Slide 1

Slide 1 text

[email protected] 
 @babalablab ߴߍੜͱେֶੜͷͨΊͷ༵ۚಛผߨ࠲ 
 ਓؒͱਓ޻஌ೳͷڠಇ ೥݄೔ 
 ౦ژେֶڭཆֶ෦ֶࡍՊֶՊഅ৔ઇ೫ https://fontawesome.com https://openmoji.org

Slide 2

Slide 2 text

2 അ৔ઇ೫ʢ͹͹Ώ͖ͷʣ ౦ژେֶ૯߹จԽݚڀՊ޿ҬՊֶઐ߈ɾ।ڭत ത࢜ʢ৘ใཧ޻ֶʣ ઐ໳˔ਓ޻஌ೳ ػցֶश ݚڀτϐοΫ˔)VNBO$PNQVUBUJPO 
 ɹɹɹɹɹɹɹ)VNBO"*$PMMBCPSBUJPO ɹɹɹɹɹɹɹ)VNBOJOUIF-PPQ.BDIJOF-FBSOJOH ஶॻ˔ʮώϡʔϚϯίϯϐϡςʔγϣϯͱΫϥ΢υιʔγϯάʯ ɹɹɹɹʢߨஊࣾαΠΤϯςΟϑΟοΫʣ

Slide 3

Slide 3 text

Ϣʔβ͕ೖྗͨ͠จষʢϓϩϯϓτʣ͔Βը૾Λੜ੒͢Δ"*͕ొ৔ 
 ʢ%"--w& .JEKPVSOFZ 4UBCMF%J ff VTJPOͳͲʣ ਓ޻஌ೳʹͰ͖Δ͜ͱ͕૿͍͑ͯΔ 3 https://huggingface.co/spaces/stabilityai/stable-di ff usion https://huggingface.co/spaces/stabilityai/stable-di ff usion 
 Λར༻ͯ͠ߨࢣ͕࡞੒

Slide 4

Slide 4 text

ਓ޻஌ೳʹͰ͖Δ͜ͱ͕૿͍͑ͯΔ 4 ྲྀெͳจষͰ࣭໰ʹ౴͑ͯ͘ΕΔର࿩"*͕ొ৔ʢ$IBU(15ͳͲʣ https://chat.openai.com/ Λར༻ͯ͠ߨࢣ͕࡞੒

Slide 5

Slide 5 text

ਓ޻஌ೳʹͰ͖Δ͜ͱ͕૿͍͑ͯΔ 5 $IBU(15͸1ZUIPOίʔυ΋ग़ྗͰ͖Δ https://chat.openai.com/ Λར༻ͯ͠ߨࢣ͕࡞੒

Slide 6

Slide 6 text

͍·ͷਓ޻஌ೳ͸໌Β͔ͳؒҧ͍Λ͢Δ͜ͱ͕͋Δ 6 https://twitter.com/mahimahi/status/1599384548571516929

Slide 7

Slide 7 text

͍·ͷਓ޻஌ೳ͸ਓ͕ؒ͠ͳ͍ؒҧ͍Λ͢Δ͜ͱ͕͋Δ 7 https://www.bbc.com/news/technology-33347866 (PPHMF1IPUPͷ"*͕ 
 ΞϑϦΧܥͷஉঁʹޡͬͯ ʮΰϦϥʯͱλά෇͚

Slide 8

Slide 8 text

͍·ͷਓ޻஌ೳ͸ภͬͨ݁ՌΛฦ͢͜ͱ͕͋Δ 8 https://twitter.com/favorite_Bonsai/status/1564754888386543616 ʮαʔϞϯϥϯʢࡪͷ૎্ʣʯͷ ࣮ࡍͷࣸਅ https://www.worldatlas.com/articles/best-places-to-see-the-salmon-run.html

