情報の世界 2023年度 第12回「個人のデータ」 #情報の世界 / Data of Individual 2023

2023年7月6日(木)に開催された北海道情報大学の講義「情報の世界」の第12回「個人のデータ 〜 ウェアラブルデバイス、行動認識、画像処理、プライバシー、パーソナルデータ etc.」の講義資料

Transcript

1. 2023.07.06 ๺ւಓ৘ใେֶ ৘ใϝσΟΞֶ෦ ৘ใϝσΟΞֶՊ ౬ଜ ཌྷ !"#$%& ()*+,-$./012345167891:+/.;%1<%=>-$%&'?@AB CDEFEGH'I'CDEFEJKL DEFEMKCNOPQORONKSBKABQT ৘ใͷੈք ୈ12ճ
2. !"#$ • Twitter΁ͷײ૝ɾ࣭໰ͷ౤ߘ׻ܴ (͜ͷतۀͷࢲͷճͷΈ) • ϋογϡλά #৘ใͷੈք • HIU-WiFi͔ΒTwitterͭͳ͕Βͳ͍ͷͰ(͝ΊΜͶ🥺)ɺɹ εϚϗ͔Β౤ߘ͍ͯͩ͘͠͞ • ޶όϨͨ͘͠ͳ͍ਓ͸तۀ༻ΞΧ΢ϯτ࡞ΔͱΑ͍͔΋ • εϥΠυͷࡱӨɾεΫγϣ΋౤ߘOKͰ͢🙆 • খςετͷεΫγϣ౤ߘ͸NG🙅 2
3. %&'() • ࠓ೔ͷߨٛςʔϚ͸ʮݸਓͷσʔλʯ • ʮݸਓʯͱ͍͍ͭͭɺ͍ΖΜͳ࿩͕͋Γ·͢ • ͨͿΜ৘ใաଟͰ͢ • ͋ͱͰPOLITE3ͷϦϯΫूΛৼΓฦ͍ͬͯͩ͘͞ • ͥͻʮࣗ෼͕ͲͷΑ͏ʹؔΘΕΔ͔ʯΛҙࣝ͠ͳ͕Β ฉ͍͍ͯͩ͘͞ 3 👀 σϞ͋Γ ಈը͋Γ
4. %&'() 4 લ൒ (40෼) ΢ΣΞϥϒϧσόΠεɺߦಈೝࣝɺը૾ॲཧ ޙ൒ (25෼) ϓϥΠόγʔɺύʔιφϧσʔλ খςετ (20෼) BreakɿTeachable Machineɹ(5෼)
5. *+,-./0123 5 ϝΨω (JINS MEME) JINS MEME https://jinsmeme.com/ Apple Watch Series 3 https://www.apple.com/jp/shop/buy-watch/apple-watch-series-3 Apple Watch
6. 456789:6; 6 Xiaomi MiεϚʔτόϯυ6-ੈքҰͷϙʔλϒϧ | Xiaomi Japa n https://www.mi.com/jp/product/mi-smart-band-6/ ¥4,870
7. <=>=?'@ABC>D=EFGHF6I6JKLKM5 7 ϞόΠϧϞʔγϣϯΩϟϓνϟʔ:Մೳੑɺແݶେɻ͍ͪͯ͘͞ɺ͔Δ͍ɺϞόΠϧϞʔγϣϯΩϟϓνϟʔ mocopiʢϞίϐʣʲιχʔެࣜʳ https://youtu.be/ovHHpralCD8 👀
8. <=>=?'@ABC>D=EFGHF6I6JKLKM5 8 👀
9. NOPQ • େྔͷσʔλΛػց(ίϯϐϡʔλ)͕ಡΈࠐΜͰ ֶश͠ɺͦͷֶशͨ݁͠Ռ(Ϟσϧ)Λ࢖ͬͯॲཧ Λߦ͏ • Ͱ͖Δ͜ͱ • ෼ྨɺ༧ଌɺҟৗݕ஌ɺੜ੒ 9
10. RSTUVWXY • ༷ʑͳσʔλΛ༻͍ͯਓؒͷߦಈΛೝࣝ͢Δ • ϔϧεέΞɺεϙʔπͳͲʹ׆༻ 10 ڑ཭ηϯα ಈըɾը૾ ߦಈೝࣝ RUNNING WALKING SITTING STANDING JUMPING σʔλ εϚʔτϑΥϯ
    ʢՃ଎౓ɾ܏͖ɾ஍࣓ؾ౳ʣ ΢ΣΞϥϒϧσόΠε
    ʢՃ଎౓ɾ຺೾౳ʣ
