWSatW9b2qf3dRqV/n9RThBE7hHHy4hDrcQgOaQGEIz/ACr+gJvaF39LFYLaD85hj+AH3+AD6wlD0=
    `3
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyqoMegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL3zXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d1GpXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AIvMlYc=
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyqoMegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL3zXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d1GpXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AIvMlYc=
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyqoMegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL3zXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d1GpXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AIvMlYc=
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKeyqoMegF48RzAOSJcxOepMxM7vLzKwQluAveNW7N/Hqt3j1S5wke9DEgoaiqptqKkgE18Z1v5zCyura+kZxs7S1vbO7V94/aOo4VQwbLBaxagdUo+ARNgw3AtuJQioDga1gdDP1W4+oNI+jezNO0Jd0EPGQM2qs1OyiEL3zXrniVt0ZyDLxclKBHPVe+bvbj1kqMTJMUK07npsYP6PKcCZwUuqmGhPKRnSAHUsjKlH72ezbCTmxSp+EsbITGTJTf19kVGo9loHdlNQM9aI3Ff/1ArmQbMIrP+NRkhqM2Dw4TAUxMZlWQfpcITNibAllitvfCRtSRZmxhZVsKd5iBcukeVb13Kp3d1GpXef1FOEIjuEUPLiEGtxCHRrA4AGe4QVenSfnzXl3PuarBSe/OYQ/cD5/AIvMlYc=
    `4
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKexKQI9BLx4jmAckS5id9CZjZmeXmVkhLMFf8Kp3b+LVb/HqlzhJ9qCJBQ1FVTfVVJAIro3rfjmFtfWNza3idmlnd2//oHx41NJxqhg2WSxi1QmoRsElNg03AjuJQhoFAtvB+Gbmtx9RaR7LezNJ0I/oUPKQM2qs1OqhEP1av1xxq+4cZJV4OalAjka//N0bxCyNUBomqNZdz02Mn1FlOBM4LfVSjQllYzrErqWSRqj9bP7tlJxZZUDCWNmRhszV3xcZjbSeRIHdjKgZ6WVvJv7rBdFSsgmv/IzLJDUo2SI4TAUxMZlVQQZcITNiYgllitvfCRtRRZmxhZVsKd5yBaukdVH13Kp3V6vUr/N6inACp3AOHlxCHW6hAU1g8ADP8AKvzpPz5rw7H4vVgpPfHMMfOJ8/jWCViA==
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKexKQI9BLx4jmAckS5id9CZjZmeXmVkhLMFf8Kp3b+LVb/HqlzhJ9qCJBQ1FVTfVVJAIro3rfjmFtfWNza3idmlnd2//oHx41NJxqhg2WSxi1QmoRsElNg03AjuJQhoFAtvB+Gbmtx9RaR7LezNJ0I/oUPKQM2qs1OqhEP1av1xxq+4cZJV4OalAjka//N0bxCyNUBomqNZdz02Mn1FlOBM4LfVSjQllYzrErqWSRqj9bP7tlJxZZUDCWNmRhszV3xcZjbSeRIHdjKgZ6WVvJv7rBdFSsgmv/IzLJDUo2SI4TAUxMZlVQQZcITNiYgllitvfCRtRRZmxhZVsKd5yBaukdVH13Kp3V6vUr/N6inACp3AOHlxCHW6hAU1g8ADP8AKvzpPz5rw7H4vVgpPfHMMfOJ8/jWCViA==
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKexKQI9BLx4jmAckS5id9CZjZmeXmVkhLMFf8Kp3b+LVb/HqlzhJ9qCJBQ1FVTfVVJAIro3rfjmFtfWNza3idmlnd2//oHx41NJxqhg2WSxi1QmoRsElNg03AjuJQhoFAtvB+Gbmtx9RaR7LezNJ0I/oUPKQM2qs1OqhEP1av1xxq+4cZJV4OalAjka//N0bxCyNUBomqNZdz02Mn1FlOBM4LfVSjQllYzrErqWSRqj9bP7tlJxZZUDCWNmRhszV3xcZjbSeRIHdjKgZ6WVvJv7rBdFSsgmv/IzLJDUo2SI4TAUxMZlVQQZcITNiYgllitvfCRtRRZmxhZVsKd5yBaukdVH13Kp3V6vUr/N6inACp3AOHlxCHW6hAU1g8ADP8AKvzpPz5rw7H4vVgpPfHMMfOJ8/jWCViA==
    AAAB/XicbVDLSgNBEOyNrxhfUY9eBoPgKexKQI9BLx4jmAckS5id9CZjZmeXmVkhLMFf8Kp3b+LVb/HqlzhJ9qCJBQ1FVTfVVJAIro3rfjmFtfWNza3idmlnd2//oHx41NJxqhg2WSxi1QmoRsElNg03AjuJQhoFAtvB+Gbmtx9RaR7LezNJ0I/oUPKQM2qs1OqhEP1av1xxq+4cZJV4OalAjka//N0bxCyNUBomqNZdz02Mn1FlOBM4LfVSjQllYzrErqWSRqj9bP7tlJxZZUDCWNmRhszV3xcZjbSeRIHdjKgZ6WVvJv7rBdFSsgmv/IzLJDUo2SI4TAUxMZlVQQZcITNiYgllitvfCRtRRZmxhZVsKd5yBaukdVH13Kp3V6vUr/N6inACp3AOHlxCHW6hAU1g8ADP8AKvzpPz5rw7H4vVgpPfHMMfOJ8/jWCViA==
    n3
    AAAB+nicbVC7SgNBFL0bXzG+opY2g0GwCrsqaBm0sYxoHpAsYXYymwyZxzIzK4Q1n2CrvZ3Y+jO2fomTZAtNPHDhcM69nMuJEs6M9f0vr7Cyura+UdwsbW3v7O6V9w+aRqWa0AZRXOl2hA3lTNKGZZbTdqIpFhGnrWh0M/Vbj1QbpuSDHSc0FHggWcwItk66l73zXrniV/0Z0DIJclKBHPVe+bvbVyQVVFrCsTGdwE9smGFtGeF0UuqmhiaYjPCAdhyVWFATZrNXJ+jEKX0UK+1GWjRTf19kWBgzFpHbFNgOzaI3Ff/1IrGQbOOrMGMySS2VZB4cpxxZhaY9oD7TlFg+dgQTzdzviAyxxsS6tkqulGCxgmXSPKsGfjW4u6jUrvN6inAEx3AKAVxCDW6hDg0gMIBneIFX78l78969j/lqwctvDuEPvM8fQESUPg==
    AAAB+nicbVC7SgNBFL0bXzG+opY2g0GwCrsqaBm0sYxoHpAsYXYymwyZxzIzK4Q1n2CrvZ3Y+jO2fomTZAtNPHDhcM69nMuJEs6M9f0vr7Cyura+UdwsbW3v7O6V9w+aRqWa0AZRXOl2hA3lTNKGZZbTdqIpFhGnrWh0M/Vbj1QbpuSDHSc0FHggWcwItk66l73zXrniV/0Z0DIJclKBHPVe+bvbVyQVVFrCsTGdwE9smGFtGeF0UuqmhiaYjPCAdhyVWFATZrNXJ+jEKX0UK+1GWjRTf19kWBgzFpHbFNgOzaI3Ff/1IrGQbOOrMGMySS2VZB4cpxxZhaY9oD7TlFg+dgQTzdzviAyxxsS6tkqulGCxgmXSPKsGfjW4u6jUrvN6inAEx3AKAVxCDW6hDg0gMIBneIFX78l78969j/lqwctvDuEPvM8fQESUPg==
    AAAB+nicbVC7SgNBFL0bXzG+opY2g0GwCrsqaBm0sYxoHpAsYXYymwyZxzIzK4Q1n2CrvZ3Y+jO2fomTZAtNPHDhcM69nMuJEs6M9f0vr7Cyura+UdwsbW3v7O6V9w+aRqWa0AZRXOl2hA3lTNKGZZbTdqIpFhGnrWh0M/Vbj1QbpuSDHSc0FHggWcwItk66l73zXrniV/0Z0DIJclKBHPVe+bvbVyQVVFrCsTGdwE9smGFtGeF0UuqmhiaYjPCAdhyVWFATZrNXJ+jEKX0UK+1GWjRTf19kWBgzFpHbFNgOzaI3Ff/1IrGQbOOrMGMySS2VZB4cpxxZhaY9oD7TlFg+dgQTzdzviAyxxsS6tkqulGCxgmXSPKsGfjW4u6jUrvN6inAEx3AKAVxCDW6hDg0gMIBneIFX78l78969j/lqwctvDuEPvM8fQESUPg==
    AAAB+nicbVC7SgNBFL0bXzG+opY2g0GwCrsqaBm0sYxoHpAsYXYymwyZxzIzK4Q1n2CrvZ3Y+jO2fomTZAtNPHDhcM69nMuJEs6M9f0vr7Cyura+UdwsbW3v7O6V9w+aRqWa0AZRXOl2hA3lTNKGZZbTdqIpFhGnrWh0M/Vbj1QbpuSDHSc0FHggWcwItk66l73zXrniV/0Z0DIJclKBHPVe+bvbVyQVVFrCsTGdwE9smGFtGeF0UuqmhiaYjPCAdhyVWFATZrNXJ+jEKX0UK+1GWjRTf19kWBgzFpHbFNgOzaI3Ff/1IrGQbOOrMGMySS2VZB4cpxxZhaY9oD7TlFg+dgQTzdzviAyxxsS6tkqulGCxgmXSPKsGfjW4u6jUrvN6inAEx3AKAVxCDW6hDg0gMIBneIFX78l78969j/lqwctvDuEPvM8fQESUPg==
    x1
    AAAB+nicbVDLSgMxFL1TX7W+qi7dBIvgqkxE0GXRjcuK9gHtUDJppg1NMkOSEcvYT3Cre3fi1p9x65eYtrPQ1gMXDufcy7mcMBHcWN//8gorq2vrG8XN0tb2zu5eef+gaeJUU9agsYh1OySGCa5Yw3IrWDvRjMhQsFY4up76rQemDY/VvR0nLJBkoHjEKbFOunvs4V654lf9GdAywTmpQI56r/zd7cc0lUxZKogxHewnNsiItpwKNil1U8MSQkdkwDqOKiKZCbLZqxN04pQ+imLtRlk0U39fZEQaM5ah25TEDs2iNxX/9UK5kGyjyyDjKkktU3QeHKUC2RhNe0B9rhm1YuwIoZq73xEdEk2odW2VXCl4sYJl0jyrYr+Kb88rtau8niIcwTGcAoYLqMEN1KEBFAbwDC/w6j15b9679zFfLXj5zSH8gff5A0z4lEY=
    AAAB+nicbVDLSgMxFL1TX7W+qi7dBIvgqkxE0GXRjcuK9gHtUDJppg1NMkOSEcvYT3Cre3fi1p9x65eYtrPQ1gMXDufcy7mcMBHcWN//8gorq2vrG8XN0tb2zu5eef+gaeJUU9agsYh1OySGCa5Yw3IrWDvRjMhQsFY4up76rQemDY/VvR0nLJBkoHjEKbFOunvs4V654lf9GdAywTmpQI56r/zd7cc0lUxZKogxHewnNsiItpwKNil1U8MSQkdkwDqOKiKZCbLZqxN04pQ+imLtRlk0U39fZEQaM5ah25TEDs2iNxX/9UK5kGyjyyDjKkktU3QeHKUC2RhNe0B9rhm1YuwIoZq73xEdEk2odW2VXCl4sYJl0jyrYr+Kb88rtau8niIcwTGcAoYLqMEN1KEBFAbwDC/w6j15b9679zFfLXj5zSH8gff5A0z4lEY=
    AAAB+nicbVDLSgMxFL1TX7W+qi7dBIvgqkxE0GXRjcuK9gHtUDJppg1NMkOSEcvYT3Cre3fi1p9x65eYtrPQ1gMXDufcy7mcMBHcWN//8gorq2vrG8XN0tb2zu5eef+gaeJUU9agsYh1OySGCa5Yw3IrWDvRjMhQsFY4up76rQemDY/VvR0nLJBkoHjEKbFOunvs4V654lf9GdAywTmpQI56r/zd7cc0lUxZKogxHewnNsiItpwKNil1U8MSQkdkwDqOKiKZCbLZqxN04pQ+imLtRlk0U39fZEQaM5ah25TEDs2iNxX/9UK5kGyjyyDjKkktU3QeHKUC2RhNe0B9rhm1YuwIoZq73xEdEk2odW2VXCl4sYJl0jyrYr+Kb88rtau8niIcwTGcAoYLqMEN1KEBFAbwDC/w6j15b9679zFfLXj5zSH8gff5A0z4lEY=
    AAAB+nicbVDLSgMxFL1TX7W+qi7dBIvgqkxE0GXRjcuK9gHtUDJppg1NMkOSEcvYT3Cre3fi1p9x65eYtrPQ1gMXDufcy7mcMBHcWN//8gorq2vrG8XN0tb2zu5eef+gaeJUU9agsYh1OySGCa5Yw3IrWDvRjMhQsFY4up76rQemDY/VvR0nLJBkoHjEKbFOunvs4V654lf9GdAywTmpQI56r/zd7cc0lUxZKogxHewnNsiItpwKNil1U8MSQkdkwDqOKiKZCbLZqxN04pQ+imLtRlk0U39fZEQaM5ah25TEDs2iNxX/9UK5kGyjyyDjKkktU3QeHKUC2RhNe0B9rhm1YuwIoZq73xEdEk2odW2VXCl4sYJl0jyrYr+Kb88rtau8niIcwTGcAoYLqMEN1KEBFAbwDC/w6j15b9679zFfLXj5zSH8gff5A0z4lEY=
    x2
    AAAB+nicbVDLSgMxFL3xWeur6tJNsAiuykwRdFl047KifUA7lEyaaUOTzJBkxDL2E9zq3p249Wfc+iWm7Sy09cCFwzn3ci4nTAQ31vO+0Mrq2vrGZmGruL2zu7dfOjhsmjjVlDVoLGLdDolhgivWsNwK1k40IzIUrBWOrqd+64Fpw2N1b8cJCyQZKB5xSqyT7h571V6p7FW8GfAy8XNShhz1Xum7249pKpmyVBBjOr6X2CAj2nIq2KTYTQ1LCB2RAes4qohkJshmr07wqVP6OIq1G2XxTP19kRFpzFiGblMSOzSL3lT81wvlQrKNLoOMqyS1TNF5cJQKbGM87QH3uWbUirEjhGrufsd0SDSh1rVVdKX4ixUsk2a14nsV//a8XLvK6ynAMZzAGfhwATW4gTo0gMIAnuEFXtETekPv6GO+uoLymyP4A/T5A06MlEc=
    AAAB+nicbVDLSgMxFL3xWeur6tJNsAiuykwRdFl047KifUA7lEyaaUOTzJBkxDL2E9zq3p249Wfc+iWm7Sy09cCFwzn3ci4nTAQ31vO+0Mrq2vrGZmGruL2zu7dfOjhsmjjVlDVoLGLdDolhgivWsNwK1k40IzIUrBWOrqd+64Fpw2N1b8cJCyQZKB5xSqyT7h571V6p7FW8GfAy8XNShhz1Xum7249pKpmyVBBjOr6X2CAj2nIq2KTYTQ1LCB2RAes4qohkJshmr07wqVP6OIq1G2XxTP19kRFpzFiGblMSOzSL3lT81wvlQrKNLoOMqyS1TNF5cJQKbGM87QH3uWbUirEjhGrufsd0SDSh1rVVdKX4ixUsk2a14nsV//a8XLvK6ynAMZzAGfhwATW4gTo0gMIAnuEFXtETekPv6GO+uoLymyP4A/T5A06MlEc=
    AAAB+nicbVDLSgMxFL3xWeur6tJNsAiuykwRdFl047KifUA7lEyaaUOTzJBkxDL2E9zq3p249Wfc+iWm7Sy09cCFwzn3ci4nTAQ31vO+0Mrq2vrGZmGruL2zu7dfOjhsmjjVlDVoLGLdDolhgivWsNwK1k40IzIUrBWOrqd+64Fpw2N1b8cJCyQZKB5xSqyT7h571V6p7FW8GfAy8XNShhz1Xum7249pKpmyVBBjOr6X2CAj2nIq2KTYTQ1LCB2RAes4qohkJshmr07wqVP6OIq1G2XxTP19kRFpzFiGblMSOzSL3lT81wvlQrKNLoOMqyS1TNF5cJQKbGM87QH3uWbUirEjhGrufsd0SDSh1rVVdKX4ixUsk2a14nsV//a8XLvK6ynAMZzAGfhwATW4gTo0gMIAnuEFXtETekPv6GO+uoLymyP4A/T5A06MlEc=
    AAAB+nicbVDLSgMxFL3xWeur6tJNsAiuykwRdFl047KifUA7lEyaaUOTzJBkxDL2E9zq3p249Wfc+iWm7Sy09cCFwzn3ci4nTAQ31vO+0Mrq2vrGZmGruL2zu7dfOjhsmjjVlDVoLGLdDolhgivWsNwK1k40IzIUrBWOrqd+64Fpw2N1b8cJCyQZKB5xSqyT7h571V6p7FW8GfAy8XNShhz1Xum7249pKpmyVBBjOr6X2CAj2nIq2KTYTQ1LCB2RAes4qohkJshmr07wqVP6OIq1G2XxTP19kRFpzFiGblMSOzSL3lT81wvlQrKNLoOMqyS1TNF5cJQKbGM87QH3uWbUirEjhGrufsd0SDSh1rVVdKX4ixUsk2a14nsV//a8XLvK6ynAMZzAGfhwATW4gTo0gMIAnuEFXtETekPv6GO+uoLymyP4A/T5A06MlEc=
    dn0
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8XPpr65QqyGw1UrdYgsmvIdd26IejCrTcc6NhojgpYouWXPwdhQrKYCk04VqrvoFR7OZaaEU6npUGmaIrJGA9p31CBY6q8fH7uFJ4ZJYRRIk0JDefqz4kcx0pN4sB0xliP1Ko3E//0gnhls47qXs5EmmkqyGJxlHGoEzjLAoZMUqL5xBBMJDO3QzLCEhNtEiuZUL4/h/+Tjms7yHZuq5Xm1TKeIjgBp+AcOOASNMENaIE2IGAMHsETeLYerBfr1XpbtBas5cwx+AXr/QsAXZZn
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8XPpr65QqyGw1UrdYgsmvIdd26IejCrTcc6NhojgpYouWXPwdhQrKYCk04VqrvoFR7OZaaEU6npUGmaIrJGA9p31CBY6q8fH7uFJ4ZJYRRIk0JDefqz4kcx0pN4sB0xliP1Ko3E//0gnhls47qXs5EmmkqyGJxlHGoEzjLAoZMUqL5xBBMJDO3QzLCEhNtEiuZUL4/h/+Tjms7yHZuq5Xm1TKeIjgBp+AcOOASNMENaIE2IGAMHsETeLYerBfr1XpbtBas5cwx+AXr/QsAXZZn
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8XPpr65QqyGw1UrdYgsmvIdd26IejCrTcc6NhojgpYouWXPwdhQrKYCk04VqrvoFR7OZaaEU6npUGmaIrJGA9p31CBY6q8fH7uFJ4ZJYRRIk0JDefqz4kcx0pN4sB0xliP1Ko3E//0gnhls47qXs5EmmkqyGJxlHGoEzjLAoZMUqL5xBBMJDO3QzLCEhNtEiuZUL4/h/+Tjms7yHZuq5Xm1TKeIjgBp+AcOOASNMENaIE2IGAMHsETeLYerBfr1XpbtBas5cwx+AXr/QsAXZZn
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8XPpr65QqyGw1UrdYgsmvIdd26IejCrTcc6NhojgpYouWXPwdhQrKYCk04VqrvoFR7OZaaEU6npUGmaIrJGA9p31CBY6q8fH7uFJ4ZJYRRIk0JDefqz4kcx0pN4sB0xliP1Ko3E//0gnhls47qXs5EmmkqyGJxlHGoEzjLAoZMUqL5xBBMJDO3QzLCEhNtEiuZUL4/h/+Tjms7yHZuq5Xm1TKeIjgBp+AcOOASNMENaIE2IGAMHsETeLYerBfr1XpbtBas5cwx+AXr/QsAXZZn
    dn4
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSYxxXZXdOOygn1AG8JkMm2HTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+YcqY0Qh9WYW19Y3OruF3a2d3bPygfHrVVkklCWyThieyGWFHOBG1ppjntppLiOOS0E46vZ37nnkrFEnGnJyn1YzwUbMAI1kbqREEuAm8alCvIrteR51UhsqvIdd2aIejCrdUd6NhojgpYohmUP/tRQrKYCk04VqrnoFT7OZaaEU6npX6maIrJGA9pz1CBY6r8fH7uFJ4ZJYKDRJoSGs7VnxM5jpWaxKHpjLEeqVVvJv7phfHKZj2o+TkTaaapIIvFg4xDncBZFjBikhLNJ4ZgIpm5HZIRlphok1jJhPL9OfyftF3bQbZz61UaV8t4iuAEnIJz4IBL0AA3oAlagIAxeARP4Nl6sF6sV+tt0VqwljPH4Bes9y8GsZZr
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSYxxXZXdOOygn1AG8JkMm2HTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+YcqY0Qh9WYW19Y3OruF3a2d3bPygfHrVVkklCWyThieyGWFHOBG1ppjntppLiOOS0E46vZ37nnkrFEnGnJyn1YzwUbMAI1kbqREEuAm8alCvIrteR51UhsqvIdd2aIejCrdUd6NhojgpYohmUP/tRQrKYCk04VqrnoFT7OZaaEU6npX6maIrJGA9pz1CBY6r8fH7uFJ4ZJYKDRJoSGs7VnxM5jpWaxKHpjLEeqVVvJv7phfHKZj2o+TkTaaapIIvFg4xDncBZFjBikhLNJ4ZgIpm5HZIRlphok1jJhPL9OfyftF3bQbZz61UaV8t4iuAEnIJz4IBL0AA3oAlagIAxeARP4Nl6sF6sV+tt0VqwljPH4Bes9y8GsZZr
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSYxxXZXdOOygn1AG8JkMm2HTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+YcqY0Qh9WYW19Y3OruF3a2d3bPygfHrVVkklCWyThieyGWFHOBG1ppjntppLiOOS0E46vZ37nnkrFEnGnJyn1YzwUbMAI1kbqREEuAm8alCvIrteR51UhsqvIdd2aIejCrdUd6NhojgpYohmUP/tRQrKYCk04VqrnoFT7OZaaEU6npX6maIrJGA9pz1CBY6r8fH7uFJ4ZJYKDRJoSGs7VnxM5jpWaxKHpjLEeqVVvJv7phfHKZj2o+TkTaaapIIvFg4xDncBZFjBikhLNJ4ZgIpm5HZIRlphok1jJhPL9OfyftF3bQbZz61UaV8t4iuAEnIJz4IBL0AA3oAlagIAxeARP4Nl6sF6sV+tt0VqwljPH4Bes9y8GsZZr
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSYxxXZXdOOygn1AG8JkMm2HTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+YcqY0Qh9WYW19Y3OruF3a2d3bPygfHrVVkklCWyThieyGWFHOBG1ppjntppLiOOS0E46vZ37nnkrFEnGnJyn1YzwUbMAI1kbqREEuAm8alCvIrteR51UhsqvIdd2aIejCrdUd6NhojgpYohmUP/tRQrKYCk04VqrnoFT7OZaaEU6npX6maIrJGA9pz1CBY6r8fH7uFJ4ZJYKDRJoSGs7VnxM5jpWaxKHpjLEeqVVvJv7phfHKZj2o+TkTaaapIIvFg4xDncBZFjBikhLNJ4ZgIpm5HZIRlphok1jJhPL9OfyftF3bQbZz61UaV8t4iuAEnIJz4IBL0AA3oAlagIAxeARP4Nl6sF6sV+tt0VqwljPH4Bes9y8GsZZr
    dn1
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8XvjP1yxVkNxqoWq1BZNeQ67p1Q9CFW2840LHRHBWwRMsvfw7ChGQxFZpwrFTfQan2ciw1I5xOS4NM0RSTMR7SvqECx1R5+fzcKTwzSgijRJoSGs7VnxM5jpWaxIHpjLEeqVVvJv7pBfHKZh3VvZyJNNNUkMXiKONQJ3CWBQyZpETziSGYSGZuh2SEJSbaJFYyoXx/Dv8nHdd2kO3cVivNq2U8RXACTsE5cMAlaIIb0AJtQMAYPIIn8Gw9WC/Wq/W2aC1Yy5lj8AvW+xcB8pZo
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8XvjP1yxVkNxqoWq1BZNeQ67p1Q9CFW2840LHRHBWwRMsvfw7ChGQxFZpwrFTfQan2ciw1I5xOS4NM0RSTMR7SvqECx1R5+fzcKTwzSgijRJoSGs7VnxM5jpWaxIHpjLEeqVVvJv7pBfHKZh3VvZyJNNNUkMXiKONQJ3CWBQyZpETziSGYSGZuh2SEJSbaJFYyoXx/Dv8nHdd2kO3cVivNq2U8RXACTsE5cMAlaIIb0AJtQMAYPIIn8Gw9WC/Wq/W2aC1Yy5lj8AvW+xcB8pZo
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8XvjP1yxVkNxqoWq1BZNeQ67p1Q9CFW2840LHRHBWwRMsvfw7ChGQxFZpwrFTfQan2ciw1I5xOS4NM0RSTMR7SvqECx1R5+fzcKTwzSgijRJoSGs7VnxM5jpWaxIHpjLEeqVVvJv7pBfHKZh3VvZyJNNNUkMXiKONQJ3CWBQyZpETziSGYSGZuh2SEJSbaJFYyoXx/Dv8nHdd2kO3cVivNq2U8RXACTsE5cMAlaIIb0AJtQMAYPIIn8Gw9WC/Wq/W2aC1Yy5lj8AvW+xcB8pZo
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8XvjP1yxVkNxqoWq1BZNeQ67p1Q9CFW2840LHRHBWwRMsvfw7ChGQxFZpwrFTfQan2ciw1I5xOS4NM0RSTMR7SvqECx1R5+fzcKTwzSgijRJoSGs7VnxM5jpWaxIHpjLEeqVVvJv7pBfHKZh3VvZyJNNNUkMXiKONQJ3CWBQyZpETziSGYSGZuh2SEJSbaJFYyoXx/Dv8nHdd2kO3cVivNq2U8RXACTsE5cMAlaIIb0AJtQMAYPIIn8Gw9WC/Wq/W2aC1Yy5lj8AvW+xcB8pZo
    dn2
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8Xvjv1yxVkNxqoWq1BZNeQ67p1Q9CFW2840LHRHBWwRMsvfw7ChGQxFZpwrFTfQan2ciw1I5xOS4NM0RSTMR7SvqECx1R5+fzcKTwzSgijRJoSGs7VnxM5jpWaxIHpjLEeqVVvJv7pBfHKZh3VvZyJNNNUkMXiKONQJ3CWBQyZpETziSGYSGZuh2SEJSbaJFYyoXx/Dv8nHdd2kO3cVivNq2U8RXACTsE5cMAlaIIb0AJtQMAYPIIn8Gw9WC/Wq/W2aC1Yy5lj8AvW+xcDh5Zp
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8Xvjv1yxVkNxqoWq1BZNeQ67p1Q9CFW2840LHRHBWwRMsvfw7ChGQxFZpwrFTfQan2ciw1I5xOS4NM0RSTMR7SvqECx1R5+fzcKTwzSgijRJoSGs7VnxM5jpWaxIHpjLEeqVVvJv7pBfHKZh3VvZyJNNNUkMXiKONQJ3CWBQyZpETziSGYSGZuh2SEJSbaJFYyoXx/Dv8nHdd2kO3cVivNq2U8RXACTsE5cMAlaIIb0AJtQMAYPIIn8Gw9WC/Wq/W2aC1Yy5lj8AvW+xcDh5Zp
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8Xvjv1yxVkNxqoWq1BZNeQ67p1Q9CFW2840LHRHBWwRMsvfw7ChGQxFZpwrFTfQan2ciw1I5xOS4NM0RSTMR7SvqECx1R5+fzcKTwzSgijRJoSGs7VnxM5jpWaxIHpjLEeqVVvJv7pBfHKZh3VvZyJNNNUkMXiKONQJ3CWBQyZpETziSGYSGZuh2SEJSbaJFYyoXx/Dv8nHdd2kO3cVivNq2U8RXACTsE5cMAlaIIb0AJtQMAYPIIn8Gw9WC/Wq/W2aC1Yy5lj8AvW+xcDh5Zp
    AAAB/nicdVDLSsNAFJ3UV62vqks3g0VwFSaxxXZXdOOygn1AG8JkMmmHTiZhZiKUUPAX3Orenbj1V9z6JU4fghY9cOFwzr3ce0+QcqY0Qh9WYW19Y3OruF3a2d3bPygfHnVUkklC2yThiewFWFHOBG1rpjntpZLiOOC0G4yvZ373nkrFEnGnJyn1YjwULGIEayN1Qz8Xvjv1yxVkNxqoWq1BZNeQ67p1Q9CFW2840LHRHBWwRMsvfw7ChGQxFZpwrFTfQan2ciw1I5xOS4NM0RSTMR7SvqECx1R5+fzcKTwzSgijRJoSGs7VnxM5jpWaxIHpjLEeqVVvJv7pBfHKZh3VvZyJNNNUkMXiKONQJ3CWBQyZpETziSGYSGZuh2SEJSbaJFYyoXx/Dv8nHdd2kO3cVivNq2U8RXACTsE5cMAlaIIb0AJtQMAYPIIn8Gw9WC/Wq/W2aC1Yy5lj8AvW+xcDh5Zp
    dn3
    AAAB/nicdVDLSsNAFJ34rPVVdelmsAiuwiRtsd0V3bisYB/QhjCZTNqhk0mYmQglFPwFt7p3J279Fbd+idOHoEUPXDiccy/33hOknCmN0Ie1tr6xubVd2Cnu7u0fHJaOjjsqySShbZLwRPYCrChngrY105z2UklxHHDaDcbXM797T6ViibjTk5R6MR4KFjGCtZG6oZ8LvzL1S2VkNxqoWq1BZNeQ67p1Q1DFrTcc6NhojjJYouWXPgdhQrKYCk04VqrvoFR7OZaaEU6nxUGmaIrJGA9p31CBY6q8fH7uFJ4bJYRRIk0JDefqz4kcx0pN4sB0xliP1Ko3E//0gnhls47qXs5EmmkqyGJxlHGoEzjLAoZMUqL5xBBMJDO3QzLCEhNtEiuaUL4/h/+Tjms7yHZuq+Xm1TKeAjgFZ+ACOOASNMENaIE2IGAMHsETeLYerBfr1XpbtK5Zy5kT8AvW+xcFHJZq
    AAAB/nicdVDLSsNAFJ34rPVVdelmsAiuwiRtsd0V3bisYB/QhjCZTNqhk0mYmQglFPwFt7p3J279Fbd+idOHoEUPXDiccy/33hOknCmN0Ie1tr6xubVd2Cnu7u0fHJaOjjsqySShbZLwRPYCrChngrY105z2UklxHHDaDcbXM797T6ViibjTk5R6MR4KFjGCtZG6oZ8LvzL1S2VkNxqoWq1BZNeQ67p1Q1DFrTcc6NhojjJYouWXPgdhQrKYCk04VqrvoFR7OZaaEU6nxUGmaIrJGA9p31CBY6q8fH7uFJ4bJYRRIk0JDefqz4kcx0pN4sB0xliP1Ko3E//0gnhls47qXs5EmmkqyGJxlHGoEzjLAoZMUqL5xBBMJDO3QzLCEhNtEiuaUL4/h/+Tjms7yHZuq+Xm1TKeAjgFZ+ACOOASNMENaIE2IGAMHsETeLYerBfr1XpbtK5Zy5kT8AvW+xcFHJZq
    AAAB/nicdVDLSsNAFJ34rPVVdelmsAiuwiRtsd0V3bisYB/QhjCZTNqhk0mYmQglFPwFt7p3J279Fbd+idOHoEUPXDiccy/33hOknCmN0Ie1tr6xubVd2Cnu7u0fHJaOjjsqySShbZLwRPYCrChngrY105z2UklxHHDaDcbXM797T6ViibjTk5R6MR4KFjGCtZG6oZ8LvzL1S2VkNxqoWq1BZNeQ67p1Q1DFrTcc6NhojjJYouWXPgdhQrKYCk04VqrvoFR7OZaaEU6nxUGmaIrJGA9p31CBY6q8fH7uFJ4bJYRRIk0JDefqz4kcx0pN4sB0xliP1Ko3E//0gnhls47qXs5EmmkqyGJxlHGoEzjLAoZMUqL5xBBMJDO3QzLCEhNtEiuaUL4/h/+Tjms7yHZuq+Xm1TKeAjgFZ+ACOOASNMENaIE2IGAMHsETeLYerBfr1XpbtK5Zy5kT8AvW+xcFHJZq
    AAAB/nicdVDLSsNAFJ34rPVVdelmsAiuwiRtsd0V3bisYB/QhjCZTNqhk0mYmQglFPwFt7p3J279Fbd+idOHoEUPXDiccy/33hOknCmN0Ie1tr6xubVd2Cnu7u0fHJaOjjsqySShbZLwRPYCrChngrY105z2UklxHHDaDcbXM797T6ViibjTk5R6MR4KFjGCtZG6oZ8LvzL1S2VkNxqoWq1BZNeQ67p1Q1DFrTcc6NhojjJYouWXPgdhQrKYCk04VqrvoFR7OZaaEU6nxUGmaIrJGA9p31CBY6q8fH7uFJ4bJYRRIk0JDefqz4kcx0pN4sB0xliP1Ko3E//0gnhls47qXs5EmmkqyGJxlHGoEzjLAoZMUqL5xBBMJDO3QzLCEhNtEiuaUL4/h/+Tjms7yHZuq+Xm1TKeAjgFZ+ACOOASNMENaIE2IGAMHsETeLYerBfr1XpbtK5Zy5kT8AvW+xcFHJZq
    P(y | x)
    AAACAXicdZDLSgMxFIYzXmu9VV26CRahbkpGam13RTcuK9gLtkPJpJk2NMkMSUYsQ1e+glvduxO3Polbn8RMW0GL/hD4+c45nJPfjzjTBqEPZ2l5ZXVtPbOR3dza3tnN7e03dRgrQhsk5KFq+1hTziRtGGY4bUeKYuFz2vJHl2m9dUeVZqG8MeOIegIPJAsYwcai23ph3BWsD+9Perk8Kp4ht1ouQ1REyC1VXGuq1YqF0LUkVR7MVe/lPrv9kMSCSkM41rrjosh4CVaGEU4n2W6saYTJCA9ox1qJBdVeMr14Ao8t6cMgVPZJA6f050SChdZj4dtOgc1QL9ZS+GfNFwubTVDxEiaj2FBJZouDmEMTwjQO2GeKEsPH1mCimL0dkiFWmBgbWtaG8v1z+L9pnhZdVHSvS/naxTyeDDgER6AAXHAOauAK1EEDECDBI3gCz86D8+K8Om+z1iVnPnMAfsl5/wJZI5cV
    AAACAXicdZDLSgMxFIYzXmu9VV26CRahbkpGam13RTcuK9gLtkPJpJk2NMkMSUYsQ1e+glvduxO3Polbn8RMW0GL/hD4+c45nJPfjzjTBqEPZ2l5ZXVtPbOR3dza3tnN7e03dRgrQhsk5KFq+1hTziRtGGY4bUeKYuFz2vJHl2m9dUeVZqG8MeOIegIPJAsYwcai23ph3BWsD+9Perk8Kp4ht1ouQ1REyC1VXGuq1YqF0LUkVR7MVe/lPrv9kMSCSkM41rrjosh4CVaGEU4n2W6saYTJCA9ox1qJBdVeMr14Ao8t6cMgVPZJA6f050SChdZj4dtOgc1QL9ZS+GfNFwubTVDxEiaj2FBJZouDmEMTwjQO2GeKEsPH1mCimL0dkiFWmBgbWtaG8v1z+L9pnhZdVHSvS/naxTyeDDgER6AAXHAOauAK1EEDECDBI3gCz86D8+K8Om+z1iVnPnMAfsl5/wJZI5cV
    AAACAXicdZDLSgMxFIYzXmu9VV26CRahbkpGam13RTcuK9gLtkPJpJk2NMkMSUYsQ1e+glvduxO3Polbn8RMW0GL/hD4+c45nJPfjzjTBqEPZ2l5ZXVtPbOR3dza3tnN7e03dRgrQhsk5KFq+1hTziRtGGY4bUeKYuFz2vJHl2m9dUeVZqG8MeOIegIPJAsYwcai23ph3BWsD+9Perk8Kp4ht1ouQ1REyC1VXGuq1YqF0LUkVR7MVe/lPrv9kMSCSkM41rrjosh4CVaGEU4n2W6saYTJCA9ox1qJBdVeMr14Ao8t6cMgVPZJA6f050SChdZj4dtOgc1QL9ZS+GfNFwubTVDxEiaj2FBJZouDmEMTwjQO2GeKEsPH1mCimL0dkiFWmBgbWtaG8v1z+L9pnhZdVHSvS/naxTyeDDgER6AAXHAOauAK1EEDECDBI3gCz86D8+K8Om+z1iVnPnMAfsl5/wJZI5cV
    AAACAXicdZDLSgMxFIYzXmu9VV26CRahbkpGam13RTcuK9gLtkPJpJk2NMkMSUYsQ1e+glvduxO3Polbn8RMW0GL/hD4+c45nJPfjzjTBqEPZ2l5ZXVtPbOR3dza3tnN7e03dRgrQhsk5KFq+1hTziRtGGY4bUeKYuFz2vJHl2m9dUeVZqG8MeOIegIPJAsYwcai23ph3BWsD+9Perk8Kp4ht1ouQ1REyC1VXGuq1YqF0LUkVR7MVe/lPrv9kMSCSkM41rrjosh4CVaGEU4n2W6saYTJCA9ox1qJBdVeMr14Ao8t6cMgVPZJA6f050SChdZj4dtOgc1QL9ZS+GfNFwubTVDxEiaj2FBJZouDmEMTwjQO2GeKEsPH1mCimL0dkiFWmBgbWtaG8v1z+L9pnhZdVHSvS/naxTyeDDgER6AAXHAOauAK1EEDECDBI3gCz86D8+K8Om+z1iVnPnMAfsl5/wJZI5cV
    Region-wise model
    0 1
    0
    0 0
    0
    1 1
    1 1
    Edward
    Pyro
    Prob Torch
    BayesFlow
    ੜ໋Պֶ/ҩɾༀɾੜ෺ Խֶ(ྔࢠԽֶɾ৮ഔԽֶɾੜԽֶɾ༗ػԽֶ) ෺࣭ɾࡐྉՊֶ