Slide 9

Slide 9 text

͍·ͷਓ޻஌ೳ͸ภͬͨ݁ՌΛฦ͢͜ͱ͕͋Δ 9 https://twitter.com/zatazata/status/1572376514364346370

Slide 10

Slide 10 text

͍·ͷਓ޻஌ೳ͸ෳࡶͳจষͷཧղ͕ۤखͰ͋Δ 10 αϦʔ͸ΧΰΛ͍࣋ͬͯ·͢ɻΞϯ͸ശΛ͍࣋ͬͯ·͢ɻαϦʔ͸ ϏʔۄΛ͍࣋ͬͯ·͢ɻαϦʔ͸ϏʔۄΛࣗ෼ͷΧΰʹ࢓෣͍·͠ ͨɻαϦʔ͸֎ʹग़͔͚·ͨ͠ɻΞϯ͸ɺΧΰ͔ΒϏʔۄΛऔΓग़ ͠ɺࣗ෼ͷശʹ࢓෣͍·ͨ͠ɻαϦʔ͕໭͖ͬͯ·ͨ͠ɻαϦʔ͸ ϏʔۄͰ༡΅͏ͱ͍ͯ͠·͢ɻͯ͞ɺαϦʔ͸Ͳ͜Λ୳͢Ͱ͠ΐ͏ ͔ɻ S. Baron-Cohen et al.: Does the Autistic Child have a “Theory of mind” ? Cognition, Vol. 21, Issue 1, 1985. ೔ຊޠ༁͸ߨࢣʹΑΔɻ

Slide 11

Slide 11 text

͍·ͷਓ޻஌ೳ͸ෳࡶͳจষͷཧղ͕ۤखͰ͋Δ 11 ࣭໰จͷग़యɿS. Baron-Cohen et al.: Does the Autistic Child have a “Theory of mind” ? Cognition, Vol. 21, Issue 1, 1985. ճ౴จ͸https://beta.openai.com/playground Λར༻ͯ͠ߨࢣ͕࡞੒ɻ

Slide 12

Slide 12 text

ྙཧతͳ໰୊͸ਓ޻஌ೳʹ͸ܾΊͯ΄͘͠ͳ͍ 12 https://www.moralmachine.net

Slide 13

Slide 13 text

13 ώϡʔϚϯίϯϐϡςʔγϣϯ ਓؒͱਓ޻஌ೳͷڠಇʹΑΓ 
 ͲͪΒ͔Ұํ͚ͩͰ͸೉͍͠໰୊Λղܾ͢Δ

Slide 14

Slide 14 text

14 ௨ৗͷਓ޻஌ೳγεςϜ ώϡʔϚϯίϯϐϡςʔγϣϯ ਓؒΛܭࢉࢿݯͱͯ͠औΓࠐΉ

Slide 15

Slide 15 text

7J[8J[ࢹ֮ো͕͍ऀͷ࣭໰Ԡ౴γεςϜ 15 -tap to take a photo. -tap to begin recording your question and again to stop. side, User ? Web Server - Database - Local Client Remote Services and Worker Interface 4UFQϢʔβ͕࣭໰Λ౤ߘ ίʔϯͷ؈͸ͲΕʁ 4UFQγεςϜ಺෦ͷਓ͕ؒճ౴ Ұ൪ӈͷ؈Ͱ͢ J. P. Bigham et al.: VizWiz: Nearly Real-time Answers to Visual Questions. In Proceedings of the 23rd annual ACM symposium on User interface software and technology (UIST), 2010.

Slide 16

Slide 16 text

;FOTPSTਓؒΛ࢖ͬͨ؂ࢹγεςϜ 16 https://www.youtube.com/watch?v=aYHzG9uXQ6k

Slide 17

Slide 17 text

;FOTPSTਓؒΛ࢖ͬͨ؂ࢹγεςϜ 17 G Laput et al. Zensors: Adaptive, Rapidly Deployable, Human-Intelligent Sensor Feeds. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems (CHI), 2015. How many glasses need a re fi ll? or… I can’t Tell https://www.dreamstime.com/empty-white- wine-glasses-table-restaurant-bar-setting- close-up-alcohol-image215682647 ෼͓͖ʹਓؒʹը૾Λૹ৴ͯ͠ 
 ໰͍߹ΘͤΔ

Slide 18

Slide 18 text

;FOTPSTਓؒΛ࢖ͬͨ؂ࢹγεςϜ 18 ਓؒͷճ౴Λ༻͍ͯ 
 ਓ޻஌ೳΛֶश 
 
 
 े෼ͳਫ਼౓ʹୡͨ͠Β 
 ਓؒͰ͸ͳ͘ਓ޻஌ೳΛར༻ G. Laput et al. Zensors: Adaptive, Rapidly Deployable, Human-Intelligent Sensor Feeds. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems (CHI), 2015.

