Turing × atmaCup #18 - 1st Place Solution

Slide 1

Slide 1 text

Slide 2

Slide 2 text

© 2024 Wantedly, Inc. ໊લɿ   ߹ా पฏ Shuhei Goda ॴଐͱ໾ׂɿ   ΢ΥϯςουϦʔגࣜձࣾ   ɾData Team Manager   ɾMachine Learning Tech Lead   ɾProduct Manager     Kaggle Tierɿ   Kaggle Competitions Grandmaster @jy_msc ࣗݾ঺հ https://www.kaggle.com/shuheigoda

Slide 3

Slide 3 text

Slide 4

Slide 4 text

Slide 5

Slide 5 text

Slide 6

Slide 6 text

Slide 7

Slide 7 text

Slide 8

Slide 8 text

Slide 9

Slide 9 text

Slide 10

Slide 10 text

Slide 11

Slide 11 text

Slide 12

Slide 12 text

Slide 13

Slide 13 text

© 2024 Wantedly, Inc. 1. ෼ੳͱํ਑ܾΊ ࣮ݧʹ࢖༻ͨ͠ϕʔεϥΠϯ ● tk@tnkcoder ͞Μ͕ެ։ͨ͠ϕʔεϥΠϯϞσϧΛར༻ ● [CV 0.2008/LB 0.2017] LightGBM + CNN stacking baseline (LightGBM + CNN) 1st-stage:   CNN 2nd-stage: GBDT Image Tabular 1st-stage Predictions Submission

Slide 14

Slide 14 text

Slide 15

Slide 15 text

Slide 16

Slide 16 text

© 2024 Wantedly, Inc. 2-1. Target ͷվળ ֤༧ଌ࣌఺ͷ଎౓΍Ճ଎౓Λ Auxiliary Target ͱֶͯ͠शɾ༧ଌ͢Δ ● ྫ͑͹ Target ͷ x_0, x_1 ͔Β vx_1 Λࢉग़͢Δ͜ͱ͕Ͱ͖Δ ● ݁Ռͱͯ͠ɺ1st-stage CV: 0.2312 → 0.2288 (-0.0024) ʹվળ ● ·ͨɺAuxiliary Target ʹର͢Δ༧ଌ஋Λޙஈͷಛ௃ྔͱͯ͠௥Ճ͢Δ͜ͱͰ ΋ɺ2nd-stage ͷείΞ͕վળʢ2nd-stage CV: 0.1963 → 0.1933ʣ

Slide 17

Slide 17 text

Slide 18

Slide 18 text

© 2024 Wantedly, Inc. 2-2. HorizontalFlip ֶश࣌ɾਪ࿦࣌ʹ HorizontalFlip ΛՃ͑Δ ● ֶश࣌ɿp=0.5 Ͱ HorizontalFlip ● ਪ࿦࣌ɿΦϦδφϧը૾ͷਪ࿦݁Ռͱਫฏ൓సͨ͠ਪ࿦݁ՌΛฏۉ͢Δ ● ͜ΕΒʹΑͬͯείΞ͸৳ͼΔ͕ɺ1st-stage Ϟσϧʹೖྗ͢Δςʔϒϧಛ௃ྔ ͷ൓సΛ๨Εͯ͠·͏ͱείΞ͕ٯʹԼ͕ͬͯ͠·͏ͷͰ஫ҙ ● 1st-stageͰѻ͏ςʔϒϧಛ௃ྔ͸ۃྗγϯϓϧʹ཈͑Δඞཁ͕͋Δɻ୅Θ Γʹ 2nd-stage ʹෳࡶͳFEΛدͤΔ͜ͱ͕Ͱ͖Δ {“steeringAngleDeg”: 15, “leftBlinker”: True, “rightBlinker”: False}   → {“steeringAngleDeg”: -15, “leftBlinker”: False, “rightBlinker”: True}

Slide 19

Slide 19 text

© 2024 Wantedly, Inc. 2-3. Scene୯Ґ ࢥͬͨ͜ͱɿಉҰγʔϯͷલޙͷࢹ֮త৘ใΛ࢖͏͜ͱͰɺ୯ҰID(t-1.0 ~ t)͚ͩͩ ͱࠔ೉ͳ༧ଌ΋Ͱ͖ΔΑ͏ʹͳΔͷͰ͸ʁΑΓ௕ظతͳӡసঢ়گͷ೺Ѳ͕ॏཁ ྫ1ɿ sec=2.0, t-0.5 sec=2.0, t-1.0 sec=2.0 sec=12.0 12secޙͷ৘ใ͔Βɺͦͷ··௚ਐ͢Ε͹   ྑ͔ͬͨ͜ͱ͕Θ͔Δ