    View full-size slide

  4. ʮ཭ࢄߏ଄ʯΛ൐͏ػցֶश
    4
    ܾఆ໦ɾܾఆDAG χϡʔϥϧωοτ ֬཰తϓϩάϥϛϯά
    • ର৅͕ʮ཭ࢄߏ଄ʯΛ࣋ͭ
    • Ϟσϧ͕ʮ཭ࢄߏ଄ʯΛ࣋ͭ
    • ର৅ͷؔ܎͕ʮ཭ࢄߏ଄ʯΛ࣋ͭ
    H
    H
    H
    H
    H
    H
    H
    H
    O
    N
    O
    O
    H
    H
    H
    O
    O
    H
    H
    N
    O
    O
    Cl
    Cl
    Cl
    ू߹ɺܥྻɺ૊߹ͤɺஔ׵ɺ෼ذ(໦)ɺωοτϫʔΫ(άϥϑ)ɺ…

    View full-size slide

  5. 5
    ౷ܭ਺ཧݚڀॴ DATߨ࠲ L-B2

    View full-size slide

  6. ౷ܭ਺ཧݚڀॴ DATߨ࠲ L-S
    6

    View full-size slide

  7. స৬ɿ2019೥4݄1೔ʙ
    7
    ๺ւಓେֶ৘ใՊֶݚڀՊͷݚڀࣨΛclose͠Լه̎૊৫ͷ

    ʮΫϩεΞϙΠϯτϝϯτʯ΁
    • ཧԽֶݚڀॴ 

    ֵ৽஌ೳ౷߹ݚڀηϯλʔ (AIP)

    iPSࡉ๔࿈ܞҩֶతϦεΫճආνʔϜ ݚڀһ
    • ๺ւಓେֶ 

    Խֶ൓Ԡ૑੒ݚڀڌ఺ (WPI-ICReDD) ಛ೚।ڭत
    70%
    30%

    View full-size slide

  8. 8
    ๺ւಓେֶ Խֶ൓Ԡ૑੒ݚڀڌ఺ (WPI-ICReDD)
    ਆాϥϘ
    ૑੒ݚڀػߏ౩ 05-116 ૑੒ݚڀػߏ౩ 02-106