Slide 19

Slide 19 text

ෆಛఆଟ਺ͷਓʹ୯७࡞ۀΛൃ஫Ͱ͖Δɼ 
 ΦϯϥΠϯϓϥοτϑΥʔϜʮΫϥ΢υιʔγϯάʯΛ࢖ͬͯ 
 ώϡʔϚϯίϯϐϡςʔγϣϯʹࢀՃ͢ΔਓؒΛूΊΔ Ϋϥ΢υιʔγϯάʹΑΓࢀՃऀΛืΔ 19 Ϋϥ΢υιʔγϯά "NB[PO.FDIBOJDBM5VSL 
 ϥϯαʔζͳͲ ࡞ۀΛड஫ ใुΛ੥ٻ ࡞ۀΛൃ஫ අ༻ͷࢧ෷͍

Slide 20

Slide 20 text

࡞ۀը໘ Ϋϥ΢υιʔγϯάʹΑΓࢀՃऀΛืΔ 20 Ϩγʔτͷॻ͖ى͜͠ ʢʣ ֆըͷײ৘ϥϕϧ෇༩ ʢʣ "NB[PO.FDIBOJDBM5VSLͷը໘ྫ ґཔҰཡ https://www.mturk.com https://www.mturk.com

Slide 21

Slide 21 text

Ϋϥ΢υιʔγϯάʹΑΓࢀՃऀΛืΔ 21 "NB[PO.FDIBOJDBM5VSLͰ͸"1*͕ఏڙ͞Ε͍ͯͯ 
 ਓؒ΁ͷ໰͍߹ΘͤΛࣗಈੜ੒Ͱ͖Δ https://qiita.com/ssmsaito/items/c0c514d76abcd532b59e

Slide 22

Slide 22 text

ώϡʔϚϯίϯϐϡςʔγϣϯͷ՝୊ 22 ҙࢥͷ͋ΔਓؒΛώϡʔϚϯίϯϐϡςʔγϣϯʹࢀՃͤ͞Δʹ͸ 
 ద੾ͳಈػ͚͕ͮඞཁ ᶃಈػ͚ͮ ਓؒ͸ʮৗʹɾ୭Ͱ΋ʯਖ਼͍͠౴͑Λฦ͢ͱ͸ݶΒͳ͍ͨΊ 
 ඼࣭Λอূ͢ΔͨΊͷ޻෉͕ඞཁ ᶄ඼࣭อূ

Slide 23

Slide 23 text

SF$"15$)"ɿॻ੶ͷจࣈೝࣝγεςϜʹਓؒΛ૊ΈࠐΉ 23 morning 4UFQ݁Ռ͕ෆҰகͷͱ͖ʹਓؒʹ໰͍߹ΘͤΔ moroiog morpipg 4UFQॻ੶தͷจࣈΛ̎ͭͷਓ޻஌ೳγεςϜʹೝࣝͤ͞Δ L. von Ahn et al.: reCAPTCHA: Human-Based Character Recognition via Web Security Measures. Science, Vol. 321, Issue 589, 2008. ਓ޻஌ೳʹ͸೉͍͠จࣈΛਓؒʹೝࣝͤ͞Δ

Slide 24

Slide 24 text

ΞΫηεݖΛใुͱͯ͠ਓؒΛจࣈೝࣝ࡞ۀʹࢀՃͤ͞Δ 24 ΞΫηεΛ 
 ڐՄͯ͠ 
 ͍ͩ͘͞ʂ ਓؒͷΞΫηε͸ڐՄ͠·͢ɽ 
 ϘοτͷΞΫηε͸ڐՄ͠·ͤΜɽ ͋ͳ͕ͨਓؒͩͱূ໌͢ΔͨΊʹ 
 ͜ͷจࣈΛೝࣝ͠ͳ͍͞