11. Z[\<=]$^_VWXY • Bayat et al., A Study on Human Activity Recognition Using Accelerometer Data from Smartphones (2014) • εϚʔτϑΥϯͷՃ଎౓ηϯ αͰߦಈೝࣝ • 2005೥ࠒ(εϚʔτϑΥϯ͕ ग़Δલʂ)͔Β͔͞Μʹ΍Β Ε͍ͯΔݚڀ 11
12. `abc<=]$^_VWXY • Laput et al., Synthetic Sensors: Towards General-Purpose Sensing (2017) • Ոఉ಺ͰͷߦಈΛΧϝϥΛ࢖ΘͣʹϚΠΫɺ੺֎ઢηϯαɺՃ଎౓ηϯαɺ࣓ྗܭͳͲΛ౥ࡌͨ͠ηϯαͰਪఆ 12 Synthetic Sensors: Towards General-Purpose Sensing https://youtu.be/aqbKrrru2co
13. defgBhij=?$^_VWkl • ߴྛΒ, ਍அπʔϧΛ༻͍ͨεϚʔτλοϓͷফඅిྗ৘ใʹΑΔਓͷߦಈਪఆਫ਼౓ ͷ࣮ূධՁ (2018) • εϚʔτλοϓͰऔಘͨ͠ফඅిྗྔ͔Βʮىচʯʮ֎ग़ʯʮؼ୐ʯʮब৸ʯΛਪఆ 13
14. m2nompCqC'rstuvwxYy • ૔ڮΒ, τΠϨοτϖʔύͷճసʹجͮ͘τΠϨ࢖༻ऀࣝผख๏ (2017) • τΠϨοτϖʔύʔͷਊʹ֯଎౓ηϯαΛઃஔ͠ɺ5ͭͷಛ௃ྔ(࢖༻ྔɺ ࣌ؒɺ଎౓ͷ࠷େ஋ɺ଎౓ͷฏۉ஋ɺ଎౓ͷ෼ࢄ)͔Β࢖༻ऀΛਪఆ 14
15. z{|}0123~•€•• 15 ͸͡Ίͯͷํ΁ | DFree - ഉᔔ༧ଌσόΠε/Toilet Timing Predicting Device https://dfree.biz/personal/beginners.html
16. VWXY‚0Ci‚G-21>C • ػցֶशʹΑΓ͞·͟·ͳߦಈೝ͕ࣝՄೳʹ • ࿙Εͯ΋໰୊ͳ͍ͱࢥͬͨσʔλ͔Βɺ༧ظ͠ͳ͍৘ใ(ߦಈཤྺͳͲ) ͕࿙ΕΔՄೳੑ΋͋Δ • σʔλͷѻ͍͸ɺ͜Ε·ͰҎ্ʹϓϥΠόγʔʹ഑ྀ͢Δඞཁ͕͋Δ 16
17. NOPQ‚ƒ„…† • ػցֶशͱը૾ॲཧ෼໺͸૬ޓʹີ઀ʹൃల • Α͘࢖ΘΕΔศརͳπʔϧΛ3ͭ঺հ • YOLO • MediaPipe • Teachable Machine 17
18. ‡ˆ‰ˆAŠ‹Œ•,/Žj•• 18 YOLO Open Images in New York https://youtu.be/69Ii3HjUiTM
19. 4•:58‘5M• • Google͕ఏڙ͢ ΔΦʔϓϯιʔε πʔϧ • Pythonɺ JavaScriptͳͲ͍ Ζ͍ΖͳݴޠͰϥ ΠϒϥϦ͕͋Δ 19 👀 Home - mediapipe https://google.github.io/mediapipe/
20. 4•:58‘5M•6’89:“'”•@ 20 ๭Ξχϝͷѱ໾͕࢖͏ʮͱ͓͖ͬͯͷख඼ʯͬΆ͍ېढ͕࢖͑Δʢؾ͕͢Δ͔΋͠Εͳ͍ʣWebΞϓϦ V3 ͷࢼ࡞ʢʁʣ https://youtu.be/wyOzbHk_tVE
21. 4•:58‘5M•6‘K“•'”•@ 21 ্൒਎Ͱԋ૗ | ProtoPedia https://protopedia.net/prototype/2840 👀