Slide 20

Slide 20 text

Slide 21

Slide 21 text

Slide 22

Slide 22 text

© 2024 Wantedly, Inc. 2-3. Scene୯Ґ ϘτϧωοΫͱͳ͍ͬͯΔʮ௕ظมԽʯΛޮՌతʹϞσϧԽ͢Δʹ͸ʁ ● ϕʔεϥΠϯͰ͸2nd-stageʹ͓͍ͯલޙͷ௕ظ৘ใΛߟྀͨ͠༧ଌ͕Մೳͩ ͕ɺ1st-stageʹ͓͍ͯ͸֤ID͕ಠཱͳ΋ͷͱͯ͠ѻ͏&ܦ࿏༧ଌͷ݁Ռͱͯ͠ͷ ৘ใΛ఻ൖͤ͞ΔܗʹͳΔͷͰඇޮ཰ʹݟ͑Δ ● 1st-stage ͷNNͷஈ֊Ͱɺ௕ظͷมԽʹجͮ͘܏޲ΛֶशͰ͖ΔΑ͏ʹ͢Δ 1st-stage:   CNN 2nd-stage: GBDT sceneA,ID1 1st-stage Predictions 1st-stage:   CNN 1st-stage Predictions FE(e.g. shift features) sceneA,ID2 Shared

Slide 23

Slide 23 text

Slide 24

Slide 24 text

© 2024 Wantedly, Inc. 2-3. Scene୯Ґ sec=20 sec=120 Pad Pad Pad Pad sec=220 sec=320 sec=520 Pad Pad Pad sec=20 sec=120 sec=220 sec=320 sec=420 sec=520 scene=A scene=B scene=C όονͷ࡞Γํ ● Scene͝ͱʹ௕͕͞ҟͳΔͷͰɺ٧ΊͯPadding ● αϯϓϧؒͰ࣌ܥྻతͳҐஔؔ܎͕ҟͳΔͷͰɺscene_sec ΍ scene_num ʢsceneͷத ͰԿ൪໨ʹొ৔ͨ͠ID͔ʣΛಛ௃ྔͱͯ͠ೖྗ

Slide 25

Slide 25 text

Slide 26

Slide 26 text

Slide 27

Slide 27 text

Slide 28

Slide 28 text

Slide 29

Slide 29 text

Slide 30

Slide 30 text

Slide 31

Slide 31 text

Slide 32

Slide 32 text

Slide 33

Slide 33 text

Slide 34

Slide 34 text

Slide 35

Slide 35 text

Slide 36

Slide 36 text

Slide 37

Slide 37 text

Slide 38

Slide 38 text

Slide 39

Slide 39 text

Slide 40

Slide 40 text

© 2024 Wantedly, Inc. Ͳ͏͍͏՝୊Λղ͘΂͖͔Λઃఆ͢Δ Ͳ͏͍͏໰͍ʢnot Ξϓϩʔνʣ͕ࠩผԽϙΠϯτʹͳΔͷ͔ߟ͑Δ ● ΞΠσΞΛεϙοτతʹݕূ͢ΔΑΓ΋ɺूத͢Δ΂͖՝୊Λઃఆͯ͠ਂ΅ͬ ͨ΄͏͕ɺద੾ͳΞϓϩʔνʹͨͲΓண͖΍͍͢ ෼ੳޙʹઃఆͨ͠՝୊ QɿલޙͷγʔϯͷมԽ͸௕ظͷ܏޲Λ֫ಘ͢Δͷʹ༗ޮ͔ʁ Qɿӡస࣌ͷঢ়گ͸༧ଌʹͲͷΑ͏ʹӨڹ͢Δͷ͔ʁʢྫ͑͹ߴ଎ಓ࿏ͩͱʁʣ

Slide 41

Slide 41 text

© 2024 Wantedly, Inc. ੜ੒AIػೳͰ࣮ݧεϐʔυΛૣ͘͢Δ ࣮ݧαΠΫϧ͕ैདྷͷ1/2~1/3ͷ࣌ؒͰճͤΔΑ͏ʹ ● ࣮ݧͷઃܭ͔ΒݕূʢσόοάʣʹࢸΔ·Ͱͷ࣌ؒͷେ෯ͳ୹ॖ ● ΊΜͲ͘͞… ͱ͍͏৺ཧతϋʔυϧΛେ෯ʹԼ͛Δʢਖ਼௚͜Ε͕େ͖͍ʣ ● ྫ͑͹ɺID୯Ґ→Scene୯Ґ΁ͷมߋɺมߋՕॴ͕ଟͯ͘ਏ͍ ● Ͳ͏ઃܭ͢Δ΂͖͔ɺͲ͏͍͏มߋՕॴ͕͋Δ͔ɺͲ͏࣮૷͢Δ͔Λ౴͑ ͯ΋Β͏ɻͦͯ͠ίέͨΒσόοάͷࡐྉΛ΋Β͏ ΞΠσΞͷ ݕ౼ ઃܭ ࣮૷ ݕূ