Slide 25

Slide 25 text

ʮਓ͔ؒϘοτ͔ͷ൑ఆʯͱʮॻ੶ͷจࣈೝࣝʯΛಉ࣌ʹߦ͏ 25 ᶃਓ͔ؒϘοτ͔ͷ൑ఆɹᶄॻ੶ͷจࣈೝࣝ SF$"15$)"ͷ໨త Type the word Type the word reCAPTCHA ೋͭͷจࣈΛೝࣝ͢ΔΑ͏ʹࢦࣔ͢Δ ˞ͲͪΒ͕ਖ਼ղط஌͔͸఻͑ͳ͍ ਖ਼ղط஌ ਖ਼ղະ஌ L. von Ahn et al.: reCAPTCHA: Human-Based Character Recognition via Web Security Measures. Science, Vol. 321, Issue 589, 2008. ᶄʮਓؒɹʯͷ݁Ռ͚ͩΛ 
 ɹೝࣝ݁Ռͱͯ͠࠾༻ ᶃɹਖ਼ղͳΒˠਓؒ 
 ɹෆਖ਼ղͳΒˠϘοτ

Slide 26

Slide 26 text

඼࣭อূͷͨΊʹʮਓؒʯͷճ౴͚ͩ࠾༻͢Δ 26 overlooks morning reCAPTCHA ਖ਼ղط஌ ਖ਼ղະ஌ L. von Ahn et al.: reCAPTCHA: Human-Based Character Recognition via Web Security Measures. Science, Vol. 321, Issue 589, 2008. ਖ਼ղͳͷͰ 
 ʮਓؒɹʯͱ൑ఆ ʮਓؒɹʯͷೝࣝ݁ՌͳͷͰ 
 ࠾༻⭕

Slide 27

Slide 27 text

ʮϘοτʯͷճ౴͸࠾༻͠ͳ͍ 27 overfooRs morpipg reCAPTCHA ਖ਼ղط஌ ਖ਼ղະ஌ L. von Ahn et al.: reCAPTCHA: Human-Based Character Recognition via Web Security Measures. Science, Vol. 321, Issue 589, 2008. ෆਖ਼ղͳͷͰ 
 ʮϘοτɹʯͱ൑ఆ ʮϘοτɹʯͷճ౴ͳͷͰ 
 ෆ࠾༻❌

Slide 28

Slide 28 text

˔ ʮਓؒʯͳΒ͹ਖ਼ղະ஌ͷ໰୊ʹ΋ਖ਼͘͠౴͑ΔՄೳੑ͕ߴ͍ ˔ Ұਓ͚ͩʹ໰͍߹ΘͤΔͱɼͦͷਓ͕ؒҧ͑ΔՄೳੑ͕͋Δ ˔ ಉ͡จࣈྻΛෳ਺ਓʹ໰͍߹ΘͤɼҰఆ਺ͷճ౴͕Ұகͨ͠ͱ͖͚ͩ 
 ࠷ऴతͳೝࣝ݁Ռͱͯ͠࠾༻͢Δʢฒྻ໰߹ͤʣ ඼࣭อূͷͨΊʹಉ͡จࣈྻΛෳ਺ਓʹ໰͍߹ΘͤΔ 28 ਓͷճ౴͕Ұகͨ͠ͷͰ 
 morning 
 Λೝࣝ݁Ռͱͯ͠࠾༻ morning morning morping morning

Slide 29

Slide 29 text

&41ήʔϜɿը૾΁ͷϥϕϧ෇͚ΛήʔϜԽ 29 ਓ޻஌ೳͷֶशʹ༻͍ΔͨΊɼը૾ͷλάΛਓ͔ؒΒूΊΔ ˔ ૬खϓϨΠϠʔͱಉ͡λάΛ౴ ͑ͨΒϙΠϯτ͕΋Β͑Δ 
 ʮήʔϜʯʹ͢Δ͜ͱͰ 
 ࢀՃΛଅ͢ ˔ σλϥϝͳճ౴ͷ๷ࢭʹ΋ 
 ໾ཱͭ L. von Ahn et al.: Designing games with a Purpose. Communication of the ACM, Vol.51, Issue 8, 2008