22. –•8L—8˜™•648L—59• • Google͕ఏڙ͢ΔػցֶशϞσϧੜ੒πʔϧɺWeb্ͰࢼͤΔ 22 👀 Teachable Machine https://teachablemachine.withgoogle.com/
23. UM?KV • Teachable MachineͰ ʮάʔʯʮνϣΩʯʮύʔʯ Λೝࣝͯ͠Έ·͠ΐ͏ • Teachable MachineͷWeb αΠτ͔Βʮ࢖ͬͯΈΔʯɹɹɹɹɹɹ ˠʮը૾ϓϩδΣΫτʯɹ ˠʮඪ४ͷը૾Ϟσϧʯ • ࡞ۀ࣌ؒ5෼ • ٳܜ΋͜ͷؒʹ͍ͯͩ͘͠͞
24. ,š?-'›œ 24 ৗઃలࣔʰΞφάϥͷ͏ͨɹʙফ͑ͨത࢜ͱ࢒͞Εͨ૷ஔʙʱ https://youtu.be/75jA7F8AXC4
25. ,š?-'›œ 25 ৗઃలࣔʮΞφάϥͷ͏ͨ―ফ͑ͨത࢜ͱ࢒͞Εͨ૷ஔʯ https://youtu.be/j1zLBXUoG3I
26. •omžCŸ ¡- • Πϯλʔωοτӽ͠ʹө૾ΛݟΔ͜ͱ͕Ͱ͖ΔΧϝϥ • ๷൜Χϝϥɺࣗ୐ϖοτ؂ࢹ etc. • ੬ऑͳύεϫʔυ΍όοΫυΞʹΑΔෆਖ਼ΞΫηεͷϦεΫ΋ 26 தࠃ੡ωοτϫʔΫΧϝϥʹ؅ཧऀݖݶΛୣऔͰ͖ΔόοΫυΞͷଘࡏ͕ใࠂ͞ΕΔ - GIGAZINE https://gigazine.net/news/20200207-xiongmai-backdoor/
27. ¢’ˆ~£¤ • Πϯλʔωοτ઀ଓػثͷݕࡧαΠτ 27
28. 59“•L8J • ύεϫʔυઃఆ͍ͯ͠ͳ͍ωοτϫʔΫΧϝϥͷө૾ΛूΊͨWebαΠτ΋͋Δ 28 ๷൜Χϝϥө૾Λͷ͖ͧݟͰ͖ͯ͠·͏ʮinsecamʯϨϏϡʔɺ೔ຊͷ๷൜Χϝϥ΋1000୆Ҏ্ϦΞϧλΠϜͰ౪ΈݟՄೳ - GIGAZINE https://gigazine.net/news/20200718-insecam/
29. ¤ˆ–¥¦§ • Πϯλʔωοτ઀ଓػثͷ੬ऑੑௐࠪ • ૯຿লɺࠃཱݚڀ։ൃ๏ਓ৘ใ௨৴ݚڀػߏ͕࣮ࢪ 29 NOTICEʛαΠόʔ߈ܸʹѱ༻͞ΕΔ͓ͦΕͷ͋ΔIoTػثͷௐࠪɺ஫ҙשىΛߦ͏ϓϩδΣΫτ https://notice.go.jp/
30. ¡-¨©'G-21>Cª« 30 JRେࡕӺϏϧͷʮإࣝผʯ࣮ূ࣮ݧɺϓϥΠόγʔ৵֐ͷݒ೦͔ΒԆظ | ೔ܦΫϩεςοΫʢxTECHʣ https://xtech.nikkei.com/it/article/NEWS/20140311/542723/ ຊ࣮ূ࣮ݧ͸ݪଇͱͯ͠ɺຽ๏ٴͼಠཱߦ੓๏ਓ౳ݸਓ৘ใอޢ๏ʹ͸ҧ൓͠ͳ͍ͱͷ݁࿦Λಘͨ ػߏ๏ʹج͖ͮઃཱ͞Εͨެతஂମͱͯ͠ͷઆ໌੹೚ΛՌ͍ͨͯ͠ͳ͍ɺͱͷ൷൑Λ໔Εͳ͍Ͱ͋Ζ͏ ө૾ηϯαʔ࢖༻େن໛࣮ূ࣮ݧݕ౼ҕһձ ௐࠪใࠂॻΑΓ
31. ¬S-®‚qC¯š/0Ci 31 ύʔιφϧσʔλɿݸਓʹؔ͢Δ৘ใ ݸਓ৘ใɿݸਓΛࣝผͰ͖Δ৘ใ ࢯ໊ ੜ೥݄೔ ॅॴ ݸਓͷଐੑ৘ใ Ҡಈɾߦಈɾߪങཤྺ ΢ΣΞϥϒϧػث͔Βऩूͨ͠৘ใ
32. qC¯š/0Ci'@AB12/°T±² • ܞଳి࿩ར༻ऀͷҐஔ৘ใΛ౷ܭతʹར༻ (ߪೖՄ) 32 ʲެࣜʳϞόΠϧۭؒ౷ܭ | Ґஔ৘ใͳͲͷϏοάσʔλΛར༻ͨ͠ਓޱ౷ܭ৘ใ https://mobaku.jp/ 👀
33. -®³V 33 ৘ใۜߦͱ͸ | ৘ใۜߦαʔϏεDprimeɿࡾඛ̛̪̟৴ୗۜߦ https://www.tr.mufg.jp/dprime/service/about/