    View full-size slide

  9. 9
    ཧԽֶݚڀॴ ֵ৽஌ೳ౷߹ݚڀηϯλʔ
    東京駅
    理研AIP
    神田ラボ
    皇居

    View full-size slide

  10. 10
    ཧԽֶݚڀॴ ֵ৽஌ೳ౷߹ݚڀηϯλʔ
    COREDO೔ຊڮ15F

    View full-size slide

  11. 11
    ۈ຿஍ɿژࡕಸ஍۠(ژ౎෎૬ָ܊ਫ਼՚ொ)
    https://www.kobe.riken.jp/about/map/keihanna/

    View full-size slide

  12. 12
    ۈ຿஍ɿژࡕಸ஍۠(ژ౎෎૬ָ܊ਫ਼՚ொ)
    礵螟٥銮加峸㖑⼒խ䒉暟ꂁ縧㔳

    View full-size slide

  13. 13
    ۈ຿஍ɿژࡕಸ஍۠(ژ౎෎૬ָ܊ਫ਼՚ொ)
    ࠃཱࠃձਤॻؗؔ੢ؗ

    View full-size slide

  14. 14
    ࠃࡍిؾ௨৴جૅٕज़ݚڀॴʢATRʣ
    ੴࠇಛݚͷ
    ΞϯυϩΠυ
    ʮΤϦΧʯ༷͕௟࠲

    (࣮෺͸ࡱӨېࢭ)
    • ೴৘ใ௨৴૯߹ݚڀॴ
    • ஌ೳϩϘςΟΫεݚڀॴ
    • దԠίϛϡχέʔγϣϯݚڀॴ
    • ೾ಈ޻ֶݚڀॴ
    • ੴࠇߒಛผݚڀॴ (ੴࠇERATO)
    • ࠤ౻ঊಙಛผݚڀॴ (ࠤ౻ERATO)

    View full-size slide

  15. 15
    ࠃࡍిؾ௨৴جૅٕज़ݚڀॴʢATRʣ
    • ೴৘ใݚڀॴ (CNS)
    • ೝ஌ػߏݚڀॴ (CMC)
    • ೴৘ใղੳݚڀॴ (NIA)
    • ܭࢉ೴Πϝʔδϯάݚڀࣨ (CBI)

    ≒ ཧݚAIP ܭࢉ೴μΠφϛΫενʔϜ (ࢁԼT)
    • ಈత೴Πϝʔδϯάݚڀࣨ (DBI)

    ≒ ཧݚAIP ೴৘ใ౷߹ղੳνʔϜ (઒ುT)
    • ೴৘ใ௨৴૯߹ݚڀॴ

    View full-size slide

  16. 16
    ATR಺ ཧݚAIP
    • ๷ࡂՊֶνʔϜ (্ా मޭ)
    • ೴৘ใ౷߹ղੳνʔϜ (઒ು Ұߊ)
    • ܭࢉ೴μΠφϛΫενʔϜ (ࢁԼ ஦ਓ)
    • iPSࡉ๔࿈ܞҩֶతϦεΫճආνʔϜ (্ా मޭ)
    ࢲͱ্ాTLҎ֎͸ژେCiRAͷํʑ...

    View full-size slide

  17. ࣗݾ঺հɿ୍઒ Ұֶ(͖͕ͨΘɾ͍͕ͪ͘)
    17
    10೥ ๺େ
    (1995ʙ2004)
    7೥ ژେ
    (2005ʙ2011)
    7೥ ๺େ
    (2012ʙ2018)
    ౷ܭత৴߸ॲཧͱύλʔϯೝࣝ (޻ֶݚڀՊ)

    "ྼܾఆ৴߸ݯ෼཭ͷL1ϊϧϜ࠷খղͷཧ࿦෼ੳ"
    όΠΦΠϯϑΥϚςΟΫε (Խֶݚڀॴ)

    έϞΠϯϑΥϚςΟΫε (ༀֶݚڀՊ)
    σʔλۦಈՊֶɾ ཭ࢄߏ଄Λ൐͏ػցֶश

    (৘ใՊֶݚڀՊ)

    + JST͖͕͚͞: ࡐྉΠϯϑΥϚςΟΫε
    ʁ೥ ཧݚ(ژ౎)
    (2019ʙ)
    AIPηϯλʔ iPSࡉ๔࿈ܞҩֶతϦεΫճආνʔϜ

    (๺େ Խֶ൓Ԡ૑੒ݚڀڌ఺ͱΫϩΞϙ)
    ઐ໳ɿػցֶशɾσʔλϚΠχϯάͱͦͷՊֶͰͷར׆༻

    ɹɹʮσʔλ͔ΒͷֶशʯΛͲ͏໰୊ղܾʹ׆༻Ͱ͖Δͷ͔ʁ

    View full-size slide

  18. REVIEW
    Inverse molecular design using
    machine learning: Generative models
    for matter engineering
    Benjamin Sanchez-Lengeling1 and Alán Aspuru-Guzik2,3,4*
    The discovery of new materials can bring enormous societal and technological progress. In this
    context, exploring completely the large space of potential materials is computationally
    intractable. Here, we review methods for achieving inverse design, which aims to discover
    tailored materials from the starting point of a particular desired functionality. Recent advances
    from the rapidly growing field of artificial intelligence, mostly from the subfield of machine
    learning, have resulted in a fertile exchange of ideas, where approaches to inverse molecular
    design are being proposed and employed at a rapid pace. Among these, deep generative models
    have been applied to numerous classes of materials: rational design of prospective drugs,
    synthetic routes to organic compounds, and optimization of photovoltaics and redox flow
    batteries, as well as a variety of other solid-state materials.
    Many of the challenges of the 21st century
    (1), from personalized health care to
    energy production and storage, share a
    common theme: materials are part of
    the solution (2). In some cases, the solu-
    tions to these challenges are fundamentally
    limited by the physics and chemistry of a ma-
    terial, such as the relationship of a materials
    bandgap to the thermodynamic limits for the
    generation of solar energy (3).
    Several important materials discoveries arose
    by chance or through a process of trial and error.
    For example, vulcanized rubber was prepared in
    the 19th century from random mixtures of com-
    pounds, based on the observation that heating
    with additives such as sulfur improved the
    rubber’s durability. At the molecular level, in-
    dividual polymer chains cross-linked, forming
    bridges that enhanced the macroscopic mechan-
    ical properties (4). Other notable examples in
    this vein include Teflon, anesthesia, Vaseline,
    Perkin’s mauve, and penicillin. Furthermore,
    these materials come from common chemical
    compounds found in nature. Potential drugs
    either were prepared by synthesis in a chem-
    ical laboratory or were isolated from plants,
    soil bacteria, or fungus. For example, up until
    2014, 49% of small-molecule cancer drugs were
    natural products or their derivatives (5).
    In the future, disruptive advances in the dis-
    covery of matter could instead come from unex-
    plored regions of the set of all possible molecular
    and solid-state compounds, known as chemical
    space (6, 7). One of the largest collections of
    molecules, the chemical space project (8), has
    mapped 166.4 billion molecules that contain at
    most 17 heavy atoms. For pharmacologically rele-
    vant small molecules, the number of structures is
    estimated to be on the order of 1060 (9). Adding
    consideration of the hierarchy of scale from sub-
    nanometer to microscopic and mesoscopic fur-
    ther complicates exploration of chemical space
    in its entirety (10). Therefore, any global strategy
    for covering this space might seem impossible.
    Simulation offers one way of probing this
    space without experimentation. The physics
    and chemistry of these molecules are governed
    by quantum mechanics, which can be solved via
    the Schrödinger equation to arrive at their ex-
    act properties. In practice, approximations are
    used to lower computational time at the cost of
    accuracy.
    Although theory enjoys enormous progress,
    now routinely modeling molecules, clusters, and
    perfect as well as defect-laden periodic solids, the
    size of chemical space is still overwhelming, and
    smart navigation is required. For this purpose,
    machine learning (ML), deep learning (DL), and
    artificial intelligence (AI) have a potential role
    to play because their computational strategies
    automatically improve through experience (11).
    In the context of materials, ML techniques are
    often used for property prediction, seeking to
    learn a function that maps a molecular material
    to the property of choice. Deep generative models
    are a special class of DL methods that seek to
    model the underlying probability distribution of
    both structure and property and relate them in a
    nonlinear way. By exploiting patterns in massive
    datasets, these models can distill average and
    salient features that characterize molecules (12, 13).
    Inverse design is a component of a more
    complex materials discovery process. The time
    scale for deployment of new technologies, from
    discovery in a laboratory to a commercial pro-
    duct, historically, is 15 to 20 years (14). The pro-
    cess (Fig. 1) conventionally involves the following
    steps: (i) generate a new or improved material
    concept and simulate its potential suitability; (ii)
    synthesize the material; (iii) incorporate the ma-
    terial into a device or system; and (iv) characterize
    and measure the desired properties. This cycle
    generates feedback to repeat, improve, and re-
    fine future cycles of discovery. Each step can take
    up to several years.
    In the era of matter engineering, scientists
    seek to accelerate these cycles, reducing the
    FRONTIERS IN COMPUTATION
    1Department of Chemistry and Chemical Biology, Harvard
    LOSKI
    on July 26, 2018
    http://science.sciencemag.org/
    Downloaded from
    REVIEW
    https://doi.org/10.1038/s41586-018-0337-2
    Machine learning for molecular and
    materials science
    Keith T. Butler1, Daniel W
    . Davies2, Hugh Cartwright3, Olexandr Isayev4* & Aron Walsh5,6*
    Here we summarize recent progress in machine learning for the chemical sciences. We outline machine-learning
    techniques that are suitable for addressing research questions in this domain, as well as future directions for the field.
    We envisage a future in which the design, synthesis, characterization and application of molecules and materials is
    accelerated by artificial intelligence.
    The Schrödinger equation provides a powerful structure–
    property relationship for molecules and materials. For a given
    spatial arrangement of chemical elements, the distribution of
    electrons and a wide range of physical responses can be described. The
    development of quantum mechanics provided a rigorous theoretical
    foundation for the chemical bond. In 1929, Paul Dirac famously proclaimed
    that the underlying physical laws for the whole of chemistry are “completely
    known”1. John Pople, realizing the importance of rapidly developing
    computer technologies, created a program—Gaussian 70—that could
    perform ab initio calculations: predicting the behaviour, for molecules
    of modest size, purely from the fundamental laws of physics2. In the 1960s,
    the Quantum Chemistry Program Exchange brought quantum chemistry
    to the masses in the form of useful practical tools3. Suddenly, experi-
    mentalists with little or no theoretical training could perform quantum
    calculations too. Using modern algorithms and supercomputers,
    systems containing thousands of interacting ions and electrons can now
    be described using approximations to the physical laws that govern the
    world on the atomic scale4–6.
    The field of computational chemistry has become increasingly pre-
    dictive in the twenty-first century, with activity in applications as wide
    ranging as catalyst development for greenhouse gas conversion, materials
    discovery for energy harvesting and storage, and computer-assisted drug
    design7. The modern chemical-simulation toolkit allows the properties
    of a compound to be anticipated (with reasonable accuracy) before it has
    been made in the laboratory. High-throughput computational screening
    has become routine, giving scientists the ability to calculate the properties
    of thousands of compounds as part of a single study. In particular, den-
    sity functional theory (DFT)8,9, now a mature technique for calculating
    the structure and behaviour of solids10, has enabled the development of
    extensive databases that cover the calculated properties of known and
    hypothetical systems, including organic and inorganic crystals, single
    molecules and metal alloys11–13.
    The emergence of contemporary artificial-intelligence methods has
    the potential to substantially alter and enhance the role of computers in
    science and engineering. The combination of big data and artificial intel-
    ligence has been referred to as both the “fourth paradigm of science”14
    and the “fourth industrial revolution”15, and the number of applications
    in the chemical domain is growing at an astounding rate. A subfield of
    artificial intelligence that has evolved rapidly in recent years is machine
    generating, testing and refining scientific models. Such techniques are
    suitable for addressing complex problems that involve massive combi-
    natorial spaces or nonlinear processes, which conventional procedures
    either cannot solve or can tackle only at great computational cost.
    As the machinery for artificial intelligence and machine learning
    matures, important advances are being made not only by those in main-
    stream artificial-intelligence research, but also by experts in other fields
    (domain experts) who adopt these approaches for their own purposes. As
    we detail in Box 1, the resources and tools that facilitate the application
    of machine-learning techniques mean that the barrier to entry is lower
    than ever.
    In the rest of this Review, we discuss progress in the application of
    machine learning to address challenges in molecular and materials
    research. We review the basics of machine-learning approaches, iden-
    tify areas in which existing methods have the potential to accelerate
    research and consider the developments that are required to enable more
    wide-ranging impacts.
    Nuts and bolts of machine learning
    With machine learning, given enough data and a rule-discovery algo-
    rithm, a computer has the ability to determine all known physical laws
    (and potentially those that are currently unknown) without human
    input. In traditional computational approaches, the computer is little
    more than a calculator, employing a hard-coded algorithm provided
    by a human expert. By contrast, machine-learning approaches learn
    the rules that underlie a dataset by assessing a portion of that data
    and building a model to make predictions. We consider the basic steps
    involved in the construction of a model, as illustrated in Fig. 1; this
    constitutes a blueprint of the generic workflow that is required for the
    successful application of machine learning in a materials-discovery
    process.
    Data collection
    Machine learning comprises models that learn from existing (train-
    ing) data. Data may require initial preprocessing, during which miss-
    ing or spurious elements are identified and handled. For example, the
    Inorganic Crystal Structure Database (ICSD) currently contains more
    than 190,000 entries, which have been checked for technical mistakes
    but are still subject to human and measurement errors. Identifying
    DNA to be sequences into distinct pieces,
    parcel out the detailed work of sequencing,
    and then reassemble these independent ef-
    forts at the end. It is not quite so simple in the
    world of genome semantics.
    Despite the differences between genome se-
    quencing and genetic network discovery, there
    are clear parallels that are illustrated in Table 1.
    In genome sequencing, a physical map is useful
    to provide scaffolding for assembling the fin-
    ished sequence. In the case of a genetic regula-
    tory network, a graphical model can play the
    same role. A graphical model can represent a
    high-level view of interconnectivity and help
    isolate modules that can be studied indepen-
    dently. Like contigs in a genomic sequencing
    project, low-level functional models can ex-
    plore the detailed behavior of a module of genes
    in a manner that is consistent with the higher
    level graphical model of the system. With stan-
    dardized nomenclature and compatible model-
    ing techniques, independent functional models
    can be assembled into a complete model of the
    cell under study.
    To enable this process, there will need to
    be standardized forms for model representa-
    tion. At present, there are many different
    modeling technologies in use, and although
    models can be easily placed into a database,
    they are not useful out of the context of their
    specific modeling package. The need for a
    standardized way of communicating compu-
    tational descriptions of biological systems ex-
    tends to the literature. Entire conferences
    have been established to explore ways of
    mining the biology literature to extract se-
    mantic information in computational form.
    Going forward, as a community we need
    to come to consensus on how to represent
    what we know about biology in computa-
    tional form as well as in words. The key to
    postgenomic biology will be the computa-
    tional assembly of our collective knowl-
    edge into a cohesive picture of cellular and
    organism function. With such a comprehen-
    sive model, we will be able to explore new
    types of conservation between organisms
    and make great strides toward new thera-
    peutics that function on well-characterized
    pathways.
    References
    1. S. K. Kim et al., Science 293, 2087 (2001).
    2. A. Hartemink et al., paper presented at the Pacific
    Symposium on Biocomputing 2000, Oahu, Hawaii, 4
    to 9 January 2000.
    3. D. Pe’er et al., paper presented at the 9th Conference
    on Intelligent Systems in Molecular Biology (ISMB),
    Copenhagen, Denmark, 21 to 25 July 2001.
    4. H. McAdams, A. Arkin, Proc. Natl. Acad. Sci. U.S.A.
    94, 814 ( 1997 ).
    5. A. J. Hartemink, thesis, Massachusetts Institute of
    Technology, Cambridge (2001).
    V I E W P O I N T
    Machine Learning for Science: State of the
    Art and Future Prospects
    Eric Mjolsness* and Dennis DeCoste
    Recent advances in machine learning methods, along with successful
    applications across a wide variety of fields such as planetary science and
    bioinformatics, promise powerful new tools for practicing scientists. This
    viewpoint highlights some useful characteristics of modern machine learn-
    ing methods and their relevance to scientific applications. We conclude
    with some speculations on near-term progress and promising directions.
    Machine learning (ML) (1) is the study of
    computer algorithms capable of learning to im-
    prove their performance of a task on the basis of
    their own previous experience. The field is
    closely related to pattern recognition and statis-
    tical inference. As an engineering field, ML has
    become steadily more mathematical and more
    successful in applications over the past 20
    years. Learning approaches such as data clus-
    tering, neural network classifiers, and nonlinear
    regression have found surprisingly wide appli-
    cation in the practice of engineering, business,
    and science. A generalized version of the stan-
    dard Hidden Markov Models of ML practice
    have been used for ab initio prediction of gene
    structures in genomic DNA (2). The predictions
    correlate surprisingly well with subsequent
    gene expression analysis (3). Postgenomic bi-
    ology prominently features large-scale gene ex-
    pression data analyzed by clustering methods
    (4), a standard topic in unsupervised learning.
    Many other examples can be given of learning
    and pattern recognition applications in science.
    Where will this trend lead? We believe it will
    lead to appropriate, partial automation of every
    element of scientific method, from hypothesis
    generation to model construction to decisive
    experimentation. Thus, ML has the potential to
    amplify every aspect of a working scientist’s
    progress to understanding. It will also, for better
    or worse, endow intelligent computer systems
    with some of the general analytic power of
    scientific thinking.
    Machine Learning at Every Stage of
    the Scientific Process
    Each scientific field has its own version of the
    scientific process. But the cycle of observing,
    creating hypotheses, testing by decisive exper-
    iment or observation, and iteratively building
    up comprehensive testable models or theories is
    shared across disciplines. For each stage of this
    abstracted scientific process, there are relevant
    developments in ML, statistical inference, and
    pattern recognition that will lead to semiauto-
    matic support tools of unknown but potentially
    broad applicability.
    Increasingly, the early elements of scientific
    method—observation and hypothesis genera-
    tion—face high data volumes, high data acqui-
    sition rates, or requirements for objective anal-
    ysis that cannot be handled by human percep-
    tion alone. This has been the situation in exper-
    imental particle physics for decades. There
    automatic pattern recognition for significant
    events is well developed, including Hough
    transforms, which are foundational in pattern
    recognition. A recent example is event analysis
    for Cherenkov detectors (8) used in neutrino
    oscillation experiments. Microscope imagery in
    cell biology, pathology, petrology, and other
    fields has led to image-processing specialties.
    So has remote sensing from Earth-observing
    satellites, such as the newly operational Terra
    spacecraft with its ASTER (a multispectral
    thermal radiometer), MISR (multiangle imag-
    ing spectral radiometer), MODIS (imaging
    Machine Learning Systems Group, Jet Propulsion Lab-
    oratory/California Institute of Technology, Pasadena,
    CA, 91109, USA.
    *To whom correspondence should be addressed. E-
    mail: [email protected]
    Table 1. Parallels between genome sequencing
    and genetic network discovery.
    Genome
    sequencing
    Genome semantics
    Physical maps Graphical model
    Contigs Low-level functional
    models
    Contig
    reassembly
    Module assembly
    Finished genome
    sequence
    Comprehensive model
    www.sciencemag.org SCIENCE VOL 293 14 SEPTEMBER 2001 2051
    C O M P U T E R S A N D S C I E N C E
    on August 29, 2018
    http://science.sciencemag.org/
    Downloaded from
    Nature, 559

    pp. 547–555 (2018)
    Science, 293
    pp. 2051-2055 (2001)
    Science, 361
    pp. 360-365 (2018)
    Science is changing, the tools of science are changing. And
    that requires different approaches. ── Erich Bloch, 1925-2016
    ʮσʔλར׆༻ٕज़ʯ͸Պֶݚڀͷಓ۩ͷҰͭʹ
    18
    ڭ܇ "low input, high throughput, no output science." (Sydney Brenner)
    → ࡶͳઃఆɾܥͰ໢ཏతͳϋΠεϧʔϓοτ࣮ݧΛ͍͘Βͯ͠΋Կ΋ಘΒΕͳ͍