Slide 30

Slide 30 text

˔ λϯύΫ࣭ͷߏ଄༧ଌ͸ͦͷػೳΛ஌Δͷʹॏཁ͕ͩਓ޻஌ೳʹ͸ࠔ೉ ˔ ήʔϜʹ͢Δ͜ͱͰେྔͷਓؒʹղ͔ͤͨͱ͜Ζ 
 ೥Ҏ্ະղܾͩͬͨ໰୊ΛϓϨΠϠʔୡ͕೔Ҏ಺ʹղ͍ͨ 'PME*UλϯύΫ࣭ͷߏ଄༧ଌΛήʔϜԽ͠ਓؒʹղ͔ͤΔ 30 ཱମߏ଄ ഑ྻ https://fold.it/ S. Cooper et al.: Predicting Protein Structures with a Multiplayer Online Game. Nature, Vol. 466, No. 7307, 2010. ϓϨΠϠʔ͸ߴείΞΛ 
 ૂ͍ߏ଄ΛมԽͤ͞Δ

Slide 31

Slide 31 text

'PME*UλϯύΫ࣭ͷߏ଄༧ଌΛήʔϜԽ͠ਓؒʹղ͔ͤΔ 31 https://www.youtube.com/watch?v=DvYFjo3vC-k

Slide 32

Slide 32 text

˔ ໺ௗͷ؍࡯ه࿥Λऩू͢ΔͷʹਓؒΛ׆༻ ˔ ࢀՃऀ͸ࣗൃతʹ໺ௗΛ؍࡯͠छྨɾ਺ɾࣸਅɾԻ੠౳Λه࿥͢Δ ˔ ࢀՃऀʹ޿ൣғΛΧόʔͤ͞ΔͨΊʹ؍࡯৔ॴΛਪન͢Δ F#JSE໺ௗѪ޷ՈΛར༻ͯ͠໺ௗͷσʔλΛूΊΔ 32 https://ebird.org/home

Slide 33

Slide 33 text

ଟ਺ܾ͸શһʹಉ͡ॏΈΛ༩͑Δ͕ɼਖ਼͍͠౴͑ΛಘΔͨΊʹ͸ 
 ৴པੑͷߴ͍ਓʹߴ͍ॏΈΛ༩͍͑ͨ ฒྻ໰߹ͤʹΑΔ඼࣭อূ 33 ࣸਅʹௗ͕ࣸͬͯ·͔͢ʁ YES NO YES YES YES ଟ਺ܾͷ݁Ռ͸YES ࣸਅʹௗ͕ࣸͬͯ·͔͢ʁ YES YES NO ଟ਺ܾͷ݁Ռ͸YES YES NO

Slide 34

Slide 34 text

ੜెͷճ౴͚͔ͩΒࢼݧͷਖ਼ղΛ༧ଌ͢ΔΑ͏ͳ΋ͷ ճ౴͔Β֤ࣗͷ৴པੑΛਪఆ֤͠໰୊ͷਖ਼ղΛ༧ଌ͢Δ 34 ਖ਼ղ ໰୊ NO YES YES YES YES YES NO YES YES YES YES NO ? ? ? “Is a bird in 
 the picture?" https://commons.wikimedia.org/wiki/File:American_ fl amingo_(Phoenicopterus_ruber).JPG

Slide 35

Slide 35 text

ti ਖ਼ղ YES ti = NO ti = yij βj ճ౴ ճ౴Ϟσϧʢ໰୊ ʹର͢Δճ౴ऀ ͷճ౴ʣ i j αj Pr[yij ∣ ti = 1] = αyij j (1 − αj )(1−yij ) Pr[yij ∣ ti = 0] = β(1−yij ) j (1 − βj )yij ճ౴ϞσϧΛߏஙͯ͠ਖ਼ղ༧ଌʹ༻͍Δ 35 ɿճ౴ऀ ͕ਖ਼ղ͕YESͷ໰୊ʹYESͱ౴͑Δ֬཰ αj j ɿճ౴ऀ ͕ਖ਼ղ͕NOͷ໰୊ʹNOͱ౴͑Δ֬཰ βj j ճ౴ऀͷ৴པੑύϥϝʔλʢࠞಉߦྻʣ ճ౴ YES NO ਖ਼ 
 ղ YES NO αj βj 1 − αj 1 − βj ࠞಉߦྻ A. P. Dawid and A. M. Skene: Maximum likelihood estimation of observer error-rates using the EM algorithm. Journal of the Royal Statistical Society, Series C (Applied Statistics), 1979.