34. -®³V • ৘ใΛ༬͚Δ(౉͢)ͷ͸ෆ҆ʁ • ೔ຊͰ࠷ॳʹͰ͖ͨۜߦ͸1873(໌࣏6)೥ͷୈҰ ࠃཱۜߦ • ೔ຊͰ࠷ॳʹͰ͖ͨ௨՟͸ 7ੈلޙ൒ͷ෋ຊમ 34
35. ´µ²¶ 35 NEC ൿີܭࢉ | NEC https://jpn.nec.com/secure-computation/index.html
36. G-21>C·<E¸j¹º‚»¼ • ๏཯ͷઢҾ͖͸֬ݻͨΔ΋ͷͰ͸ͳ͍ • ղऍʹ෯͕͋Δ(ͳͷͰࡋ൑ͰܾΊΔ) • ࣌୅എܠͷӨڹΛड͚ͯҐஔ͚͕ͮมΘΔ • Coinhiveࣄ݅ • ޿ࠂΛࡌͤΔ୅ΘΓʹ҉߸௨՟ϚΠχϯάΛߦ͏ɺWebαΠ τͷ৽͍͠ऩӹγεςϜCoinhive͕ҧ๏ͱ͞Εݕڍ • 1৹ɿແࡑɹɹ2৹ɿ༗ࡑɹɹ࠷ߴࡋɿແࡑ 36
37. ¢¤¢ • ଟ͘ͷύʔιφϧσʔλؚ͕·ΕΔ • ςΩετɺࣸਅɺಈը • ॴଐɺډॅ஍ɺझຯɺަ༑ؔ܎౳ 37
38. ½¾¿¢¤¢‚ÀÁ¿¢¤¢ 38 αʔό αʔό αʔό αʔό ूதܕ(طଘ)SNS Twitter, Instagram, Facebook౳ ෼ࢄܕSNS Mastodon, Misskey, Bluesky౳
39. ÀÁ¿¢¤¢ 39 misskey.io trpger.us nijimiss.moe • αʔόؒͰ৘ใΛ࿈ܞ = ҧ͏αʔόͷϢʔβ৘ใ΋औಘՄೳ Misskeyαʔό Misskeyαʔό Misskeyαʔό Misskey αʔόʔҰཡ https://join.misskey.page/ja-JP/instances
40. £LÂ5Ã5ÂđŘ6R48“ÂK:K9Æ645““Ç•ÄU • ActivityPubͱ͍͏ ϓϩτίϧͰαʔ όؒͷ࿈ܞ͕Մೳ • ूதܕͱಉ༷ͷ՝ ୊΋ൃੜ 40 Segment, AT Protocol (BlueSky Social)࢓༷ղઆ ʙ W3C DID࢓༷Λఴ͑ͯ ʙ - Speaker Dec k https://speakerdeck.com/gpsnmeajp/at-protocol-bluesky-social-shi-yang-jie-shuo-w3c-didshi-yang-wotian-ete
41. £–‘€KÂKLK™6R7™Å•“ÇÄU 41 Segment, AT Protocol (BlueSky Social)࢓༷ղઆ ʙ W3C DID࢓༷Λఴ͑ͯ ʙ - Speaker Dec k https://speakerdeck.com/gpsnmeajp/at-protocol-bluesky-social-shi-yang-jie-shuo-w3c-didshi-yang-wotian-ete • ATProtocolͱ͍͏ ϓϩτίϧͰ࿈ܞ • ΞΧ΢ϯτͷҠߦ ΋Մೳ • ࢓૊Έ͸ෳࡶɾɾɾ • ݱࡏઈࢍ։ൃத
42. È‚# • ৽͍͠΋ͷ ← ෆ҆ ← ࢓૊Έ͕Θ͔Βͳ͍ • ύʔιφϧσʔλ • AI • ϒϩοΫνΣʔϯ • etc. • ࢓૊ΈΛษڧͯ͠ෆ҆Λ෷১͠Α͏ 42
43. ÉÊÊ#-® 43 @shiropen2 AIͷҨిࢠ
44. ÉËt$ • খςετʹऔΓ૊ΜͰ͍ͩ͘͞ • 2ճडݧͰ͖·͢ • క੾ɿ࣍ճߨٛ։࢝࣌·Ͱ • ײ૝ॻ͍͍ͯͩ͘͞ (೚ҙɺ੒੷ʹؔ܎ͳ͠) 44