    View full-size slide

  19. ຊ೔ͷ࿩ͷલஔ͖
    19
    ࠓ೔͸ɺͦͷޙͷ࿩Λ൓өͨ͠࠶ʑฤ൛Ͱ͢ɻ

    View full-size slide

  20. Deepͳ৘ใܥͷํʹ͸
    20
    ͱ͜ΖͰɺࠓ݄ͷਓ޻஌ೳֶձࢽͷ୍઒͞ΜͷSIG-FPAIهࣄɺ
    ͨͩͷݚڀձͷ঺հهࣄͷϋζͷͳͷʹɺ͑Β͍ݟࣝʹͳ͍ͬͯͯɺ
    ଞͷݚڀձͷهࣄͱͷམࠩʹࠔ࿭͠·ͨ͠ɻ
    Έͳ͞·΋ੋඇɻ kashi_pong ڭत (ژ౎େֶ)
    ਓ޻஌ೳֶձࢽ
    Vol.34 No.5 (2019/9)
    ಛूɿʮݚڀձ঺հʯ
    ਓ޻஌ೳجຊ໰୊ݚڀձʢ'1"*ʣ
    ΦʔϓϯΞΫηε
    https://sig-fpai.org/

    View full-size slide

  21. ਓ޻஌ೳֶձ ਓ޻஌ೳجຊ໰୊ݚڀձ(SIG-FPAI)
    21

    View full-size slide

  22. ষཱͯ
    1. ͸͡Ίʹ
    2. ͦͷ্ʹ෺ͷݐͨͳ͍΋ͷ͸جૅͱ͸͍Θͳ͍
    3. มΘΔ΋ͷɼมΘΒͳ͍΋ͷ
    4. The Hard Thing about Hard Things
    5. ػցֶशͱࣗಈϓϩάϥϛϯά: બ୒ͱֶशͷؒ
    6. ૊߹ͤͷ൚Խ: ཭ࢄͱ࿈ଓͷؒ
    7. ػցൃݟͱࣗಈԽͷເ: ֶशͱൃݟͷؒ
    8. දݱͱհೖ: ܦݧ࿦ͱ߹ཧ࿦ͷؒ
    9. աఔͱ࣮ࡏ: ༗ݶͱແݶͷؒ
    ݹ͍จݙαʔϕΠ

    ʹΑΔFAI/FPAIͷྺ࢙
    ౰࣌ͷ࿩୊ͱݱ୅ͷ

    ࿩୊ͷࢲͳΓͷϦϯΫ
    झຯతࡶײͱల๬
    (ࢀߟจݙ100݅)
    ਓ޻஌ೳֶձࢽ Vol.34 No.5 (2019/9)
    ಛूɿʮݚڀձ঺հʯ ਓ޻஌ೳجຊ໰୊ݚڀձʢ'1"*ʣ
    ΦʔϓϯΞΫηε Permalink : http://id.nii.ac.jp/1004/00010296/
    https://sig-fpai.org/

    View full-size slide

  23. Take Home Message
    24
    Պֶ͕ٻΊΔ͜ͱ: ෼͔Βͳ͍͜ͱ͕෼͔Δ(Պֶతൃݟ)
    ൃݟ
    ཧղ ݪҼͱ݁Ռ(ҼՌؔ܎)Λݟग़͢
    ࠓ·Ͱݟग़͞Ε͍ͯͳ͍ྑ͍ର৅Λݟग़͢
    ࠓ೔఻͍͑ͨͨͬͨ3ͭͷ͜ͱ
    1. ୯७ʹػցֶशΛ࢖͏͚ͩͰ͸͍ͣΕ΋ղ͚ͳ͍
    2. ػցֶशҎ֎ͷ΋ͷ(հೖ΍υϝΠϯ஌ࣝ)͕ݪཧ্ඞਢ
    3. ࠷ۙ·͞ʹݚڀ͕ਐߦதͷະղܾྖҬ͕ͩݚڀ͸৭ʑ͋Δ
    ͷաఔΛཧղ͠(ਓ͕ؒ)ൃݟ͢Δ
    x!y
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    Λར༻ͯ͠ྑ͍ Λ࣋ͭ Λൃݟ͢Δ
    x
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    x!y
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5

    View full-size slide

  24. ࠓ೔ͷ಺༰
    25
    1. Πϯτϩ

    ػցֶशͱՊֶ(͋Δ͍͸"΋ͷͮ͘Γ")
    2. ػցֶशͰԿ͔ΛʮཧղʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    3. ػցֶशͰԿ͔ΛʮൃݟʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    4. ͡Ό͋Ͳ͏͢ΜͷʂʁԿ͕͍Δͷʂʁ

    Answer:ʮදݱʯͱʮհೖʯ

    2ͱ3Λલఏʹػցֶश෼໺ͷτϐοΫΛ؆୯ʹ঺հ

    View full-size slide

  25. ࠓ೔ͷ಺༰
    25
    1. Πϯτϩ

    ػցֶशͱՊֶ(͋Δ͍͸"΋ͷͮ͘Γ")
    2. ػցֶशͰԿ͔ΛʮཧղʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    3. ػցֶशͰԿ͔ΛʮൃݟʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    4. ͡Ό͋Ͳ͏͢ΜͷʂʁԿ͕͍Δͷʂʁ

    Answer:ʮදݱʯͱʮհೖʯ

    2ͱ3Λલఏʹػցֶश෼໺ͷτϐοΫΛ؆୯ʹ঺հ

    View full-size slide

  26. Empirical optimization or "Edisonian empiricism"
    26
    ط஌ͷ஌ݟɾ
    ؍ଌ(σʔλ)
    • ࣮ݧ
    • γϛϡϨʔγϣϯ
    Ծઆܗ੒
    ݁Ռͷ֬ೝͱ
    ݕূ
    ࣍ͷ࣮ݧܭը΁feedback
    Thomas Edisonઌੜ
    • Genius is 1% inspiration and 99% perspiration.
    • There is no substitute for hard work.
    • I have not failed. I've just found 10,000 ways that won't work.
    :
    ໰୊ɿ࣌ؒͱίετ͸༗ݶʂʂ

    ཧ࿦తʹՄೳͳ͋ΒΏΔީิΛ

    ͜ͷํࣜͰݕূ͢Δ͜ͱ͸ෆՄೳ
    Α͘ߟ͑ΔͱϒϥοΫͳ͜ͱ͔͠ݴͬͯͳ͍ʂ
    Ծઆݕূ
    "؍࡯ͱؼೲ (empirical/inductive)" "࿦ཧͱԋ៷ (rational/deductive)"

    View full-size slide

  27. ՊֶతൃݟͱηϨϯσΟϐςΟ
    27
    ط஌ͷ஌ݟɾ
    ؍ଌ(σʔλ)
    • ࣮ݧ
    • γϛϡϨʔγϣϯ
    Ծઆܗ੒
    ݁Ռͷ֬ೝͱ
    ݕূ
    ࣍ͷ࣮ݧܭը΁feedback
    • ͦΕΏ͑ʮݚڀऀͷηϯεɾ࿹ͷݟͤॴʯʴʮ޾ӡ(ηϨϯσΟϐςΟ)ʯʹґଘ͢Δ

    ےͷྑͦ͞͏ͳީิΛબͿɺࠓ·Ͱࢼ͞Εͯͳ͍શ͘৽͍͠΍ΓํΛࢥ͍ͭ͘ɺetc
    • ީิ͕͋·Γʹ๲େ(࣮࣭΄΅ແݶ)ͳͷͰ(਺ଟ͘ࢼ͢ͷ͸༗རͩͱ͸ݴ͑...)

    ඞͣ͠΋ʮྗٕͱ͓ۚͱਓւઓज़Ͱ਺ଟ͘ࢼͨ͠ऀ͕উͭʯͱ͸ݶΒͳ͍
    ໰୊ɿ࣌ؒͱίετ͸༗ݶʂʂ

    ཧ࿦తʹՄೳͳ͋ΒΏΔީิΛ

    ͜ͷํࣜͰݕূ͢Δ͜ͱ͸ෆՄೳ
    Ծઆݕূ

    View full-size slide

  28. ػցֶश͸γϛϡϨʔγϣϯɾ࣮ݧͱ૬ิత
    28
    ط஌ͷ஌ݟɾ
    ؍ଌ(σʔλ)
    ݁Ռͷ֬ೝͱ
    ݕূ
    ߴ଎ɾߴਫ਼౓ͳ
    Data-Driven༧ଌ
    ࣍ͷ࣮ݧܭը΁feedback
    Ծઆܗ੒
    (γϛϡϨʔγϣϯ+࣮ݧ)
    • ࠶ݱੑΛ୲อ͢Δߴਫ਼౓ɾߴ଎࣮ݧܥ
    • Ծ૝Խݕূ͕ՄೳͳҼࢠͷγϛϡϨʔ
    γϣϯ(ܭࢉՊֶ)ʹΑΔ୳ࡧ

    → ๬·͍͠ର৅ͷ͞ΒͳΔߜΓࠐΈ
    Ծઆݕূ
    (ػցֶशɾσʔλϚΠχϯά)
    • Ͳ͏͍͏࣮ݧɾγϛϡϨʔγϣϯΛ

    ࣍ʹߦ͏͔ͷܭըཱҊ
    • ࣌ؒͷ͔͔Δܭࢉͷߴਫ਼౓ߴ଎ۙࣅ
    • ᐆດͳҼࢠ΍࣮ݧ৚݅ͷ࠷దԽ
    • Multilevelͷ৘ใ౷߹

    View full-size slide

  29. ࠓ೔ͷ಺༰
    29
    1. Πϯτϩ

    ػցֶशͱՊֶ(͋Δ͍͸"΋ͷͮ͘Γ")
    2. ػցֶशͰԿ͔ΛʮཧղʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    3. ػցֶशͰԿ͔ΛʮൃݟʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    4. ͡Ό͋Ͳ͏͢ΜͷʂʁԿ͕͍Δͷʂʁ

    Answer:ʮදݱʯͱʮհೖʯ

    2ͱ3Λલఏʹػցֶश෼໺ͷτϐοΫΛ؆୯ʹ঺հ

    View full-size slide

  30. ʮػցֶशʯతγνϡΤʔγϣϯ
    30
    ࣄྫɿࣸਅΛA͞Μ͔B͞Μ͔ʹ෼ྨ͢Δίϯϐϡʔλ

    ϓϩάϥϜΛ࡞Γ͍ͨɻ(େྔͷࣸਅΛਓख෼ྨ͢Δͷݏ)
    ? ? ?
    ? ? ?
    • ਓؒ͸؆୯ʹͰ͖Δ
    • ͕ɺͲ͏΍ͬͯ΍͍ͬͯΔͷ͔ݪཧ͸ෆ໌֬
    • ൅ܕɺ֯౓ɺর໌ɺഎܠɺද৘ɺԽহɺ೥ྸɺͳͲ

    Λߟ͑Δͱ໌ࣔతͳϓϩάϥϛϯά͸ͱͯ΋೉͍͠
    ..
    ?

    View full-size slide

  31. ػցֶश: ৽͍͠ϓϩάϥϛϯάͷ͔ͨͪ
    31
    Ұൠ෺ମೝࣝ
    ήʔϜϓϨΠ
    “͋Γ͕ͱ͏”
    J’aime
    la
    musiqu
    e
    I love music
    ೖग़ྗͷؔ܎͕Α͘෼͔Βͳ͍ม׵աఔ(ؔ਺)Λେྔͷೖग़ྗͷ
    ݟຊྫ͔Β໌ࣔతʹϓϩάϥϛϯά͢Δ͜ͱͳ͘ߏ੒͢Δٕ๏
    Ի੠ೝࣝ
    ػց຋༁
    ௒ղ૾

    View full-size slide

  32. ػցֶशͰ͖ͨͱ͖ͷ̎छྨͷظ଴
    32
    1. ಘΒΕͨม׵աఔ(ؔ਺)ʹΑΔ༧ଌΛ৭ʑͳ໨తʹ࢖͏
    2. ಘΒΕͨม׵աఔ(ؔ਺)Λ෼ੳͯ͠എܠաఔͷ࢓૊ΈΛ஌Δ
    Ұൠ෺ମೝࣝ
    ήʔϜϓϨΠ
    “͋Γ͕ͱ͏”
    J’aime
    la
    musiqu
    e
    I love music
    Ի੠ೝࣝ
    ػց຋༁
    ௒ղ૾

    View full-size slide

  33. ʮཧղʯฤͷϙΠϯτʂ
    33
    1͸OK͕ͩ
    2͸ͱͯ΋ඍົʂ
    ͦͯͦ͠ͷཧ༝΋·ͨඍົͳͷͰ͕͢
    ͋ͨͬͯ͞͠...
    1. ಘΒΕͨม׵աఔ(ؔ਺)ʹΑΔ༧ଌΛ৭ʑͳ໨తʹ࢖͏
    2. ಘΒΕͨม׵աఔ(ؔ਺)Λ෼ੳͯ͠എܠաఔͷ࢓૊ΈΛ஌Δ

    View full-size slide

  34. ػցֶश༧ଌͱͦͷཧղͷʮਂ͍ߔʯ
    34
    Լه͸ͲΕ΋ػցֶशͰ͔ͳΓߴਫ਼౓ͳ༧ଌ͕Ͱ͖·͕͢ɺ

    Ռͨͯͦ͠ͷ࢓૊Έͷཧղ͕ಘΒΕͨͷͰ͠ΐ͏͔...?
    Ͳ͏΍ͬͯݟ͍͑ͯΔ΋ͷΛೝࣝʁ
    Ͳ͏΍ͬͨΒғޟউͯΔʁ
    “͋Γ͕ͱ͏”
    J’aime
    la
    musiqu
    e
    I love music
    Ͳ͏΍ͬͯԻ೾ͷҙຯΛཧղʁ
    Ͳ͏΍ͬͯଟݴޠΛཧղʁ
    ࣮৘ɿ༧ଌ͸౰ͨΔΜ͚ͩͲཧ༝͸Α͘Θ͔Βͳ͍ʂ

    View full-size slide

  35. ༧ଌ͕͋ͨΕ͹ཧ༝͸Θ͔Βͳͯ͘΋OKʁ
    35
    Ͳ͏ߟ͑ͯ΋OKͳΘ͚ͳ͍΍Ζ...😫ͱࢥ͏͔΋͠Ε·ͤΜ
    ͕
    • ͦΕ͸༻్ʹΑΔ (༧ଌ͕ߴ͍ਫ਼౓Ͱ౰ͨΔͱ͍͏લఏͰ)
    ݱࡏͷ঎ۀత੒ޭΛݗҾ͢Δଟ͘ͷ༻్Ͱ͸ཁΒͳ͍৔߹΋ɻ
    ݕࡧɺ޿ࠂɺਪનɺηϯαʔ/IoTɺը૾ɾԻ੠ೝࣝɺܳज़ͳͲ
    • ࠷ۙɺద༻ઌ͕޿͕Γʮཧ༝ʯΛٻΊΒΕΔΑ͏ʹ
    ࣾձతʹΫϦςΟΧϧͳ໨తʹ౤ೖ͢ΔͳΒʮཧ༝ʯඞཁɻ
    ҩྍɺࣗಈ੍ޚɺΠϯϑϥ੍ޚɺޏ༻ɺ੓ࡦܾఆɺ༥ࢿͳͲ
    આ໌੹೚ɾಁ໌ੑɾެฏੑɾ҆શੑɾྙཧΛ୲อՄೳʁ

    View full-size slide

  36. ࿦จɿUse and Abuse of Regression (1966)
    36
    ࠓͷػցֶश 😉
    ίϯϐϡʔλ͕ॳΊͯ൰ۙͳಓ۩ʹͳΓɺਓʑ͸͋ΒΏΔ໨తʹ

    Data-driven(ճؼ෼ੳ)Λଟ༻͢ΔΑ͏ʹͳͬͯ͠·ͬͨ… 😅

    View full-size slide

  37. ࿦จɿUse and Abuse of Regression (1966)
    37
    "one of the great statistical minds of the 20th century"
    େ౷ܭֶऀ George E. P. Box (1919-2013)
    https://en.wikipedia.org/wiki/All_models_are_wrong
    "Essentially, all models are wrong,
    but some are useful"
    1. આ໌ม਺Λ؍ଌͨ͠ͱ͖ͷ໨తม਺ͷ༧ଌ
    2. આ໌ม਺ʹ֎తૢ࡞ΛՃ͑ͨͱ͖ͷ

    ໨తม਺΁ͷҼՌతޮՌͷൃݟ
    ճؼ෼ੳͷ໨త: (ڭࢣ෇ֶ͖शҰൠʹ౰ͯ͸·Δ)
    Use 😆
    Abuse 😫
    ...ۤݴʁ😅

    View full-size slide

  38. ૬ؔؔ܎͸ඞͣ͠΋ҼՌؔ܎Λҙຯ͠ͳ͍
    38
    ମॏ
    ਎௕
    Ԡ༻౷ܭֶͷΠϩϋɿCorrelation does not imply causation
    ମॏΛ૿΍ͤ͹

    ਎௕΋৳ͼΔʂʁ🤔
    ೔ຊϓϩ໺ٿ։ນҰ܉બखͷ਎௕ɾମॏσʔλ
    (2016೥ٿஂެࣜαΠτબखσʔλΑΓࣗ࡞)
    ͜Ε͕͓͔͍͠ͱ
    ͍͏͜ͱ͸͜ͷσʔλ
    ͚͔ͩΒ͸෼͔Βͳ͍
    ͦͯ͠ػցֶश͸σʔλʹ಺ࡏ͢Δ૬ؔؔ܎ͷར׆༻ٕज़

    View full-size slide

  39. N Engl J Med 2012; 367:1562-1564
    39
    νϣίϨʔτফඅྔ
    IF(2018) 70.670😳ͷ࠷΋ྺ࢙ͱݖҖͷ͋Δҩֶࢽ
    ਓޱ1ઍສਓ͋ͨΓͷ
    ϊʔϕϧ৆ड৆ऀ਺

    View full-size slide

  40. ؍࡯σʔλ͚͔ͩΒ͸ҼՌ͸෼͔Βͳ͍
    40
    http://phenomena.nationalgeographic.com/2015/09/11/nick-cage-movies-
    vs-drownings-and-more-strange-but-spurious-correlations/
    ΞϝϦΧͷՊֶ༧ࢉ vs
    ट௻ΓʹΑΔࣗࡴऀ਺
    ϓʔϧͰͷṆࢮऀ਺ vs
    χίϥεέΠδͷөըग़ԋ਺

    View full-size slide

  41. ͍ͭ૬ؔͱҼՌ͸ဃ཭͠͏Δͷ͔ʁ
    41
    ☺ؠ೾σʔλαΠΤϯε Vol. 3, ಛूɿҼՌਪ࿦――࣮ੈքͷσʔλ͔ΒҼՌΛಡΉ, 2016ΑΓ
    ૬ؔʹ
    ͍ͭͯ
    ͦΕҎ֎

    (Check೉)
    ҼՌؔ܎൑ఆͷHillͷΨΠυϥΠϯ (Hill, 1965)

    View full-size slide

  42. ަབྷҼࢠͱݟ͔͚ͤͷ૬ؔ
    42
    ྫɿҿञ͸ഏ͕ΜͷϦεΫཁҼͰ͋Δ(?)
    ٤Ԏ
    ަབྷҼࢠ(cofounders)
    ഏ͕Μ
    ҿञ
    ݟ͔͚ͤͷ૬ؔ(spurious correlation)
    Ҽࢠʮ٤Ԏʯ͕ަབྷ͍ͯ͠Δ
    ަབྷʹͲ͏ରॲ͢Δ͔ʁ
    ཧ૝ʮ࣮ݧ͢Δ(հೖ͢Δ)ʯɿ

    հೖ͢Δ͔൱͔Λແ࡞ҝʹׂΓ෇͚ΔϥϯμϜԽൺֱࢼݧ(RCT)
    ؍࡯ݚڀͰ͸Ͱ͖ͳ͍: ٤Ԏ͢Δ͔Ͳ͏͔ΛׂΓ෇͚Ͱ͖ͳ͍
    ҼՌͷ্ྲྀʹڞ௨Ҽࢠ͕ଘࡏ

    View full-size slide

  43. ؍࡯ݚڀʹΑΔҼՌਪ࿦ͷجຊ
    43
    ᶃ ૚ผ(Stratification)
    ᶄ ճؼ෼ੳͷར༻
    ٤Ԏ=༗ͷ܈ͱɺ٤Ԏ=ແͷ܈ʹ෼͚ɺ֤ʑղੳͨ͋͠ͱ౷߹
    ʮ٤ԎʯΛઆ໌ม਺ʹؚΊͯճؼ෼ੳͰ༗ҙੑݕఆ
    ඞཁͳલఏɿڵຯͷର৅ͷؔ܎ҼࢠͱަབྷҼࢠ͕͢΂ͯ

    ଌఆ͞Ε͍ͯΔ (͞ΒʹҼࢠͷؒͷҼՌߏ଄΋෼͔͍ͬͯΔ)
    ަབྷͦ͠͏ͳҼࢠ͸શͯઆ໌ม਺ʹೖΕ͓͚ͯ͹ྑ͍͕

    αϯϓϧ਺ʹΑͬͯ͸ճؼ෼ੳ͕ഁ୼ͯ͠͠·͏
    • ʮ܏޲είΞʯʹΑͬͯଟ਺ͷڞมྔΛ̍࣍ݩʹม׵͢Δ
    • ʮόοΫυΞج४ʯʹΑͬͯऔΓೖΕΔ΂͖આ໌ม਺ΛબͿ
    Α͘෼͔Βͳ͍ର৅Ͱ͸ݱ࣮తʹຬͨ͞ΕͮΒ͍...

