TrendOwl - Speaker Deck

Slide 1

Slide 1 text

No content

Slide 2

Slide 2 text

.FNCFST Α͏͔Μ ͿΜͿΜ ͍ʔͪΌΜ 5ZMFS Ωϯδϣ΢ ͱΈͬ͘͢ ࣎լ ژ౎ ౦ژ ෱ౡ aaࢲ͕͸ͳ͍ͯ͠·͢

Slide 3

Slide 3 text

No content

Slide 4

Slide 4 text

IUUQTRJJUBDPNBCDTIPUBSPJUFNTGDCBDF

Slide 5

Slide 5 text

ເΛඳ͖͍ͨ ʮϑΝΫτνΣοΫͷຽओԽʯ Λ໨ࢦ͍ͯ͠·͢ʜ

Slide 6

Slide 6 text

੍࡞෺ 🤔ϑΣΠΫχϡʔεͱ͸

Slide 7

Slide 7 text

໋΍੓࣏ʹؔΘΔ͜ͱ΋ ۽ຊ஍਒ ೥ถબڍ

Slide 8

Slide 8 text

ϫΫνϯ͸ࣗ෼ͷҙࢥͰଧͯΔ ͷʹɺଧͭͳͱڧ੍͞Εͯ͠ ·͍ͬͯΔʜ ࣮ࡍͷମݧஊ ਌͕৘ใݯΛͭʹཔΓ͗ͯ͢ɺ ภͬͨ৘ใͰ࿩͢Α͏ʹͳΓɺ ձ࿩͕੒Γཱͨͳ͍ ϝϯόʔͷ5͘Μ ༑ୡͷ/͞Μ

Slide 9

Slide 9 text

૯຿লͷൃදͰ͸ϲ݄ʹҎ্ͷਓ͸ϑΣΠΫχϡʔεʹ৮Ε͍ͯΔ Ҿ༻ɿIUUQTXXXTPVNVHPKQNBJO@TPTJLJKPIP@UTVTJOE@TZPIJJIPZVHBJ@IUNM   ʢຊݚڀձʢୈճʣࢿྉ̍ʮʰϑΣΠΫχϡʔεʱʹؔ͢ΔΞϯέʔτௐࠪ݁Ռʯʢ/3*ʣʣ

Slide 10

Slide 10 text

ίϩφʹؔ͢Δ৘ใͷਅِΛ൑அͰ͖ͳ͔ͬͨਓ͕ଟ͍܏޲ ɾίϩφՒͷ೔ຊͰ͸ɺ   ʮ͓౬Ͱ৽ܕίϩφ΢Πϧε͕ࢮ໓͢Δʯ   ΍ʮ৽ܕίϩφ΢ΠϧεͷӨڹͰ   τΠϨοτϖʔύʔ͕ෆ଍͢Δʯ   ͳͲͷِ৘ใ͕֦ࢄͨ͠ɻ   ɾ૯຿লͷௐࠪͰɺ͜ΕΒͷِ৘ใͷ͏ͪ   ҰͭͰ΋৴ͨ͡ͱ౴͑ͨਓ͸ ೥ྸ૚͕௿͘ͳΔ΄Ͳଟ͘ɺ   ̍̑ʙ̍̕ࡀͰ͸̏̒ʹ΋ٴΜͰ͍Δɻ Ҿ༻ɿIUUQTXXXTPVNVHPKQNBJO@DPOUFOUQEG ʢ৽ܕίϩφ΢Πϧεײછ঱ʹؔ͢Δ৘ใྲྀ௨ௐࠪʣ

Slide 11

Slide 11 text

Ҿ༻ɿIUUQTXXXOBUVSFDPNBSUJDMFTT ʢ5IFPOMJOFDPNQFUJUJPOCFUXFFOQSPBOEBOUJWBDDJOBUJPOWJFXTʣ ϫΫνϯ൓ର೿͕͜ͷ··૿͑ଓ͚Δͱ͍ۙকདྷऔΓࠐ·Εͯ͠·͏ةݥੑ͕͋Δͱ͍͏ݚڀ݁Ռ😱

Slide 12

Slide 12 text

͍ۙকདྷ΁ͷةػײ ৘ใੜଶܥͰੜ੒͞ΕΔ৘ใ͕૿େ͠ɺ   ͔ͭෆ͔֬ͳ৘ใͷׂ߹͕ߴ͍ʢͭ·ΓϊΠζ͕ ଟ͍ʣ৘ใաଟͷঢ়ଶ͕ੜ͡Δ͜ͱͰɺِ৘ใΛ ࣄ࣮ͱޡೝͨ͠Γɺ͋Δ͍͸ࣄ࣮Λࣄ࣮ͩͱ৴͡ ΒΕͳ͍Α͏ͳ͜ͱ͕සൟʹى͜ΔΑ͏ʹͳΕ ͹ɺ೔ৗੜ׆΍ܦࡁ׆ಈɺ͞Βʹ͸ຽओओٛʹ΋ ෛͷӨڹ͕ٴͿՄೳੑ͕͋Δɻ