Slide 36

Slide 36 text

˔ ଞਓͷ౴͑Λ࢖͍ղ͔ͤΔ͜ͱͰਖ਼͍͠౴͑ΛҾ͖ग़͢ ˙ 4PZMFOUਓؒΛ࢖ͬͨจষߍਖ਼γεςϜ 'JOE໰୊ͷ͋ΔՕॴͷݕग़ 'JYߍਖ਼ͷ࣮ࢪ 7FSJGZߍਖ਼͕ਖ਼͍͔֬͠ೝ ௚ྻ໰߹ͤʹΑΔ඼࣭อূ 36 When the crowd is finished, Soylent calls out the edited sections with a purple dashed underline. If the user clicks on the error, a drop-down menu explains the problem and offers a list of alternatives. By clicking on the desired alter- native, the user replaces the incorrect text with an option of Figure 2. Crowdproof is a human-augmented proofreader. The drop-down explains the problem (blue title) and suggests fixes (gold selection). M. S. Bernstein et al.: Soylent: a Word Processor with a Crowd Inside. Communications of the ACM, Vol. 58, Issue 8, 2015. Find Fix Verify

Slide 37

Slide 37 text

˔ %"31"3FE#BMMPPO$IBMMFOHF੺͍෩ધΛ୳͢ίϯςετ ˙ શถՕॴʹ੺͍෩ધ🎈Λઃஔ ˙ ࠷ॳʹશͯͷ੺͍෩ધΛݟ͚ͭͨνʔϜʹ৆͕ۚࢧ෷ΘΕΔ ˙ ༏উνʔϜ͸ʮ෩ધΛݟ͚ͭͦ͏ͳਓʯΛଞऀ͔ΒͷਪનͰൃݟ ਪનʹΑΔ඼࣭อূ 37 contributed articles platform for viral collaboration that used recursive incentives to align the public’s interest with the goal of win- ning the Challenge. This approach was inspired by the work of Peter S. Dodds et al.5 that found that success in us- ing social networks to tackle widely distributed search problems depends on individual incentives. The work of Mason and Watts7 also informed the use of financial incentives to motivate crowdsourcing productivity. The MIT team’s winning strategy was to use the prize money as a finan- cial incentive structure rewarding not only the people who correctly located balloons but also those connecting the finder to the MIT team. Should the team win, they would allocate $4,000 in prize money to each balloon. They promised $2,000 per balloon to the Figure 1. Locations in the DARPA Red Balloon Challenge. Figure 2. Example recursive incentive-structure process for the MIT team. J. C. Tang et al.: Re fl ecting on the DARPA Red Balloon Challenge. Communications of the ACM, Vol. 54, Issue 11, 2011.

Slide 38

Slide 38 text

ඃਪનऀ͕෩ધΛൃݟͨ͠Βਪનऀʹ΋ใु͕෼഑͞ΕΔΑ͏ʹઃܭ ਪનʹΑΔ඼࣭อূ 38 ਪન ਪન ਪન ෩ધΛൃݟʂ ͷใु ͷใु ෩ધΛൃݟʂ ͷใु ͷใु ͷใु

Slide 39

Slide 39 text

˔ ώϡʔϚϯίϯϐϡςʔγϣϯ͸ɼਓ޻஌ೳͱਓؒͷڠಇʹΑΓ 
 ͜ͱͰͲͪΒ͔Ұํ͚ͩͰ͸ղܾ͕೉͍͠໰୊Λղܾ͢Δ ˔ ώϡʔϚϯίϯϐϡςʔγϣϯͷ՝୊͸ಈػ͚ͮͱ඼࣭อূ ˙ ಈػ͚ͮ 
 ήʔϜԽ΍ɼΞΫηε੍ޚɾ޷ح৺ɾۚમͳͲΛใुʹ͢Δ͜ͱͰ ਓؒͷࢀՃΛଅ͢ ˙ ඼࣭อূ 
 ௚ྻ໰߹ͤɾฒྻ໰߹ͤʹΑΓෳ਺ਓΛ૊Έ߹ΘͤͨΓɼ 
 ਪનʹΑΓ৴པͰ͖ΔਓΛݟ͚ͭͯɼਖ਼͍͠ճ౴ͷ֫ಘΛ໨ࢦ͢ ·ͱΊ 39