    View full-size slide

  44. ౷ܭֶͱػցֶशͷʮߔʯ
    44
    • ౷ܭֶͱػցֶशͷҧ͍: σʔλ΍ม਺ʹର͢ΔԾఆ͕ҧ͏
    ౷ܭֶ: ੍ޚ͞Ε࣮ͨݧܭը (ྟচࢼݧ, ࣾձௐࠪ, ೶ۀࢼݧ,...)
    ಛ௃ྔ vs આ໌ม਺: Ҽࢠతҙຯ͸ͳ͍৔߹΋(ը૾ͷϐΫηϧ)
    • ૬ؔؔ܎ͷར׆༻Ͱ"༧ଌ͕ΊͬͪΌ౰ͨΔΜͳΒ͑͑΍Μ..."
    ͠͹͠͹ʮ͑͑Θ͚ͳ͍΍Ζ😫ʯͱ͍͏᫁᫟ΛੜΜͰ͖ͨ...
    e.g. ݴޠֶͷڊਓ Chomsky vs Googleݚڀ෦໳௕ Norvig
    http://norvig.com/chomsky.html
    ☺Statistical Modeling: The Two Cultures (Breiman, Statist. Sci. 16(3), 199-231, 2001)
    ػցֶश: ੍ޚ͞Εͳ͍σʔλ (ը૾, Ի੠, ςΩετ, ৴߸, ...)
    ஫ҙɿػցֶश԰͸ҼՌΛ͋·Γؾʹ͠ͳ͍

    View full-size slide

  45. σʔλۦಈՊֶ: Պֶ΋ҼՌ(ཧ༝)͕ओͨΔؔ৺
    45
    Պֶͷؔ৺͸ʮ࢓૊Έ΍ݪཧ͕Α͘෼͔Βͳ͍ݱ৅ʯ
    ? ؍ଌʗσʔλ
    • Theory-driven / Hypothesis-driven (ࣗવՊֶ)
    ExplicitͳԾઆɾཧ࿦ ؍ଌʗσʔλ
    • Data-driven (ػցֶशɺਓ޻஌ೳɺ౷ܭֶͳͲ)
    ৭ʑͳؔ਺ΛදݱͰ͖Δ൚༻Ϟσϧ ؍ଌʗσʔλ
    ԋ៷
    ؼೲ
    σʔλʹ࠷΋ద߹͢ΔΑ͏ʹϑΟοςΟϯά
    γϛϡϨʔγϣϯ & ਖ਼͍͔͠Ͳ͏͔͸(հೖ)࣮ݧͰ֬ೝ

    View full-size slide

  46. Data-driven vs Theory-driven
    46
    David Hand
    Data-driven͸Theory-drivenͱߟ͑ํɾ໾ׂ͕ҟͳΔͷͰ஫ҙ
    All models are wrong, but some are useful
    (George Box)
    Theory-driven models can be wrong
    But data-driven models cannot be wrong
    http://videolectures.net/kdd2018_hand_data_science/

    View full-size slide

  47. Data-driven vs Theory-driven
    46
    David Hand
    Data-driven͸Theory-drivenͱߟ͑ํɾ໾ׂ͕ҟͳΔͷͰ஫ҙ
    All models are wrong, but some are useful
    (George Box)
    Theory-driven models can be wrong
    But data-driven models cannot be wrong
    or right
    http://videolectures.net/kdd2018_hand_data_science/

    View full-size slide

  48. Data-driven vs Theory-driven
    46
    David Hand
    Data-driven͸Theory-drivenͱߟ͑ํɾ໾ׂ͕ҟͳΔͷͰ஫ҙ
    All models are wrong, but some are useful
    (George Box)
    Theory-driven models can be wrong
    But data-driven models cannot be wrong
    or right
    Data-driven are not trying to describe an underlying
    reality.
    so they could be poor or useless, but not
    wrong
    But are merely intended to be useful
    http://videolectures.net/kdd2018_hand_data_science/

    View full-size slide

  49. With enough data, the
    numbers speak for
    themselves.
    Chris Anderson (2008)
    cf.

    View full-size slide

  50. ʮཧղʯฤ: ·ͱΊ
    48
    σʔλͷ૬ؔؔ܎ͷར׆༻ٕज़Ͱ͋Δػցֶश͚ͩͰ͸

    ର৅ݱ৅ͷഎޙʹ͋Δ࢓૊ΈΛཧղ͢Δͷ͸ݪཧ্ࠔ೉
    • ૬ؔؔ܎͸ඞͣ͠΋ҼՌؔ܎Λҙຯ͠ͳ͍
    • ҼՌͷݕূʹ͸؍࡯ݚڀͰ͸ͳ͘հೖݚڀ͕ඞཁ
    • ҩྍ΍೴ՊֶͳͲྙཧతʹհೖݚڀ͕೉͍͠৔߹΋ଟ͘
    ҼՌਪ࿦ͷཧ࿦ɾख๏͸௕Β͘ݚڀ͞Ε͖͍ͯͯΔ
    • ҼՌਪ࿦Ͱ͸ؔ࿈Ҽࢠ΍ҼՌߏ଄͕͢΂ͯ෼͔͍ͬͯΔ
    ͳͲͷݱ࣮తʹ͸೉͍͠લఏ͕ຬͨ͞ΕΔඞཁ͕͋Δ
    • ૬ؔؔ܎͸ҼՌͷࣔࠦͰ͸͋ΔͷͰ஫ҙਂ͘ߟ͑Α͏ʂ

    View full-size slide

  51. ࠓ೔ͷ಺༰
    49
    1. Πϯτϩ

    ػցֶशͱՊֶ(͋Δ͍͸"΋ͷͮ͘Γ")
    2. ػցֶशͰԿ͔ΛʮཧղʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    3. ػցֶशͰԿ͔ΛʮൃݟʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    4. ͡Ό͋Ͳ͏͢ΜͷʂʁԿ͕͍Δͷʂʁ

    Answer:ʮදݱʯͱʮհೖʯ

    2ͱ3Λલఏʹػցֶश෼໺ͷτϐοΫΛ؆୯ʹ঺հ

    View full-size slide

  52. ʮൃݟʯ͸ֶशͷൃలܥʁ
    50
    ൃݟ = ͍··Ͱʹͳ͍΋ͷɾ͜ͱΛݟ͚ͭΔ
    • ࠓ·Ͱʹͳ͍ըظతͳ৽ༀ
    • ࠓ·ͰͷσʔλͷͲΕΑΓ΋௕࣋ͪ͢Δి஑ࡐྉ
    • ࠓ·Ͱ୭΋ࢼ͞ͳ͔ͬͨըظతͳձࣾܦӦઓུ
    • ࠓ·ͰະൃݟͩͬͨըظతͳՊֶ๏ଇ΍Պֶཧ࿦
    • ࠓ·Ͱରઓͨ͠୭ΑΓ΋ڧ͍ϘʔυήʔϜউརઓུ
    ൃݟ͸;ͭ͏׬શʹߦ͖౰ͨΓ͹ͬͨΓͰ͸ͳ͍ɻ

    ʮצͱܦݧʯ͕ඇৗʹେ੾ ʮ޾ӡ͸४උ͞Εͨऀʹ߱ΓΔʯ
    ܦݧ(աڈͷσʔλ)͔Βֶशͨ͠צ(๏ଇੑ) = ػցֶश

    ͱߟ͑ΔͱͳΜ͔ͩΠέͦ͏ͳؾ͕͢Δʙʁ

    View full-size slide

  53. Ұํɺػցֶशͱ͸Կ͔ͩͬͨ
    51
    Ұൠ෺ମೝࣝ
    ήʔϜϓϨΠ
    “͋Γ͕ͱ͏”
    J’aime
    la
    musiqu
    e
    I love music
    ೖग़ྗͷؔ܎͕Α͘෼͔Βͳ͍ม׵աఔ(ؔ਺)Λେྔͷೖग़ྗͷ
    ݟຊྫ͔Β໌ࣔతʹϓϩάϥϛϯά͢Δ͜ͱͳ͘ߏ੒͢Δٕ๏
    Ի੠ೝࣝ
    ػց຋༁
    ௒ղ૾

    View full-size slide

  54. ػցֶशͷ࢓૊Έ = ߴ࣍ݩͰͷۂ໘͋ͯ͸Ί
    52
    ίϯϐϡʔλ
    ϓϩάϥϜ
    ೖྗ
    ग़ྗ
    ೖग़ྗͷݟຊ
    ڭࢣσʔλ

    ೖग़ྗͷݟຊ {(x1, y1), (x2, y2), . . . , (xn, yn)}
    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
    Ұൠʹ͸ߴ࣍ݩ

    View full-size slide

  55. ػցֶशͷ࢓૊Έ = ߴ࣍ݩͰͷۂ໘͋ͯ͸Ί
    52
    ίϯϐϡʔλ
    ϓϩάϥϜ
    ೖྗ
    ग़ྗ
    ೖग़ྗͷݟຊ
    ڭࢣσʔλ

    ೖग़ྗͷݟຊ {(x1, y1), (x2, y2), . . . , (xn, yn)}
    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
    Ұൠʹ͸ߴ࣍ݩ

    View full-size slide

  56. ػցֶशͷ࢓૊Έ = ߴ࣍ݩͰͷۂ໘͋ͯ͸Ί
    52
    ίϯϐϡʔλ
    ϓϩάϥϜ
    ೖྗ
    ग़ྗ
    ೖग़ྗͷݟຊ
    ڭࢣσʔλ

    ิؒ
    ೖग़ྗͷݟຊ {(x1, y1), (x2, y2), . . . , (xn, yn)}
    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
    Ұൠʹ͸ߴ࣍ݩ

    View full-size slide

  57. ػցֶशͷ࢓૊Έ = ߴ࣍ݩͰͷۂ໘͋ͯ͸Ί
    53
    Inputs Outputs
    ML model
    x
    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
    y
    AAACq3ichVG7SgNBFD1ZX/EdtRFsxBCxMUxUUKxEG0tNzAMTkd11Ekf3xe4kEEN+wMpO1ErBQvwMG3/Awk8Qywg2Ft7dLIgG9S67c+bce+6c2as5hvAkY88Rpau7p7cv2j8wODQ8MhobG895dtXVeVa3DdstaKrHDWHxrBTS4AXH5aqpGTyvHW/4+XyNu56wrR1Zd/ieqVYsURa6KonKlTSzUW/ux+IsyYKY7gSpEMQRxpYde0QJB7ChowoTHBYkYQMqPHqKSIHBIW4PDeJcQiLIczQxQNoqVXGqUIk9pm+FdsWQtWjv9/QCtU6nGPS6pJxGgj2xO9Zij+yevbCPX3s1gh6+lzqtWlvLnf3R08nM+78qk1aJwy/Vn54lylgJvAry7gSMfwu9ra+dnLcyq+lEY5bdsFfyf82e2QPdwKq96bfbPH31hx+NvPz+x/x8WEEjTP0cWCfILSRTi8mF7aX42no4zCimMIM5mtgy1rCJLWTphCOc4QKXyrySUXaVUrtUiYSaCXwLhX8CgAGZrA==
    A function best fitted to
    a given set of example input-output pairs
    (the training data).
    (x1, y1), (x2, y2), . . . , (xn, yn)
    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
    f(x; ✓)
    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
    x
    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
    y
    AAACq3ichVG7SgNBFD1ZX/EdtRFsxBCxMUxUUKxEG0tNzAMTkd11Ekf3xe4kEEN+wMpO1ErBQvwMG3/Awk8Qywg2Ft7dLIgG9S67c+bce+6c2as5hvAkY88Rpau7p7cv2j8wODQ8MhobG895dtXVeVa3DdstaKrHDWHxrBTS4AXH5aqpGTyvHW/4+XyNu56wrR1Zd/ieqVYsURa6KonKlTSzUW/ux+IsyYKY7gSpEMQRxpYde0QJB7ChowoTHBYkYQMqPHqKSIHBIW4PDeJcQiLIczQxQNoqVXGqUIk9pm+FdsWQtWjv9/QCtU6nGPS6pJxGgj2xO9Zij+yevbCPX3s1gh6+lzqtWlvLnf3R08nM+78qk1aJwy/Vn54lylgJvAry7gSMfwu9ra+dnLcyq+lEY5bdsFfyf82e2QPdwKq96bfbPH31hx+NvPz+x/x8WEEjTP0cWCfILSRTi8mF7aX42no4zCimMIM5mtgy1rCJLWTphCOc4QKXyrySUXaVUrtUiYSaCXwLhX8CgAGZrA==

    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
    x
    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
    y
    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
    x
    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
    y
    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
    x
    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
    y
    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
    interpolative
    prediction
    (High-dimensional)
    High
    Low Model Complexity
    Underfitting
    (High bias, Low variance)
    Overfitting
    (Low bias, High variance)
    "The bias-variance tradeoff"
    The training data
    f(x; ✓)
    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
    extrapolative
    prediction

    View full-size slide

  58. ஫ҙɿػցֶश͸ʮൃݟʯʹ޲͍͍ͯͳ͍
    54
    ػցֶश = ܇࿅σʔλͷฏۉత๏ଇੑΛͱΒ͑Δ
    ൃݟ = ݟຊσʔλͷதʹͳ͍΋ͷΛݟ͚͍ͭͨ
    ༧ଌϞσϧͱͷޡࠩͷʮظ଴஋ʯΛ࠷খԽ = ൚Խ
    ೖྗ (Ұൠʹ͸ߴ࣍ݩ)
    ग़ྗ
    x
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    ฏۉతڍಈ
    ͷϞσϧ
    ܇࿅σʔλͷ࠷େҎ্ͷ
    ܇࿅σʔλͷ࠷େ
    (ظ଴ޡࠩ࠷খԽ)
    ໨త෺͕֎Ε஋(֎ૠత)Ͱ

    ෼෍ͷ੄ʹདྷͯ͠·͏
    ໨త͕

    ෆ੔߹
    ʮ֎Ε஋ʯ
    ฏۉత(ຌ༱)ͳͷ͸
    Α͘౰ͨΔ...

    View full-size slide

  59. ػցֶश͸༩͑ͨ܇࿅σʔλΛ୅ද͢Δ͚ͩ
    55
    Highly Inaccurate Model Predictions
    from Extrapolation (Lohninger 1999)
    ༩͑ͨσʔλͷ܏޲Λ(ۂઢ͋ͯ͸ΊͰ)
    ද͚ͩ͢Ͱσʔλ͕ͳ͍֎ૠྖҬͰ͸
    ແࠜڌͳ༧ଌΛฦ͢
    ར༻ "exploitation"
    ୳ࡧ "exploration"
    ৽͍͠஌ࣝ/σʔλΛ֫ಘ ֫ಘͨ͠஌ࣝ/σʔλΛར༻
    τϨʔυΦϑ

    View full-size slide

  60. հೖ࣮ݧͷܭըɿ࣮ݧܭը๏ͱԠ౴ۂ໘๏
    56
    ౰વBox͸ͲͷΑ͏ʹճؼ෼ੳΛ࢖͑͹ྑ͍͔୳ٻࡁΈʂ☺
    Empirical Model-Building and
    Response Surfaces (1987)
    Statistics for Experimenters: Design,
    Innovation, and Discovery (2005)

    View full-size slide

  61. Ԡ౴ۂ໘๏ (Box & Wilson, 1951)
    ม਺ͷ਺͕গͳ͘౷ܭֶతͳԾఆ͕͋Δఔ౓༗ޮͳΒ͜ΕͰOKɻ
    ୳ࡧۭؒ(ؔ৺ྖҬ)ͷ಺ૠʹͳΔΑ͏࣮ݧܭըͰࣄྫ఺ΛಘΔ
    1. Ԡ౴ۂ໘(Response Surface)ΛϞσϧԽ

    (e.g. ೋ࣍ଟ߲ࣜճؼ)
    2. ্هϞσϧΛ౰ͯ͸ΊΔͨΊͷ࣮ݧܭը(e.g. த৺ෳ߹ܭը)Ͱ
    ݕࠪ఺ΛಘΔ
    3. Ԡ౴ۂ໘Λݕࠪ఺ʹ౰ͯ͸ΊͦΕ͕࠷େʹͳΔ఺ΛٻΊΔ
    ͔͠͠ɺݱ୅ͷ࣮໰୊ͷσʔλ͸...
    Box-WilsonͷԠ౴ۂ໘๏

    View full-size slide

  62. ೖྗ͕ଟ༷ + গ͠ͷมԽͰग़ྗ͕มΘΓಘΔ
    58
    J. Med. Chem. 2012, 55, 2932−2942
    ෺ੑ΍׆ੑͷϥϯυεέʔϓ͸ඇฏ׈త

    (গ͠ͷߏ଄มԽ͕ٸफ़ͳӨڹΛ΋ͨΒ
    ͢)
    Activity
    cliffs
    Selectivity
    cliffs

    View full-size slide

  63. ͞Βʹߴ࣍ݩͰ͸֎ૠ͔಺ૠ͔ͷ൑ఆ͢Β೉͍͠...
    59
    ߴ࣍ݩۭؒ͸ඇ௚ײతͳੑ࣭Λ࣋ͭ
    • ِ૬ؔɿม਺ͷ਺͕͋·Γʹଟ͍ͱ܇࿅σʔλ͢΂ͯΛ

    ͱ͓Δۂ໘͕ࣗ༝ʹ࡞Εͯ͠·ِ͍૬͕ؔੜ͡΍͘͢ͳΔ
    • ଌ౓ͷूதݱ৅ɿݟຊ఺ͷؒͷڑ཭͕શͯ΄΅ಉ͡ʹͳΔ
    ༩͑ΒΕͨ܇࿅σʔλ ಺ૠ or ֎ૠʁ

    View full-size slide

  64. ੒ޭྫ΋಺ૠతͩͱ௚ײ͢Δͷ͸ඇৗʹࠔ೉
    60
    QJYQJY
    $ZDMF("/
    :0-0
    FH%FFQ'BLF

    View full-size slide

  65. ੒ޭྫ΋಺ૠతͩͱ௚ײ͢Δͷ͸ඇৗʹࠔ೉
    60
    QJYQJY
    $ZDMF("/
    :0-0
    FH%FFQ'BLF

    View full-size slide

  66. ͋ͯ͸ΊΔۂ໘ͷ΄͏΋ߴ࣍ݩͰ͸ඇ௚ײత
    61
    http://www.evolvingai.org/fooling
    https://arxiv.org/pdf/
    1610.06940.pdf
    https://towardsdatascience.com/know-your-adversary-
    understanding-adversarial-examples-part-1-2-63af4c2f5830
    e.g. Adversarial examples (GANͷൃ૝)ɺ಺ૠ or ֎ૠͷ൑ఆͷ೉͠͞

    View full-size slide

  67. ʮൃݟʯฤ: ·ͱΊ
    62
    σʔλͷۂ໘͋ͯ͸ΊʹΑΔ಺ૠͰ͋Δػցֶश͚ͩͰ͸

    ܇࿅σʔλʹશ͘ͳ͍৽نൃݟΛ͢Δͷ͸ݪཧ্ࠔ೉
    • ػցֶश͸ۂ໘͋ͯ͸ΊͰ܇࿅σʔλΛ୅ද͢Δ͚ͩ
    • ۂ໘͸ݟຊσʔλʹ͋͏Α͏ϑΟοτ͞ΕΔͷͰ

    ֎ૠతͳτϨϯυ͸༧ଌࠜڌ͕͖ΘΊͯബ͘ͳΔ
    • ൃݟ(஌ࣝ֫ಘ)ʹ͸஌ࣝͷར༻ͱ୳ࡧͷτϨʔυΦϑͷ
    ߟྀ͕ඞਢ
    • ࠷ۙͷσʔλ͸ෳࡶͰଟ༷Ͱ੍ޚ͞Ε͍ͯͳ͍ͷͰݹయ
    తͳ࣮ݧܭը΍ۂ໘Ԡ౴ํͰ͸ͳ͔ͳ͔े෼Ͱ͸ͳ͍