Slide 13

Slide 13 text

ʁ Έͳ͞ΜɺϑΣΠΫχϡʔε ରࡦ͸΍ͬͯ·͔͢ʁ ʁ

Slide 14

Slide 14 text

͍͍͑ ͸͍ ΠϯελάϥϜʹͯνʔϜϝϯόʔௐ΂ ʮϑΣΠΫχϡʔεʹԿ͔͠ΒରࡦΛ͍ͯ͠Δ͔ʯ ߴߍੜɾେֶੜਓ

Slide 15

Slide 15 text

͔͠͠ϑΣΠΫχϡʔε ʹର͢ΔҙࣝΛ͍࣋ͬͯ Δਓ͸·ͩ·ͩগͳ͍ ௐࠪͷ݁Ռ ϑΣΠΫχϡʔεʹ৮ΕΔස౓͸೔ʹ೔ʹ૿͍͑ͯΔ

Slide 16

Slide 16 text

ײ৘΍ओ؍ʹࠨӈ͞Εͣ ٬؍తͳࠜڌ Λ࣋ͬͯ΋Β͍͍ͨ Ծઆɾ໨ඪ ؆୯ʹϑΣΠΫχϡʔε൑ఆʹ৮ΕΒΕΔ   ؀ڥ͕ඞཁͰ͋ΔͷͰ͸ͳ͍͔ɻ

Slide 17

Slide 17 text

੍࡞෺ ✌੍࡞෺

Slide 18

Slide 18 text

IUUQTUXJUUFSDPNUSFOE@PXM

Slide 19

Slide 19 text

σβΠϯͷϙΠϯτ ը૾ʹ਺஋Λॻ͔ͳ͍ ஫ҙࣄ߲ΛಡΜͰ΋Β͏ͨΊ ʹɺ͋͑ͯςΩετʹ͔݁͠ ՌΛॻ͖·ͤΜͰͨ͠ πΠʔτΛݟΔͱ͖ʹ ਓؒ͸ը૾͔ΒݟΔ 👎

Slide 20

Slide 20 text

੍࡞෺ 👩🎓෼ੳख๏

Slide 21

Slide 21 text

఻೻Ϟσϧºػցֶश Ҿ༻ɿIUUQTXXXSJLFMBCKQTUVEZ ਖ਼͍͠৘ใ '",&

Slide 22

Slide 22 text

IUUQTXXXBBBJPSHPDTJOEFYQIQ"""*"""*QBQFSWJFX'JMF ൑அख๏ ɾΞΧ΢ϯτ͕ొ࿥͞Ε͔ͯΒͲΕ͙Β͍͔ ɾݩπΠʔτ͔πΠʔτ͞Ε͔ͯΒͲΕ͙Β͍ܦա͍ͯ͠Δ͔   ɾೝূࡁΈΞΧ΢ϯτ͔Ͳ͏͔   ɾϓϩϑΟʔϧը૾Λࢦఆ͍ͯ͠Δ͔Ͳ͏͔   ɾ͍͍Ͷ͍ͯ͠Δ਺   ɾϑΥϩϫʔͷ਺   ɾϑΥϩʔ͍ͯ͠Δ਺   ɾ໊લͷจࣈ਺   ɾࣗݾ঺հͷจࣈ਺ πΠʔτจষʹҰ੾ͱΒΘΕͣ෼ੳΛߦ͏ 📌 ࢀߟ࿦จ

Slide 23

Slide 23 text

੍࡞෺ ✍Ϟσϧͷਫ਼౓ɾಁ໌ੑ

Slide 24

Slide 24 text

ֶशϞσϧ͸ਫ਼౓௒͑ Ϟσϧͷ༧ଌ ࣮ ࡍ

Slide 25

Slide 25 text

ಁ໌ੑϨϙʔτΛ࡞੒ެ։ tensorboardXΛ༻͍ͨϞσϧͷՄࢹԽ

Slide 26

Slide 26 text

෼ੳ 🧠 πΠʔτऩू IUUQTUXJUUFSDPNUSFOE@PXM ϑΣΠΫ෼ੳ 🤖 ใࠂ

Slide 27

Slide 27 text

IUUQTUSFOEPXMTVHPLVOBSJUBJEFW ϙʔλϧαΠτ աڈͷ෼ੳ͕ݟΕΔ ෼ੳϦΫΤετ͕ૹΕΔ

Slide 28

Slide 28 text

"NB[PO$MPVE8BUDI&WFOU πΠʔτ "NB[PO4 "NB[PO-BNCEB "NB[PO%ZOBNP%# ఆظ࣮ߦ ෼ੳ ؅ཧը໘ ϩά ը૾Ξοϓϩʔυ ٕज़ߏ੒ (PPHMF"QQ4DSJQU ϑϩϯτΤϯυ