    View full-size slide

  68. ࠓ೔ͷ಺༰
    63
    1. Πϯτϩ

    ػցֶशͱՊֶ(͋Δ͍͸"΋ͷͮ͘Γ")
    2. ػցֶशͰԿ͔ΛʮཧղʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    3. ػցֶशͰԿ͔ΛʮൃݟʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    4. ͡Ό͋Ͳ͏͢ΜͷʂʁԿ͕͍Δͷʂʁ

    Answer:ʮදݱʯͱʮհೖʯ

    2ͱ3Λલఏʹػցֶश෼໺ͷτϐοΫΛ؆୯ʹ঺հ

    View full-size slide

  69. Պֶతʮཧղɾൃݟʯʹඞཁͳೋେཁૉ?
    64
    ʮදݱʯ
    ʮհೖʯ
    • ର৅ΛͲ͏දݱ͢Δ͔ʁԿΛଌΔ͔ʁ
    • ໰୊Λ಺ૠతʹ͢Δදݱͷֶश

    (͍·ͷͱ͜ΖઃܭʹཁυϝΠϯ஌ࣝ)
    • എܠաఔʹ͍ͭͯ෼͔͍ͬͯΔ

    ͜ͱͷ൓ө΍׆༻ (ؼೲόΠΞε)
    • ػցֶशʹʮ࣮ࡍʹ௥ՃσʔλΛ
    औΓʹߦ͘ʯ࢓૊ΈΛ༥߹
    • ࣍ʹԿΛ࣮ݧ͢Δ͔ͷ࠷దܭը
    Պֶతʮཧղʯ΍ʮൃݟʯͱ͸Կ͔(ԿͰ͋Δ΂͖͔)͸

    Պֶ఩ֶͷ໰୊

    View full-size slide

  70. 65
    ػցֶशͰ͖ͨͱ͖ͷ̎छྨͷظ଴
    1. ಘΒΕͨม׵աఔ(ؔ਺)ʹΑΔ༧ଌΛ৭ʑͳ໨తʹ࢖͏
    2. ಘΒΕͨม׵աఔ(ؔ਺)Λ෼ੳͯ͠എܠաఔͷ࢓૊ΈΛ஌Δ
    1.ʹ͍ͭͯ
    • ໰୊͕಺ૠతʹͳΔΑ͏޻෉ (දݱֶशɾӅΕߏ଄ಉఆ)
    • ಺ૠɾ֎ૠ൑ఆ (༧ଌͷ৴པ౓ܭࢉ)
    • Ϟσϧϕʔε࠷దԽͱ୳ࡧ (࠷ద࣮ݧܭը)
    2.ʹ͍ͭͯ
    • ϙετϗοΫղੳͱղऍੑϞσϧ (ֶशࡁΈϞσϧ෼ੳ)

    View full-size slide

  71. 65
    ػցֶशͰ͖ͨͱ͖ͷ̎छྨͷظ଴
    1. ಘΒΕͨม׵աఔ(ؔ਺)ʹΑΔ༧ଌΛ৭ʑͳ໨తʹ࢖͏
    2. ಘΒΕͨม׵աఔ(ؔ਺)Λ෼ੳͯ͠എܠաఔͷ࢓૊ΈΛ஌Δ
    1.ʹ͍ͭͯ
    • ໰୊͕಺ૠతʹͳΔΑ͏޻෉ (දݱֶशɾӅΕߏ଄ಉఆ)
    • ಺ૠɾ֎ૠ൑ఆ (༧ଌͷ৴པ౓ܭࢉ)
    • Ϟσϧϕʔε࠷దԽͱ୳ࡧ (࠷ద࣮ݧܭը)
    2.ʹ͍ͭͯ
    • ϙετϗοΫղੳͱղऍੑϞσϧ (ֶशࡁΈϞσϧ෼ੳ)

    View full-size slide

  72. ਂ૚ֶशʹΑΔදݱֶश
    66
    ೖྗͷʮଟ࣍ݩͷ਺஋૊(ϕΫτϧ)ʯΛগͮͭ͠
    ผͷʮଟ࣍ݩͷ਺஋૊(ϕΫτϧ)ʯ΁ม׵͢Δϓϩηε
    ม׵͞Εͨ࠷ऴྔ
    ʹ͍ͭͯ༧ଌ
    End

    (खʹೖΔ··)
    End
    (๬Ήग़ྗ)
    End to End (׬શͳσʔλۦಈ)
    Raw Pixel Values

    (RGB Color Image)

    .01 Bill
    .96 Ichi
    .01 Jeff
    .02 Larry

































    ೖྗม਺ͷஈ֊Ͱ͸
    ಺ૠత͡Όͳͯ͘΋...
    ༧ଌʹ࢖͏தؒදݱ(ӅΕ

    ߏ଄)Ͱ಺ૠతͰ͋Ε͹OK!

    View full-size slide

  73. ྫ) ਂ૚ֶश͸ೖྗͷଟஈͷม׵ϓϩηεΛֶश
    67
    IUUQTDPMBIHJUIVCJPQPTUT//.BOJGPMET5PQPMPHZ
    ઢܗ෼཭Ͱ͖ΔΑ͏ͳม׵Λֶश ࣦഊྫ(ৗʹ੒ޭ͢Δͱ͸ݶΒͳ͍)

    View full-size slide

  74. ྫ) ਂ૚ֶश͸ೖྗͷଟஈͷม׵ϓϩηεΛֶश
    67
    IUUQTDPMBIHJUIVCJPQPTUT//.BOJGPMET5PQPMPHZ
    ઢܗ෼཭Ͱ͖ΔΑ͏ͳม׵Λֶश ࣦഊྫ(ৗʹ੒ޭ͢Δͱ͸ݶΒͳ͍)

    View full-size slide

  75. ྫ) ਂ૚ֶश͸ೖྗͷଟஈͷม׵ϓϩηεΛֶश
    67
    IUUQTDPMBIHJUIVCJPQPTUT//.BOJGPMET5PQPMPHZ
    ઢܗ෼཭Ͱ͖ΔΑ͏ͳม׵Λֶश ࣦഊྫ(ৗʹ੒ޭ͢Δͱ͸ݶΒͳ͍)

    View full-size slide

  76. ྫ) ਂ૚ֶश͸ೖྗͷଟஈͷม׵ϓϩηεΛֶश
    67
    IUUQTDPMBIHJUIVCJPQPTUT//.BOJGPMET5PQPMPHZ
    ઢܗ෼཭Ͱ͖ΔΑ͏ͳม׵Λֶश ࣦഊྫ(ৗʹ੒ޭ͢Δͱ͸ݶΒͳ͍)

    View full-size slide

  77. Ϟσϧߏ଄΁ͷυϝΠϯ஌ࣝͷΤϯίʔυ
    68
    Model
    https://www.kaggle.com/c/recursion-cellular-image-classification/discussion/
    110543#latest-637352
    ͱʹ͔͘ඍ෼Մೳͳԋࢉͷ(ਂ͍)߹੒ؔ਺ʹ͑͞ͳ͍ͬͯΕ͹

    Ϟσϧύϥϝλ͸ࣗಈඍ෼(aka backprop)ͰֶशͰ͖ΔͷͰ

    എܠաఔΛཧղͦ͠ΕʹͦͬͯॊೈʹϞσϧߏ଄Λઃܭ
    Head
    Backbone Neck
    ࣄલֶश͔ΒͷసҠ΋༗ޮ

    View full-size slide

  78. ࣄલֶश͕ޮ͔ͳ͍ͱ͞ΕͨݴޠλεΫͰ΋...
    69
    ͔ͭͯݴޠλεΫͰ͸RNN(LSTM/GRU)→CNNͷྲྀΕ͕ͩͬͨ...
    GoogleͷݴޠϞσϧBERT
    OpenAIͷݴޠϞσϧGPT-2
    GLUEϕϯνϚʔΫͷશݴޠཧղλεΫͰͿͬͪ͗ΓͷSOTAʂ
    ࣭ٙԠ౴λεΫͷSQuADͰ΋SOTAʂ
    CMUͷXLNet
    MicrosoftͷݴޠϞσϧMT-DNN
    2018/10/18
    2019/01/31
    ࡞จੑೳ͕ߴ͗ͯ͢Φʔϓϯιʔεͱͯ͠ެ։ͯ͠͠·͏ͱϑΣΠΫχϡʔε
    ͕࡞Γ์୊ʹͳͬͯ͠·͏ݒ೦͔Βݚڀऀ޲͚ʹن໛ॖখ൛ͷΈΛެ։
    2019/02/14
    2019/06/19
    Googleͷ՚ྷͳ࿦จΛܖػʹRNN΍
    CNNΑΓTransformer͕ࢧ഑తʹ!?
    ௒ڊେͳࣄલֶशϞσϧͷ

    (Attentiveͳ)ʮసҠֶशʯ΁

    View full-size slide

  79. ֎ૠతྖҬͰͷػցֶशͷ׆༻
    70
    ݟຊ఺͕ͳ͍֎ૠతྖҬͰػցֶश(ۂ໘͋ͯ͸Ί)Λ׆༻͢Δɻ
    • ྨࣅͨ͠໰୊ɾσʔλͰͷؔ܎ੑΛԣஅతʹ׆༻ͯ͠ྨਪ 

    (సҠֶशɺ൒ڭࢣֶ͖ͭशɺϚϧνλεΫֶशɺ஫ࢹతֶश)
    • ۙ͞ͷଌΓํΛద੾ʹֶशͯ͠໰୊Λ಺ૠతʹ (ܭྔֶश)
    • എܠաఔͷୈҰݪཧϞσϧ(γϛϡϨʔγϣϯ)Λෆ࣮֬ͳҼࢠ

    ΛؚΊཱͯͯͦͷ࠷దͳਪఆ஋Λػցֶश͢Δ (σʔλಉԽ)
    • γϛϡϨʔγϣϯσʔλ΍ܦݧऀͷڭࣔΛ༻͍ͯσʔλΛ

    ૿΍͠Ͱ͖Δ͚ͩ໰୊Λ಺ૠతʹ (ఢରతੜ੒ɺ໛฿ֶश)
    • γϛϡϨʔγϣϯ݁Ռ͔Β࣮ݱ৅ͷΪϟοϓΛػցֶश͢Δ

    (సҠֶशɺϝλֶश)

    View full-size slide

  80. 71
    ػցֶशͰ͖ͨͱ͖ͷ̎छྨͷظ଴
    1. ಘΒΕͨม׵աఔ(ؔ਺)ʹΑΔ༧ଌΛ৭ʑͳ໨తʹ࢖͏
    2. ಘΒΕͨม׵աఔ(ؔ਺)Λ෼ੳͯ͠എܠաఔͷ࢓૊ΈΛ஌Δ
    1.ʹ͍ͭͯ
    • ໰୊͕಺ૠతʹͳΔΑ͏޻෉ (දݱֶशɾӅΕߏ଄ಉఆ)
    • ಺ૠɾ֎ૠ൑ఆ (༧ଌͷ৴པ౓ܭࢉ)
    • Ϟσϧϕʔε࠷దԽͱ୳ࡧ (࠷ద࣮ݧܭը)
    2.ʹ͍ͭͯ
    • ϙετϗοΫղੳͱղऍੑϞσϧ (ֶशࡁΈϞσϧ෼ੳ)

    View full-size slide

  81. ֎ૠ൑ఆ΋͘͠͸৴པྖҬਪఆ
    72
    • ύλʔϯೝࣝʹ͓͚Δغ٫Φϓγϣϯ
    • ৴པ۠ؒ(Confidence Interval)
    • ৴པྖҬ(Trust Region)
    • ϕΠζʹ͓͚Δ༧ଌ෼෍
    • Cheminfoʹ͓͚ΔApplicability Domain(AD)
    ࣗൢػ
    χ
    η

    ☺Պֶͱػցֶशͷ͍͋ͩɿมྔͷઃܭɾม׵ɾબ୒ɾަޓ࡞༻ɾઢܗੑ

    https://www.slideshare.net/itakigawa/ss-69269618
    ༧ଌ͍ͨ͠ݕࠪೖྗ఺ͷۙ͘ʹݟຊ఺͕શવͳ͍(or ඇৗʹ
    গͳ͍)৔߹͸جຊతʹ֎ૠత

    View full-size slide

  82. 73
    ػցֶशͰ͖ͨͱ͖ͷ̎छྨͷظ଴
    1. ಘΒΕͨม׵աఔ(ؔ਺)ʹΑΔ༧ଌΛ৭ʑͳ໨తʹ࢖͏
    2. ಘΒΕͨม׵աఔ(ؔ਺)Λ෼ੳͯ͠എܠաఔͷ࢓૊ΈΛ஌Δ
    1.ʹ͍ͭͯ
    • ໰୊͕಺ૠతʹͳΔΑ͏޻෉ (දݱֶशɾӅΕߏ଄ಉఆ)
    • ಺ૠɾ֎ૠ൑ఆ (༧ଌͷ৴པ౓ܭࢉ)
    • Ϟσϧϕʔε࠷దԽͱ୳ࡧ (࠷ద࣮ݧܭը)
    2.ʹ͍ͭͯ
    • ϙετϗοΫղੳͱղऍੑϞσϧ (ֶशࡁΈϞσϧ෼ੳ)

    View full-size slide

  83. ػցֶशͷར׆༻ʹΑΔ࠷ద࣮ݧܭըʁ
    74
    ط஌ͷ஌ݟɾ
    ؍ଌ(σʔλ)
    ݁Ռͷ֬ೝͱ
    ݕূ
    ߴ଎ɾߴਫ਼౓ͳ
    Data-Driven༧ଌ
    ࣍ͷ࣮ݧܭը΁feedback
    Ծઆܗ੒
    (γϛϡϨʔγϣϯ+࣮ݧ)
    • ࠶ݱੑΛ୲อ͢Δߴਫ਼౓ɾߴ଎࣮ݧܥ
    • Ծ૝Խݕূ͕ՄೳͳҼࢠͷγϛϡϨʔ
    γϣϯ(ܭࢉՊֶ)ʹΑΔ୳ࡧ

    → ๬·͍͠ର৅ͷ͞ΒͳΔߜΓࠐΈ
    Ծઆݕূ
    (ػցֶशɾσʔλϚΠχϯά)
    • Ͳ͏͍͏࣮ݧɾγϛϡϨʔγϣϯΛ

    ࣍ʹߦ͏͔ͷܭըཱҊ
    • ࣌ؒͷ͔͔Δܭࢉͷߴਫ਼౓ߴ଎ۙࣅ
    • ᐆດͳҼࢠ΍࣮ݧ৚݅ͷ࠷దԽ
    • Multilevelͷ৘ใ౷߹

    View full-size slide

  84. 2000-2010೥ࠒ͔Β૑ༀ/ੜ໋ՊֶͰઌߦ
    75

    View full-size slide

  85. ࡐྉ։ൃ΁΋ల։
    76
    ݚڀ։ൃମ੍
    ੜ࢈ମ੍
    ࡐྉ։ൃɺηϧσβΠϯʹ͓͍ͯ

    ࠷ઌ୺γϛϡϨʔγϣϯٕज़Λ׆༻
    ׬શࣗಈԽʹΑΔϑϨΩγϒϧͳੜ࢈

    IoT΍ϏοάσʔλΛ׆༻ͨ͠ੜ࢈؅ཧ
    ↓඼࣭Λ୲อ͢Δେن໛Խͷٕज़తඞਢཁૉɿ
    ੡଄ϥΠϯʹਓ͕(΄ͱΜͲ)͍ͳ͍

    View full-size slide

  86. ࡐྉ։ൃ΁΋ల։
    76
    ݚڀ։ൃମ੍
    ੜ࢈ମ੍
    ࡐྉ։ൃɺηϧσβΠϯʹ͓͍ͯ

    ࠷ઌ୺γϛϡϨʔγϣϯٕज़Λ׆༻
    ׬શࣗಈԽʹΑΔϑϨΩγϒϧͳੜ࢈

    IoT΍ϏοάσʔλΛ׆༻ͨ͠ੜ࢈؅ཧ
    ↓඼࣭Λ୲อ͢Δେن໛Խͷٕज़తඞਢཁૉɿ
    ੡଄ϥΠϯʹਓ͕(΄ͱΜͲ)͍ͳ͍

    View full-size slide

  87. Խֶʹ΋೾ٴ
    77

    View full-size slide

  88. Keyɿ஝ੵ͞Εͨʮܭࢉɾ࣮ݧσʔλʯͷར׆༻
    78
    (ػցֶशɾσʔλϚΠχϯά)
    Ծઆܗ੒ ݕূ
    (γϛϡϨʔγϣϯ+࣮ݧ)
    • ࠶ݱੑΛ୲อ͢Δߴਫ਼౓ɾߴ଎࣮ݧܥ
    • Ծ૝Խݕূ͕ՄೳͳҼࢠͷγϛϡϨʔ
    γϣϯ(ܭࢉՊֶ)ʹΑΔ୳ࡧ

    → ๬·͍͠ର৅ͷ͞ΒͳΔߜΓࠐΈ
    ࣮ݧσʔλɾܭࢉσʔλɾϑΝΫτͷ஝ੵ
    In-Houseσʔλ + Publicσʔλ + ஌ࣝϕʔε 

    + ͦͷQuality Control / Annotations
    • Ͳ͏͍͏࣮ݧɾγϛϡϨʔγϣϯΛ

    ࣍ʹߦ͏͔ͷܭըཱҊ
    • ࣌ؒͷ͔͔Δܭࢉͷߴਫ਼౓ߴ଎ۙࣅ
    • ᐆດͳҼࢠ΍࣮ݧ৚݅ͷ࠷దԽ
    • Multilevelͷ৘ใ౷߹

    View full-size slide

  89. ʮσʔλར׆༻ٕज़ʯ͸Պֶݚڀͷಓ۩ͷҰͭʹ
    79
    REVIEW
    Inverse molecular design using
    machine learning: Generative models
    for matter engineering
    Benjamin Sanchez-Lengeling1 and Alán Aspuru-Guzik2,3,4*
    The discovery of new materials can bring enormous societal and technological progress. In this
    context, exploring completely the large space of potential materials is computationally
    intractable. Here, we review methods for achieving inverse design, which aims to discover
    tailored materials from the starting point of a particular desired functionality. Recent advances
    from the rapidly growing field of artificial intelligence, mostly from the subfield of machine
    learning, have resulted in a fertile exchange of ideas, where approaches to inverse molecular
    design are being proposed and employed at a rapid pace. Among these, deep generative models
    have been applied to numerous classes of materials: rational design of prospective drugs,
    synthetic routes to organic compounds, and optimization of photovoltaics and redox flow
    batteries, as well as a variety of other solid-state materials.
    Many of the challenges of the 21st century
    (1), from personalized health care to
    energy production and storage, share a
    common theme: materials are part of
    the solution (2). In some cases, the solu-
    tions to these challenges are fundamentally
    limited by the physics and chemistry of a ma-
    terial, such as the relationship of a materials
    bandgap to the thermodynamic limits for the
    generation of solar energy (3).
    Several important materials discoveries arose
    by chance or through a process of trial and error.
    For example, vulcanized rubber was prepared in
    the 19th century from random mixtures of com-
    pounds, based on the observation that heating
    with additives such as sulfur improved the
    rubber’s durability. At the molecular level, in-
    dividual polymer chains cross-linked, forming
    bridges that enhanced the macroscopic mechan-
    ical properties (4). Other notable examples in
    this vein include Teflon, anesthesia, Vaseline,
    Perkin’s mauve, and penicillin. Furthermore,
    these materials come from common chemical
    compounds found in nature. Potential drugs
    either were prepared by synthesis in a chem-
    ical laboratory or were isolated from plants,
    soil bacteria, or fungus. For example, up until
    2014, 49% of small-molecule cancer drugs were
    natural products or their derivatives (5).
    In the future, disruptive advances in the dis-
    covery of matter could instead come from unex-
    plored regions of the set of all possible molecular
    and solid-state compounds, known as chemical
    space (6, 7). One of the largest collections of
    molecules, the chemical space project (8), has
    mapped 166.4 billion molecules that contain at
    most 17 heavy atoms. For pharmacologically rele-
    vant small molecules, the number of structures is
    estimated to be on the order of 1060 (9). Adding
    consideration of the hierarchy of scale from sub-
    nanometer to microscopic and mesoscopic fur-
    ther complicates exploration of chemical space
    in its entirety (10). Therefore, any global strategy
    for covering this space might seem impossible.
    Simulation offers one way of probing this
    space without experimentation. The physics
    and chemistry of these molecules are governed
    by quantum mechanics, which can be solved via
    the Schrödinger equation to arrive at their ex-
    act properties. In practice, approximations are
    used to lower computational time at the cost of
    accuracy.
    Although theory enjoys enormous progress,
    now routinely modeling molecules, clusters, and
    perfect as well as defect-laden periodic solids, the
    size of chemical space is still overwhelming, and
    smart navigation is required. For this purpose,
    machine learning (ML), deep learning (DL), and
    artificial intelligence (AI) have a potential role
    to play because their computational strategies
    automatically improve through experience (11).
    In the context of materials, ML techniques are
    often used for property prediction, seeking to
    learn a function that maps a molecular material
    to the property of choice. Deep generative models
    are a special class of DL methods that seek to
    model the underlying probability distribution of
    both structure and property and relate them in a
    nonlinear way. By exploiting patterns in massive
    datasets, these models can distill average and
    salient features that characterize molecules (12, 13).
    Inverse design is a component of a more
    complex materials discovery process. The time
    scale for deployment of new technologies, from
    discovery in a laboratory to a commercial pro-
    duct, historically, is 15 to 20 years (14). The pro-
    cess (Fig. 1) conventionally involves the following
    steps: (i) generate a new or improved material
    concept and simulate its potential suitability; (ii)
    synthesize the material; (iii) incorporate the ma-
    terial into a device or system; and (iv) characterize
    and measure the desired properties. This cycle
    generates feedback to repeat, improve, and re-
    fine future cycles of discovery. Each step can take
    up to several years.
    In the era of matter engineering, scientists
    seek to accelerate these cycles, reducing the
    FRONTIERS IN COMPUTATION
    1Department of Chemistry and Chemical Biology, Harvard
    University 12 Oxford Street, Cambridge, MA 02138, USA.
    2Department of Chemistry and Department of Computer
    Science, University of Toronto, Toronto Ontario, M5S 3H6,
    Canada. 3Vector Institute for Artificial Intelligence, Toronto,
    Ontario M5S 1M1, Canada. 4Canadian Institute for Advanced
    Fig. 1. Schematic comparison of material discovery paradigms. The current paradigm is
    APTED BY K. HOLOSKI
    on July 26, 2018
    http://science.sciencemag.org/
    Downloaded from
    REVIEW
    https://doi.org/10.1038/s41586-018-0337-2
    Machine learning for molecular and
    materials science
    Keith T. Butler1, Daniel W
    . Davies2, Hugh Cartwright3, Olexandr Isayev4* & Aron Walsh5,6*
    Here we summarize recent progress in machine learning for the chemical sciences. We outline machine-learning
    techniques that are suitable for addressing research questions in this domain, as well as future directions for the field.
    We envisage a future in which the design, synthesis, characterization and application of molecules and materials is
    accelerated by artificial intelligence.
    The Schrödinger equation provides a powerful structure–
    property relationship for molecules and materials. For a given
    spatial arrangement of chemical elements, the distribution of
    electrons and a wide range of physical responses can be described. The
    development of quantum mechanics provided a rigorous theoretical
    foundation for the chemical bond. In 1929, Paul Dirac famously proclaimed
    that the underlying physical laws for the whole of chemistry are “completely
    known”1. John Pople, realizing the importance of rapidly developing
    computer technologies, created a program—Gaussian 70—that could
    perform ab initio calculations: predicting the behaviour, for molecules
    of modest size, purely from the fundamental laws of physics2. In the 1960s,
    the Quantum Chemistry Program Exchange brought quantum chemistry
    to the masses in the form of useful practical tools3. Suddenly, experi-
    mentalists with little or no theoretical training could perform quantum
    calculations too. Using modern algorithms and supercomputers,
    systems containing thousands of interacting ions and electrons can now
    be described using approximations to the physical laws that govern the
    world on the atomic scale4–6.
    The field of computational chemistry has become increasingly pre-
    dictive in the twenty-first century, with activity in applications as wide
    ranging as catalyst development for greenhouse gas conversion, materials
    discovery for energy harvesting and storage, and computer-assisted drug
    design7. The modern chemical-simulation toolkit allows the properties
    of a compound to be anticipated (with reasonable accuracy) before it has
    been made in the laboratory. High-throughput computational screening
    has become routine, giving scientists the ability to calculate the properties
    of thousands of compounds as part of a single study. In particular, den-
    sity functional theory (DFT)8,9, now a mature technique for calculating
    the structure and behaviour of solids10, has enabled the development of
    extensive databases that cover the calculated properties of known and
    hypothetical systems, including organic and inorganic crystals, single
    molecules and metal alloys11–13.
    The emergence of contemporary artificial-intelligence methods has
    the potential to substantially alter and enhance the role of computers in
    science and engineering. The combination of big data and artificial intel-
    ligence has been referred to as both the “fourth paradigm of science”14
    and the “fourth industrial revolution”15, and the number of applications
    in the chemical domain is growing at an astounding rate. A subfield of
    artificial intelligence that has evolved rapidly in recent years is machine
    learning. At the heart of machine-learning applications lie statistical algo-
    rithms whose performance, much like that of a researcher, improves with
    training. There is a growing infrastructure of machine-learning tools for
    generating, testing and refining scientific models. Such techniques are
    suitable for addressing complex problems that involve massive combi-
    natorial spaces or nonlinear processes, which conventional procedures
    either cannot solve or can tackle only at great computational cost.
    As the machinery for artificial intelligence and machine learning
    matures, important advances are being made not only by those in main-
    stream artificial-intelligence research, but also by experts in other fields
    (domain experts) who adopt these approaches for their own purposes. As
    we detail in Box 1, the resources and tools that facilitate the application
    of machine-learning techniques mean that the barrier to entry is lower
    than ever.
    In the rest of this Review, we discuss progress in the application of
    machine learning to address challenges in molecular and materials
    research. We review the basics of machine-learning approaches, iden-
    tify areas in which existing methods have the potential to accelerate
    research and consider the developments that are required to enable more
    wide-ranging impacts.
    Nuts and bolts of machine learning
    With machine learning, given enough data and a rule-discovery algo-
    rithm, a computer has the ability to determine all known physical laws
    (and potentially those that are currently unknown) without human
    input. In traditional computational approaches, the computer is little
    more than a calculator, employing a hard-coded algorithm provided
    by a human expert. By contrast, machine-learning approaches learn
    the rules that underlie a dataset by assessing a portion of that data
    and building a model to make predictions. We consider the basic steps
    involved in the construction of a model, as illustrated in Fig. 1; this
    constitutes a blueprint of the generic workflow that is required for the
    successful application of machine learning in a materials-discovery
    process.
    Data collection
    Machine learning comprises models that learn from existing (train-
    ing) data. Data may require initial preprocessing, during which miss-
    ing or spurious elements are identified and handled. For example, the
    Inorganic Crystal Structure Database (ICSD) currently contains more
    than 190,000 entries, which have been checked for technical mistakes
    but are still subject to human and measurement errors. Identifying
    and removing such errors is essential to avoid machine-learning
    algorithms being misled. There is a growing public concern about
    the lack of reproducibility and error propagation of experimental data
    DNA to be sequences into distinct pieces,
    parcel out the detailed work of sequencing,
    and then reassemble these independent ef-
    forts at the end. It is not quite so simple in the
    world of genome semantics.
    Despite the differences between genome se-
    quencing and genetic network discovery, there
    are clear parallels that are illustrated in Table 1.
    In genome sequencing, a physical map is useful
    to provide scaffolding for assembling the fin-
    ished sequence. In the case of a genetic regula-
    tory network, a graphical model can play the
    same role. A graphical model can represent a
    high-level view of interconnectivity and help
    isolate modules that can be studied indepen-
    dently. Like contigs in a genomic sequencing
    project, low-level functional models can ex-
    plore the detailed behavior of a module of genes
    in a manner that is consistent with the higher
    level graphical model of the system. With stan-
    dardized nomenclature and compatible model-
    ing techniques, independent functional models
    can be assembled into a complete model of the
    cell under study.
    To enable this process, there will need to
    be standardized forms for model representa-
    tion. At present, there are many different
    modeling technologies in use, and although
    models can be easily placed into a database,
    they are not useful out of the context of their
    specific modeling package. The need for a
    standardized way of communicating compu-
    tational descriptions of biological systems ex-
    tends to the literature. Entire conferences
    have been established to explore ways of
    mining the biology literature to extract se-
    mantic information in computational form.
    Going forward, as a community we need
    to come to consensus on how to represent
    what we know about biology in computa-
    tional form as well as in words. The key to
    postgenomic biology will be the computa-
    tional assembly of our collective knowl-
    edge into a cohesive picture of cellular and
    organism function. With such a comprehen-
    sive model, we will be able to explore new
    types of conservation between organisms
    and make great strides toward new thera-
    peutics that function on well-characterized
    pathways.
    References
    1. S. K. Kim et al., Science 293, 2087 (2001).
    2. A. Hartemink et al., paper presented at the Pacific
    Symposium on Biocomputing 2000, Oahu, Hawaii, 4
    to 9 January 2000.
    3. D. Pe’er et al., paper presented at the 9th Conference
    on Intelligent Systems in Molecular Biology (ISMB),
    Copenhagen, Denmark, 21 to 25 July 2001.
    4. H. McAdams, A. Arkin, Proc. Natl. Acad. Sci. U.S.A.
    94, 814 ( 1997 ).
    5. A. J. Hartemink, thesis, Massachusetts Institute of
    Technology, Cambridge (2001).
    V I E W P O I N T
    Machine Learning for Science: State of the
    Art and Future Prospects
    Eric Mjolsness* and Dennis DeCoste
    Recent advances in machine learning methods, along with successful
    applications across a wide variety of fields such as planetary science and
    bioinformatics, promise powerful new tools for practicing scientists. This
    viewpoint highlights some useful characteristics of modern machine learn-
    ing methods and their relevance to scientific applications. We conclude
    with some speculations on near-term progress and promising directions.
    Machine learning (ML) (1) is the study of
    computer algorithms capable of learning to im-
    prove their performance of a task on the basis of
    their own previous experience. The field is
    closely related to pattern recognition and statis-
    tical inference. As an engineering field, ML has
    become steadily more mathematical and more
    successful in applications over the past 20
    years. Learning approaches such as data clus-
    tering, neural network classifiers, and nonlinear
    regression have found surprisingly wide appli-
    cation in the practice of engineering, business,
    and science. A generalized version of the stan-
    dard Hidden Markov Models of ML practice
    have been used for ab initio prediction of gene
    structures in genomic DNA (2). The predictions
    correlate surprisingly well with subsequent
    gene expression analysis (3). Postgenomic bi-
    ology prominently features large-scale gene ex-
    pression data analyzed by clustering methods
    (4), a standard topic in unsupervised learning.
    Many other examples can be given of learning
    and pattern recognition applications in science.
    Where will this trend lead? We believe it will
    lead to appropriate, partial automation of every
    element of scientific method, from hypothesis
    generation to model construction to decisive
    experimentation. Thus, ML has the potential to
    amplify every aspect of a working scientist’s
    progress to understanding. It will also, for better
    or worse, endow intelligent computer systems
    with some of the general analytic power of
    scientific thinking.
    Machine Learning at Every Stage of
    the Scientific Process
    Each scientific field has its own version of the
    scientific process. But the cycle of observing,
    creating hypotheses, testing by decisive exper-
    iment or observation, and iteratively building
    up comprehensive testable models or theories is
    shared across disciplines. For each stage of this
    abstracted scientific process, there are relevant
    developments in ML, statistical inference, and
    pattern recognition that will lead to semiauto-
    matic support tools of unknown but potentially
    broad applicability.
    Increasingly, the early elements of scientific
    method—observation and hypothesis genera-
    tion—face high data volumes, high data acqui-
    sition rates, or requirements for objective anal-
    ysis that cannot be handled by human percep-
    tion alone. This has been the situation in exper-
    imental particle physics for decades. There
    automatic pattern recognition for significant
    events is well developed, including Hough
    transforms, which are foundational in pattern
    recognition. A recent example is event analysis
    for Cherenkov detectors (8) used in neutrino
    oscillation experiments. Microscope imagery in
    cell biology, pathology, petrology, and other
    fields has led to image-processing specialties.
    So has remote sensing from Earth-observing
    satellites, such as the newly operational Terra
    spacecraft with its ASTER (a multispectral
    thermal radiometer), MISR (multiangle imag-
    ing spectral radiometer), MODIS (imaging
    Machine Learning Systems Group, Jet Propulsion Lab-
    oratory/California Institute of Technology, Pasadena,
    CA, 91109, USA.
    *To whom correspondence should be addressed. E-
    mail: [email protected]
    Table 1. Parallels between genome sequencing
    and genetic network discovery.
    Genome
    sequencing
    Genome semantics
    Physical maps Graphical model
    Contigs Low-level functional
    models
    Contig
    reassembly
    Module assembly
    Finished genome
    sequence
    Comprehensive model
    www.sciencemag.org SCIENCE VOL 293 14 SEPTEMBER 2001 2051
    C O M P U T E R S A N D S C I E N C E
    on August 29, 2018
    http://science.sciencemag.org/
    Downloaded from
    Nature, 559

    pp. 547–555 (2018)
    Science, 293
    pp. 2051-2055 (2001)
    Science, 361
    pp. 360-365 (2018)
    Science is changing, the tools of science are
    changing. And that requires different approaches.
    ─── Erich Bloch, 1925-2016
    ҰํͰɺੜ໋ՊֶͰ΋ಘΒΕͨڭ܇ͱͯ͠ɺɺɺ(͓ۚ΋͔͔ΔͷͰɺɺɺ)

    ࣮ޮੑΛͱ΋ͳ͏ํࣜͷཱ֬ʹ͸·ͩ·ͩཁૉٕज़ͷվྑͱʮྑ͍ʯσʔλͷ஝ੵ͕ඞཁ
    "low input, high throughput, no output science." (Sydney Brenner)
    → ࡶͳઃఆɾܥͰ໢ཏతͳϋΠεϧʔϓοτ࣮ݧΛ͍͘Βͯ͠΋Կ΋ಘΒΕͳ͍

    View full-size slide

  90. ʮ(ےͷྑ͍)Ծઆܗ੒ʯͱػցֶशɾσʔλϚΠχϯά
    80
    • ࣗಈԽͷ΋͏ҰͭͷԸܙ͸ɺʮ࠶ݱੑʯʮ݁Ռͷ࣭ʯͷ୲อ

    ଐਓੑ͕࢒͍ͬͯΔͱσʔλͷ࣭ʹ΋(༧ଌʹ΋)͹Β͖͕ͭੜ͡Δ
    • ۙ೥ͷࣗಈԽͰʮߴ଎ʹͰ͖Δ͚ͩͨ͘͞Μࢼ͢ʯͷ͸

    ୳ࡧΛޮ཰Խ͢ΔԦಓ͕ͩɺߟ͑ΒΕΔީิ͕΄΅ແݶʹ͋Γ͑Δ

    ͷͰɺʮԿΛࢼ͔͢ʯͷબ୒ͷ໰୊͸࢒Δ (શ෦͸ࢼͤͳ͍...)
    • ࣗಈԽΛ͢Δ͔͠ͳ͍͔ʹΑΒͣɺ࣮ݧͰ΋ܭࢉͰ΋ɺےͷྑ͍

    λʔήοτɺ࣮ݧ৚݅ɺύϥϝλΛܾΊΔεςοϓ͸ϘτϧωοΫ
    ࣮ݧσʔλɾܭࢉσʔλɾϑΝΫτͷ஝ੵ
    In-Houseσʔλ + Publicσʔλ + ஌ࣝϕʔε 

    + ͦͷQuality Control / Annotations)
    ػցֶशɾσʔλϚΠχϯά
    +

    View full-size slide

  91. Ϟσϧϕʔε࠷దԽ (୅ཧϞσϧ࠷దԽ)
    81
    ߤۭӉ஦ػͷΑ͏ͳྲྀମػցઃܭͳͲɺܭࢉ͕͔͔࣌ؒΔ

    γϛϡϨʔγϣϯΛ༻͍ͨઃܭ࠷దԽٕज़ͱͯ͠ൃల
    ܭࢉ͕͔͔࣌ؒΔγϛϡϨʔγϣϯͷ୅ཧ(surrogate)

    ͱͯ͠ɺػցֶशɹɹɹΛ׆༻͢Δ
    x!y
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    1. Initial Sampling
    2. Loop:
    1. Construct a Surrogate Model.
    2. Search Infill Criterion.
    3. Add new samples.
    ػցֶश
    దԠαϯϓϦϯά
    ࣮ݧܭը
    e.g.

    Latin hypercube sampling (LHS)
    e.g.

    Expected improvement (EI)
    ☺Recent advances in surrogate-based optimization (Forrester & Keane, 2009)

    https://doi.org/10.1016/j.paerosci.2008.11.001

    View full-size slide

  92. Infillج४ͱ࠷ద࣮ݧܭը
    82
    x
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    ML༧ଌ
    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
    Ͳ͏΍ͬͯόϥϯε੍ޚ͢Δͷ͔ʹ͸৭ʑͳ΍Γ͔͕ͨ͋ΓHot Topics
    • ΞΫςΟϒϥʔχϯά
    • ଟ࿹όϯσΟοτ
    • ਐԽܭࢉ
    • ήʔϜཧ࿦ (CFRͳͲ)
    "exploitation" ͱ "exploration" ͷόϥϯε੍ޚʹ΋MLΛ༻͍Δ
    هड़ࢠ(Ұൠʹ͸ߴ࣍ݩ)
    ׆ੑ
    • ڧԽֶश + ୳ࡧ
    • ϒϥοΫϘοΫε࠷దԽ
    • ϕΠζ࠷దԽ
    • ஞ࣮࣍ݧܭը

    View full-size slide

  93. Infillج४ͱ࠷ద࣮ݧܭը
    82
    x
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    ML༧ଌ
    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
    ༧ଌͷʮෆ࣮֬͞ʯ
    ྫ) ༧ଌ෼ࢄ, ༧ଌ෼෍
    Ͳ͏΍ͬͯόϥϯε੍ޚ͢Δͷ͔ʹ͸৭ʑͳ΍Γ͔͕ͨ͋ΓHot Topics
    • ΞΫςΟϒϥʔχϯά
    • ଟ࿹όϯσΟοτ
    • ਐԽܭࢉ
    • ήʔϜཧ࿦ (CFRͳͲ)
    "exploitation" ͱ "exploration" ͷόϥϯε੍ޚʹ΋MLΛ༻͍Δ
    هड़ࢠ(Ұൠʹ͸ߴ࣍ݩ)
    ׆ੑ
    • ڧԽֶश + ୳ࡧ
    • ϒϥοΫϘοΫε࠷దԽ
    • ϕΠζ࠷దԽ
    • ஞ࣮࣍ݧܭը

    View full-size slide

  94. Infillج४ͱ࠷ద࣮ݧܭը
    82
    x
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    ML༧ଌ
    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
    ༧ଌͷʮෆ࣮֬͞ʯ
    ྫ) ༧ଌ෼ࢄ, ༧ଌ෼෍
    e.g.