Slide 29

Slide 29 text

੍࡞෺ 😁5SFOE0XM΁ͷظ଴

Slide 30

Slide 30 text

ϑΣΠΫ χ ϡ ʔε ͸ ૣ ͘ ޿ ͕ Δ

Slide 31

Slide 31 text

ιʔείʔυ΍Ϟσϧɺ஌ݟͷެ։   ڵຯΛ࣋ͬͯͩ͘͞ΔํΛ૿΍͠   ք۾શମͷٕज़ྗΛ্͛Δ ࠓޙͷల๬ ҙݟΛ΋Β͏ɾ޿ΊΔ ݚڀػؔ΍اۀͱͷڞಉݚڀ   ༷ʑੈ୅΍૚ʹ֦ࢄ ެ։ɾڞ༗͢Δ ਫ਼౓ΛߴΊΔ ࣗવݴޠॲཧͱ఻೻Ϟσϧͷྑ఺Λ ༻͍ͨϞσϧͷ։ൃ ΞϓϩʔνΛ૿΍͢ 5XJUUFS΍8FCҎ֎ͰͷϦϦʔε   ༷ʑͳΞόλʔઃఆͰݕূΛߦ͏

Slide 32

Slide 32 text

No content

Slide 33

Slide 33 text

༧උࢿྉ

Slide 34

Slide 34 text

ϑϩϯτΤϯυ Ϧʔμʔ ػցֶशɾαʔόʔαΠυ Α͏͔Μ ͿΜͿΜ ͍ʔͪΌΜ 5ZMFS Ωϯδϣ΢ ͱΈͬ͘͢ αʔϕΠɾσβΠϯ αʔϕΠɾϢʔβΠϯλϏϡʔ αʔόʔαΠυ

Slide 35

Slide 35 text

4DSBQCPY νʔϜ࿈ܞ /PUJPO 4MBDL

Slide 36

Slide 36 text

ͳͥɺʮϑΝΫτνΣοΫʯͱ͍͏   ΞϓϩʔνΛऔ͍ͬͯΔͷ͔ʁ

Slide 37

Slide 37 text

৘ใੜଶܥ͸ਐԽ͠ଓ͚͍ͯΔ Ҿ༻ɿIUUQTXXXEFMPJUUFDPNKQKBQBHFTTUSBUFHZBSUJDMFTDCTJOGPSNBUJPOFQJEFNJDIUNM   ʢੈلͰສഒʹ૿େͨ͠৘ใ఻ୡྗʙ৘ใͷٸ଎ͳ఻છʮΠϯϑΥσϛοΫʯͱ͸ʣ

Slide 38

Slide 38 text

No content

Slide 39

Slide 39 text

օ͞ΜͷྗΛି͍ͯͩ͘͠͞ 5XJ UU FS ͷதͷਓͱ͓஌Γ߹͍ͷํ ޙ΄Ͳ͓ܨ͍͖͍͗ͨͩͨͰ͢ʂ

Slide 40

Slide 40 text

ࢲୡ͕΍͍ͬͯΔ͜ͱ͸Կ͕೔ຊൃɺੈքॳͷϓϩμΫτͳͷ͔ ैདྷ΍ڝ߹ͱԿ͕ҧ͏ͷ طଘͷ࢓૊Έͱͷҧ͍ /-1 ࣗવݴޠॲཧ ͰͰ͖Δ͜ͱ طଘͷࣙॻɾίʔύεΛ΋ͱʹ จ຺Λղੳ͢Δ ٯʹݴ͑͹ ৽͍͠෺ࣄɾະ஌ͷ໰୊ʹ ରॲͰ͖ͳ͍͔΋͠Εͳ͍

Slide 41

Slide 41 text

ࢲୡ͕΍͍ͬͯΔ͜ͱ͸Կ͕೔ຊൃɺੈքॳͷϓϩμΫτͳͷ͔ ैདྷ΍ڝ߹ͱԿ͕ҧ͏ͷ طଘͷ࢓૊Έͱͷҧ͍ ఻೻Ϟσϧ͸ղܾ͠·͢ɻ ະ஌ͷࣄ࣮ʹରͯ͠΋ ٬؍తͳ෼ੳ͕Մೳ

Slide 42

Slide 42 text

ࢲୡ͕΍͍ͬͯΔ͜ͱ͸Կ͕೔ຊൃɺੈքॳͷϓϩμΫτͳͷ͔ ैདྷ΍ڝ߹ͱԿ͕ҧ͏ͷ ػցֶशϞσϧͷ࢓૊Έ 3// -45. ࠶ؼܕχϡʔϥϧ ωοτϫʔΫ ϩʔΧϧͳ෼ੳ $// ৞ΈࠐΈχϡʔϥϧ ωοτϫʔΫ άϩʔόϧͳ෼ੳ /-1Ͱকདྷతʹਫ਼౓޲্Λʂ [ࢀߟจݙ] Yang Liu, Yi-Fang Brook Wu (2018) Early Detection of Fake News on Social Media Through Propagation Path Classi fi cation with Recurrent and Convolutional Networks