    "expected improvement
    Ͳ͏΍ͬͯόϥϯε੍ޚ͢Δͷ͔ʹ͸৭ʑͳ΍Γ͔͕ͨ͋ΓHot Topics
    • ΞΫςΟϒϥʔχϯά
    • ଟ࿹όϯσΟοτ
    • ਐԽܭࢉ
    • ήʔϜཧ࿦ (CFRͳͲ)
    "exploitation" ͱ "exploration" ͷόϥϯε੍ޚʹ΋MLΛ༻͍Δ
    هड़ࢠ(Ұൠʹ͸ߴ࣍ݩ)
    ׆ੑ
    • ڧԽֶश + ୳ࡧ
    • ϒϥοΫϘοΫε࠷దԽ
    • ϕΠζ࠷దԽ
    • ஞ࣮࣍ݧܭը

    View full-size slide

  95. ػցֶश෼໺ࣗମͰ΋Hot Research Topic
    83
    AlphaGo

    (Nature, Jan 2016)
    "VUP.- શࣗಈػցֶश

    AlphaGo Zero

    (Nature, Oct 2017)
    AlphaZero

    (Science, Dec 2018)
    w "MHPSJUIN$POpHVSBUJPO
    w )ZQFSQBSBNFUFS0QUJNJ[BUJPO )10

    w /FVSBM"SDIJUFDUVSF4FBSDI /"4

    w .FUB-FBSOJOH-FBSOJOHUP-FBSO Amazon
    SageMaker
    MuZero

    (arXiv, Nov 2019)

    View full-size slide

  96. ྫ) Model-based RL (Toward sample-efficient RL)
    ࠷ద੍ޚ (Optimal Control)
    ର৅ͷಈతγεςϜͷڍಈ(෺ཧ๏ଇͳͲ)͕Θ͔͍ͬͯΔ৔߹ɺ࠷ྑߦಈΛܾఆՄೳ
    Կ΋Θ͔Βͳ͍৔߹ (Model-free RL)
    ࣮ࡍʹ؀ڥ͔ΒಘΒΕΔߦಈɾঢ়ଶܥྻ͔Β௚઀తʹํࡦ΍Ձ஋ؔ਺Λਪఆ͢Δ
    গ͠౰ͨΓ͕͚ͭΒΕΔ(?)৔߹ (Model-based RL or ݹయతͳγεςϜಉఆͷઃఆ)
    ࣮ࡍͷߦಈɾঢ়ଶܥྻ͔Β·ͣಈతγεςϜͷڍಈΛਪఆ͠ɺͦͷਪఆͨ͠Ϟσϧ

    Λ༻͍ͯ࠷దߦಈΛܭը͢Δ (e.g. কع͢Δͱ͖૬खͷखΛ಄ʹதͰγϛϡϨʔτ͢Δ)
    Planning
    ํࡦ
    ֶश
    ํࡦ+Ձ஋ؔ਺

    ֶश
    Ձ஋ؔ਺

    ֶश

    View full-size slide

  97. ྫ) Model-based RL or Planning with Models
    Deep Planning Network (PlaNet)
    Hafner+ Learning Latent Dynamics for Planning from Pixels.
    arXiv:1811.04551 (Jun 2019)
    MuZero
    Schrittwieser,+ Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model.
    arXiv:1911.08265 (Nov, 2019)
    Simulated Policy Learning (SimPLe)
    Kaiser+ Model-Based Reinforcement Learning for Atari.
    arXiv:1903.00374 (Jun 2019)
    Stochastic Latent Actor-Critic (SLAC)
    Lee+ Stochastic Latent Actor-Critic: Deep Reinforcement Learning with a Latent
    Variable Model.
    arXiv:1907.00953 (Jul 2019)

    View full-size slide

  98. 86
    ػցֶशͰ͖ͨͱ͖ͷ̎छྨͷظ଴
    1. ಘΒΕͨม׵աఔ(ؔ਺)ʹΑΔ༧ଌΛ৭ʑͳ໨తʹ࢖͏
    2. ಘΒΕͨม׵աఔ(ؔ਺)Λ෼ੳͯ͠എܠաఔͷ࢓૊ΈΛ஌Δ
    1.ʹ͍ͭͯ
    • ໰୊͕಺ૠతʹͳΔΑ͏޻෉ (දݱֶशɾӅΕߏ଄ಉఆ)
    • ಺ૠɾ֎ૠ൑ఆ (༧ଌͷ৴པ౓ܭࢉ)
    • Ϟσϧϕʔε࠷దԽͱ୳ࡧ (࠷ద࣮ݧܭը)
    2.ʹ͍ͭͯ
    • ϙετϗοΫղੳͱղऍੑϞσϧ (ֶशࡁΈϞσϧ෼ੳ)

    View full-size slide

  99. (ҼՌͷཧղ͸ఘΊͯ?)ղऍ/આ໌/Ծઆੜ੒΁
    87
    Explainable AI (XAI), Interpretable ML, Causal ML
    ʮղऍʯvsʮཧղʯ
    ਓ(ख๏)ͷ
    ਺͚ͩ͋Δ
    ਅ࣮͸

    ͻͱͭʁ
    എޙͷਅͷ๏ଇʹؔ͢Δ৘ใ͕ಘΒΕΔͱ͸ݶΒͳ͍ͷͰ஫ҙ
    ☺ࢲͷϒοΫϚʔΫɿػցֶशʹ͓͚Δղऍੑ (ݪ ૱, ਓ޻஌ೳ 33(3), 366-369, 2018೥5݄)
    • ถDARPAͷExplainable AI (XAI)ϓϩάϥϜ
    • ػցֶशۀքʹ͓͚ΔInterpretable ML
    • CausalML: ػցֶश for Causal Inference, Counterfactual Prediction, and
    Autonomous Action
    ֤ख๏ʹΑͬͯҟͳΔԾઆܗ੒΍ࣔࠦͷఏڙ

    View full-size slide

  100. ػցֶशϞσϧͷղऍɺػցֶशʹΑΔղऍ
    88
    • ਂ૚ֶश

    ܭࢉάϥϑͱͯ͠දݱ (൚༻తύϥϝλਪఆɿٯϞʔυࣗಈඍ෼)
    • ֬཰తϓϩάϥϛϯά (ੜ੒త౷ܭϞσϦϯά)

    ֬཰ม਺ͷ֊૚తੜ੒ؔ܎Ͱදݱ (൚༻తύϥϝλਪఆ: MCMC/ࣗಈVI)
    • ճؼ෼ੳʹ͓͚ΔཁҼ෼ੳ 

    (ճؼ܎਺ͷ༗ҙੑݕఆ)
    • ϕΠζ༧ଌ෼෍
    • ม਺ॏཁ౓ɾ෦෼ैଐ౓plot
    • άϩʔόϧ/ϩʔΧϧײ౓ղੳ (Sobol'๏, FAST๏, etc)
    • ਂ૚ֶशʹ͓͚ΔSaliency΍Attentionͷར༻
    • ہॴઆ໌ੜ੒: LIME (KDD16), SHAP (NIPS17)
    • ͷ֊૚త෼ղʹΑΔӅΕҼࢠ΍֊૚ߏ଄ͷಉఆ
    • ͷؔ਺ͷPost-hocղੳ ౷ܭֶख๏͸ཁલఏͷݕূ
    • આ໌ม਺ͷબ୒
    • ઢܗͷԾఆ
    • ଟॏڞઢܗੑ "Ϛϧνί"
    • ࢒ࠩͷݕ౼

    :
    x!y
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    x!y
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5

    View full-size slide

  101. ࠓ೔ͷ಺༰
    89
    1. Πϯτϩ

    ػցֶशͱՊֶ(͋Δ͍͸"΋ͷͮ͘Γ")
    2. ػցֶशͰԿ͔ΛʮཧղʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    3. ػցֶशͰԿ͔ΛʮൃݟʯͰ͖Δ͔ʁ

    Answer: ௚઀తʹ͸ݪཧ্ࠔ೉
    4. ͡Ό͋Ͳ͏͢ΜͷʂʁԿ͕͍Δͷʂʁ

    Answer:ʮදݱʯͱʮհೖʯ

    2ͱ3Λલఏʹػցֶश෼໺ͷτϐοΫΛ؆୯ʹ঺հ

    View full-size slide

  102. ࠶ߟ ౷ܭతཧղͱՊֶͷจ๏
    90
    Պֶͷจ๏ (1892) 

    "Statistics is the grammar of science." (Karl Pearson)
    Պֶ͸ʮσʔλͷݟํʯͱແԑͰ͸͍ΒΕͳ͍ʂ
    ݱࡏͷՊֶతٙ໰ͷଟ͘͸100%YES/NOͳ౴͕͑ແ͍໰͍ʂ
    • ͜ͷༀΛҿΊ͹ࢲͷපؾ͸࣏Δͷʁ
    • ͜ͷ݈߁৯඼৯΂͍ͯΕ͹௕ੜ͖Ͱ͖Δͷʁ
    • ͜ͷԽহ඼͚͍ͭͯΕ͹গ͠Ͱ΋ए͍͘ΒΕΔͷʁ
    • ͜ͷ৯඼ͨ΂Ε͹μΠΤοτͰ͖Δͷʁ
    • ݪࢠྗ͸҆શͳͷʁ
    YES and NOͷ
    ؒʹ͖ͬͱਅ࣮͕
    Ն໨ᕸੴΑΓ10࠽೥্ͷେ౷ܭֶऀ
    Պֶͱ͍͏΋ͷʹ͸ɺຊདྷݶք͕͋ͬͯɺ޿͍ҙຯͰͷ࠶ݱՄೳͷݱ৅Λɺ
    ࣗવք͔Βൈ͖ग़ͯ͠ɺͦΕΛ౷ܭֶతʹڀ໌͍ͯ͘͠ɺͦ͏͍͏ੑ࣭ͷ

    ֶ໰ͳͷͰ͋ΔɻʮՊֶͷํ๏ (த୩Ӊ٢࿠)ʯ

    View full-size slide

  103. Impossible to model everything...?
    91
    • ਅͷ๏ଇ͕ਓؒʹཧղՄೳͳ΄ͲγϯϓϧͳϞσϧʹཁૉؐݩͰ͖Δ

    อূ͸Ͳ͜ʹ΋ͳ͍ɻ
    • σʔλ͕༗ݶͳΒͦΕΛઆ໌Ͱ͖ΔϞσϧ͸Ұൠʹແ਺ʹ͋Δɻ
    Ϟσϧͱ͸Կ͔Λࣺ৅ͨ͠΋ͷͰ͋Γ࣮ੈք(ෳࡶܥ)ͱ͸ҧ͏ɻ
    ࿦ड़తཧղ/ཁૉؐݩ͔Βෳࡶܥͷ"౷ܭతཧղ"΁?
    "one of the great statistical minds of the 20th century"
    େ౷ܭֶऀ George E. P. Box (1919-2013)
    https://en.wikipedia.org/wiki/All_models_are_wrong
    "Essentially, all models are wrong,
    but some are useful"

    View full-size slide

  104. ల๬ɿTheory-driven vs Data-drivenͷղফͱ༥߹
    92
    Theory-driven
    Data-driven
    • ର৅ݱ৅ͷෳࡶԽ
    • γϛϡϨʔγϣϯٕ๏΋ෳࡶԽ
    • "ܦݧతʹܾΊΔ"ύϥϝλ΍ॳظ஋
    • ൚ؔ਺ɺަ׵૬߲ؔͷઃܭ
    • খαϯϓϧɾ௿Χ΢ϯτͷ໰୊
    • ֎ૠͷෆՄೳੑͷ໰୊
    • ؼೲόΠΞεͷϞσϧΤϯίʔυ
    • Blackboxੑɾղऍੑͷ໰୊
    • ஌ࣝϕʔεͱ࿦ཧਪ࿦(ه߸AI)ͷݶք
    • ݫີਪ࿦΍୳ࡧͷܭࢉരൃ(NPࠔ೉ੑ)
    • େྔσʔλͷ஌ࣝԽͷ໰୊
    • ੍໿ϓϩάϥϛϯά΍૊߹ͤ࠷దԽ
    (ਓ޻஌ೳ෼໺)
    (ਓ޻஌ೳ෼໺)
    • Data-Drivenख๏(ػցֶश)ͱਓؒͷ

    ࿦ཧతࢥߟͱͷେ͖ͳΪϟοϓ
    • Data͕ͳ͍ྖҬͷ୳ࡧ΍ʮͻΒΊ͖ʯ
    • Ϟσϧద༻ൣғͱ৴པੑɾ҆શੑ
    ৽ͨͳํ๏࿦΁ʁ
    σʔλಉԽɺ໛฿ֶशɺ࿦ཧ߹੒ɺetc
    Ϟσϧϕʔε࠷దԽɺڧԽֶशɺϝλ
    ֶशɺυϝΠϯదԠɺੜ੒Ϟσϧɺetc
    ʲ߹ཧ࿦ʳ
    ʲܦݧ࿦ʳ

    View full-size slide

  105. ੈքτοϓϨϕϧݚڀڌ఺ϓϩάϥϜʢWPIʣ
    93
    ࠷େԯԁ೥º೥

    View full-size slide

  106. ๺ւಓେֶ Խֶ൓Ԡ૑੒ݚڀڌ఺(ICReDD)
    94

    View full-size slide

  107. ڌ఺ͷػٕؔज़ɿԽֶ൓Ԡܦ࿏ͷࣗಈ୳ࡧ
    95
    θ1
    θ2
    O
    H H
    Energy θ = 104.45°
    1
    θ = 95.84 pm
    2
    θ1
    Schrödinger equation
    Potential Energy Surface
    EQ EQ
    T
    ADDF
    Ohno & Maeda, Chem
    Phys Lett, 2004
    Reaction Path
    AFIR
    Maeda & Morokuma, J
    Chem Phys, 2010
    θ2

    View full-size slide

  108. Խֶ൓Ԡͷઃܭͱ୳ࡧ
    96
    Chemical reactions = recombinations of atoms and chemical bonds
    subjected to the laws of nature
    • Intractably large chemical space: A intractably large number of
    "theoretically possible" candidates for reactions and compounds...
    • Scalability issue: Simulating an Avogadro-constant number of atoms is
    utterly infeasible... (After all, we need some compromise here)
    • Complexity and uncertainty of real-world systems:
    Many uncertain factors and arbitrary parameters are involved...
    • Known and unknown imperfections of currently established theories:
    Current theoretical calculations have many exceptions and limitations...

    View full-size slide

  109. 97
    Cause-and-Effect

    Relationship
    Related factors

    (and their states)
    Outcome
    Reactions
    Some
    mechanism
    [Inputs] [Outputs]
    Theory-driven methods try to explicitly model the inner workings
    of a target phenomenon (e.g. through first-principles simulations)
    Data-driven methods try to precisely approximate its outer behavior
    (the input-output relationship) observable as "data". 

    (e.g. through machine learning from a large collection of data)
    governing equation?
    Խֶ൓Ԡͷઃܭͱ୳ࡧ

    View full-size slide

  110. 98
    Խֶ൓Ԡͷઃܭͱ୳ࡧ
    Brc1cncc(Br)c1 C[O-] CN(C)C=O Na+ COc1cncc(Br)c1
    SMILES
    Structural Formla
    Steric Structures
    Electronic States
    Reactants Reagents Products
    As pattern languages (e.g. known facts in textbooks/databases)
    As physical entities (e.g. quantum chemical calculations)

    View full-size slide

  111. 99
    Խֶ൓Ԡͷઃܭͱ୳ࡧ
    Computer-assisted synthetic planning
    (path search on knowledge bases)
    or AI-Assisted Synthesis?
    (with Machine Learning)

    View full-size slide

  112. https://www.chemistryworld.com/machine-learning/1616.tag

    View full-size slide

  113. ML-based chemical reaction predictions
    3N-MCTS/AlphaChem
    Segler+ Nature 2018
    Molecular Transformer
    Schwaller+ ACS Cent Sci
    2019
    seq2seq
    Liu+ ACS Cent Sci 2017
    WLDN
    Jin+ NeurIPS 2017
    ELECTRO
    Bradshaw+ICLR 2019
    WLN
    Coley+ Chem Sci 2019
    GPTN
    Do+ KDD 2019
    Graph NN Sequence NN Combined or Other
    IBM RXN
    Schwaller+ Chem Sci 2018
    Molecule Chef
    Bradshaw+ DeepGenStruct
    (ICLR WS) 2019
    Neural-Symbolic ML
    Segler+ Chemistry 2017
    Similarity-based
    Coley+ ACS Cent Sci 2017
    Fermionic Neural Network
    Pfau+ Ab-Initio Solution of the Many-Electron Schrödinger Equation with Deep Neural Networks. 

    arXiv:1909.02487, Sep 2019.
    ML + First-principle simulations
    Hamiltonian Graph Networks with ODE Integrators
    Sanchez-Gonzalez+ Hamiltonian Graph Networks with ODE Integrators. 

    arXiv:1909.12790, Sep 2019.
    Both from
    Խֶ൓Ԡͷઃܭͱ୳ࡧ

    View full-size slide

  114. Խֶ൓Ԡͷઃܭͱ୳ࡧ
    102

    View full-size slide

  115. ࠢ͸ࡉ෦ʹ॓Δɿಓ۩ͱͯ͠ͷػցֶश
    103
    • ϓϩʹͱͬͯ࢖͏ಓ۩͸໋ɻ
    • ಓ۩ͷಛੑʹਫ਼௨͠ɺஸೡʹѻ͍ɺ
    खೖΕΛଵΒͳ͍ɻ
    • ಓ۩ശͷதΛͻͱΊݟΔ͚ͩͰͦͷ
    ৬ਓͷؾ࣭ͱϨϕϧ͕෼͔Δɻ
    ৬ਓࠢ (ٕज़ऀࠢ)
    ΑΓཧղ͢ΔͨΊʹٕ๏(ಓ۩)Λ੔උ͢Δ
    ʮྑ͍࢓ࣄ͸ྑ͘खೖΕ͞Εͨಓ۩͔Βʯ
    http://www.900910.com/mies.php
    ʮػցֶशʯݚڀͷҙٛ

    View full-size slide

  116. ࠷ޙʹɿհೖɾ࣮ݧ΋ؚΉ࠷దԽ΁
    104
    ݅ͷBoxͷ1966೥ͷ࿦จʮUse and Abuse of regressionʯ͸

    ඇৗʹ༗໊ͳ͜ΜͳҰจͰకΊ͘͘ΒΕΔɻ
    ཧ۶͔Βݴͬͯ΋ػցֶश԰ͱσʔλʮ͚ͩʯͰ͸Կ΋Ͱ͖ͳ͍
    ͱ͍͏͜ͱͰ͢ɻ
    To find out what happens to a system when you
    interfere with it you have to interfere with it
    (not just passively observe it).
    υϝΠϯ஌ࣝΛ࣋ͬͨઐ໳Ոͱͷڠಇ͕ඞਢͰ͢ʂ
    Ͳ͏ͧΑΖ͓͘͠ئ͍͠·͢(?)

    View full-size slide

  117. Take Home Message
    105
    Պֶ͕ٻΊΔ͜ͱ: ෼͔Βͳ͍͜ͱ͕෼͔Δ(Պֶతൃݟ)
    ൃݟ
    ཧղ ݪҼͱ݁Ռ(ҼՌؔ܎)Λݟग़͢
    ࠓ·Ͱݟग़͞Ε͍ͯͳ͍ྑ͍ର৅Λݟग़͢
    ࠓ೔఻͍͑ͨͨͬͨ3ͭͷ͜ͱ
    1. ୯७ʹػցֶशΛ࢖͏͚ͩͰ͸͍ͣΕ΋ղ͚ͳ͍
    2. ػցֶशҎ֎ͷ΋ͷ(հೖ΍υϝΠϯ஌ࣝ)͕ݪཧ্ඞਢ
    3. ࠷ۙ·͞ʹݚڀ͕ਐߦதͷະղܾྖҬ͕ͩݚڀ͸৭ʑ͋Δ
    ͷաఔΛཧղ͠(ਓ͕ؒ)ൃݟ͢Δ
    x!y
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    Λར༻ͯ͠ྑ͍ Λ࣋ͭ Λൃݟ͢Δ
    x
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    AAAB/XicbVDLSgMxFL2pr1pfVZdugkVwVWZE0GXRjcsK9gHtUDJppo1NMkOSEctQ/AW3uncnbv0Wt36JaTsLbT1w4XDOvZzLCRPBjfW8L1RYWV1b3yhulra2d3b3yvsHTROnmrIGjUWs2yExTHDFGpZbwdqJZkSGgrXC0fXUbz0wbXis7uw4YYEkA8UjTol1UrMbyuxx0itXvKo3A14mfk4qkKPeK393+zFNJVOWCmJMx/cSG2REW04Fm5S6qWEJoSMyYB1HFZHMBNns2wk+cUofR7F2oyyeqb8vMiKNGcvQbUpih2bRm4r/eqFcSLbRZZBxlaSWKToPjlKBbYynVeA+14xaMXaEUM3d75gOiSbUusJKrhR/sYJl0jyr+l7Vvz2v1K7yeopwBMdwCj5cQA1uoA4NoHAPz/ACr+gJvaF39DFfLaD85hD+AH3+ADzJlfc=
    x!y
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5
    AAACBXicbVC7SgNBFL0bXzG+opY2E4NgFXZF0DJoYxnBPCC7htnJJBkys7PMzIrLktpfsNXeTmz9Dlu/xEmyhSYeuHA4517O5YQxZ9q47pdTWFldW98obpa2tnd298r7By0tE0Vok0guVSfEmnIW0aZhhtNOrCgWIaftcHw99dsPVGkmozuTxjQQeBixASPYWOneD0X2OPErvpF+Je2Vq27NnQEtEy8nVcjR6JW//b4kiaCRIRxr3fXc2AQZVoYRTiclP9E0xmSMh7RraYQF1UE2+3qCTqzSRwOp7EQGzdTfFxkWWqcitJsCm5Fe9Kbiv14oFpLN4DLIWBQnhkZkHjxIODISTStBfaYoMTy1BBPF7O+IjLDCxNjiSrYUb7GCZdI6q3luzbs9r9av8nqKcATHcAoeXEAdbqABTSCg4Ble4NV5ct6cd+djvlpw8ptD+APn8wfECJj5

    View full-size slide