Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
凄い力
Search
Sponsored
·
Your Podcast. Everywhere. Effortlessly.
Share. Educate. Inspire. Entertain. You do you. We'll handle the rest.
→
yuhr
December 17, 2017
Research
300
0
Share
凄い力
人工言語学のための世界劇場論
yuhr
December 17, 2017
Other Decks in Research
See All in Research
Self-Hosted WebAssembly Runtime for Runtime-Neutral Checkpoint/Restore in Edge–Cloud Continuum
chikuwait
0
470
"主観で終わらせない"定性データ活用 ― プロダクトディスカバリーを加速させるインサイトマネジメント / Utilizing qualitative data that "doesn't end with subjectivity" - Insight management that accelerates product discovery
kaminashi
16
24k
FUSE-RSVLM: Feature Fusion Vision-Language Model for Remote Sensing
satai
3
470
台湾モデルに学ぶ詐欺広告対策:市民参加の必要性
dd2030
0
300
「車1割削減、渋滞半減、公共交通2倍」を 熊本から岡山へ@RACDA設立30周年記念都市交通フォーラム2026
trafficbrain
1
970
Collective Predictive Coding and World Models in LLMs: A System 0/1/2/3 Perspective on Hierarchical Physical AI (IEEE SII 2026 Plenary Talk)
tanichu
1
360
Can We Teach Logical Reasoning to LLMs? – An Approach Using Synthetic Corpora (AAAI 2026 bridge keynote)
morishtr
1
200
東京大学工学部計数工学科、計数工学特別講義の説明資料
kikuzo
0
210
ペットのかわいい瞬間を撮影する オートシャッターAIアプリへの スマートラベリングの適用
mssmkmr
0
460
AIを叩き台として、 「検証」から「共創」へと進化するリサーチ
mela_dayo
0
240
世界モデルにおける分布外データ対応の方法論
koukyo1994
7
2.1k
Unified Audio Source Separation (Defense Slides)
kohei_1979
1
580
Featured
See All Featured
Stewardship and Sustainability of Urban and Community Forests
pwiseman
0
180
First, design no harm
axbom
PRO
2
1.2k
Accessibility Awareness
sabderemane
0
100
brightonSEO & MeasureFest 2025 - Christian Goodrich - Winning strategies for Black Friday CRO & PPC
cargoodrich
3
680
Impact Scores and Hybrid Strategies: The future of link building
tamaranovitovic
0
260
Become a Pro
speakerdeck
PRO
31
5.9k
Jess Joyce - The Pitfalls of Following Frameworks
techseoconnect
PRO
1
130
HU Berlin: Industrial-Strength Natural Language Processing with spaCy and Prodigy
inesmontani
PRO
0
320
Lightning talk: Run Django tests with GitHub Actions
sabderemane
0
170
Leveraging LLMs for student feedback in introductory data science courses - posit::conf(2025)
minecr
1
230
A better future with KSS
kneath
240
18k
Rebuilding a faster, lazier Slack
samanthasiow
85
9.5k
Transcript
ਓݴޠֶͷͨΊͷੈքܶ ੌ͍ྗᶗ
͜ͷੈɺ உঁΈͳऀͩɻ rrrr8γΣΠΫεϐΞʰ͓ؾʹঌ͢··ʱ
͢͜ͱ ▸ ϋΠσΨʔͷاɾམ ▸ ΰοϑϚϯͷׂԋٕɾׂڑ ▸ ϑϩΠτͱϥΧϯͷਫ਼ਆੳ ▸ ψʔιϩδʔ ▸
ΰʔϧυϚϯͷੌ͍ྗ ▸ ձੳɾஊੳ ▸ ରऀɺ࡞͍ͯ͠ΔੈքɺจԽɺऀੜ͕͋ΔਓɻਓؒͱͲ͏͍͏ͷ͔ Λ͍͖ࣔͯ͠ɺͦͷ্Ͱɺ͋ͳͨͷ࡞ੈքͰͲ͏ͳͷ͔ɺͱ͏ ▸ ݴޠͦͷͷʹ͍ͭͯ࣍ճʹճͯ͠ɺࠓճਓؒʹ͍ͭͯͷ
اͱམ
اͱམ&/5863'7&3'"--&/ ඃ ▸ ݱଘࡏ ▸ ଘࡏʢଘࡏ͢Δ͜ͱʣͱଘࡏऀʢଘࡏ͢Δͷʣͷ݁અɻ ਓؒͷຊੑͷΑ͏ͳͷɻݱଘࡏͷଘࡏऀత༷ଶ͕͍ΘΏ Δਓؒʹ͋ͨΔ ▸ ਓ͕ؒࣗબΜͩΓ࡞ͬͨΓͨ͠Θ͚Ͱͳ͍ੈքʹ൱Ԡ
ͳ͛͘ࠐ·Εͯ͠·͍ͬͯΔଘࡏͰ͋Δ ▸ ݱଘࡏͷඃੑͷͨΊɺ͔ͦ͜Βੜ͡Δࢮɺ͓ΑͼͦΕͷ ֮ޛෆ҆ਓؒͷ݅Ͱ͋Δ
اͱམ&/5863'7&3'"--&/ اͱམ ▸ ඃੑΛ౿·্͑ͨͰɺͦͷͳ͔ͰࣗΒͷՄೳੑΛ୳Δ͜ͱ Λاʢاʣͱ͍͏ ▸ ฏͨ͘ݴ͑ࣗݾ࣮ݱͷ͜ͱ ▸ ࣗΒ͕ඃతଘࡏͰ͋Δ͜ͱʢࢮͷ֮ޛɾઌۦత֮ޛੑʣ ΛΕɺͦΕΛΕͨ͜ͱ͑͞Εɺ༩͑ΒΕׂͨʢ
ৗੜ׆ʣʹຒ͢Δ͜ͱΛམʢᰑམʣͱ͍͏ ▸ ʮੈքͱৗతݱଘࡏ͕ԋ͍ͯ͡ΔԋܶͰ͋Δʯ
υϥϚπ ϧΪʔ
υϥϚπϧΪʔ%3"."563(: υϥϚπϧΪʔͱ͍͏ޠͷ༻๏ ▸ ʮ࡞ܶ๏ʯʮԋܶతͳٕ๏ʯͳͲͱ͍͏ҙຯͷҰൠ໊ࢺͩ ͕ɺࣾձֶʹ͓͍ͯಛʹΰοϑϚϯͷ࣋ͪࠐΜͩԋܶత ੳํ๏Λࢦ͢ɻࠓճ͜ͷ ▸ ࣾձతঢ়گΛܶʹݟཱͯɺਓؒͷࣾձత૬ޓߦҝΛɺԋܶ Λੳ͢Δ͔ͷΑ͏ʹੳ͢Δ
υϥϚπϧΪʔ%3"."563(: ׂԋٕͱׂڑ ▸ ׂظ ▸ ϨετϥϯͷΣΠλʔɾΣΠτϨεɺҩऀɺڭࢣͳͲɺਓԿΒ͔ͷׂͱ͠ ͯͷৼ͍Λৗʹظ͞Ε͍ͯΔɻͯ͢ʹԠ͑Α͏ͱ͢Εࣗݾໃ६ʹؕΔෳ ͷׂظʹׂ͓͍ͯᷤ౻͕ੜ͡Δ ▸ ׂԋٕ
▸ ͋Δ৬ۀʹब͘͜ͱɺͦͷ৬ۀͷਓͱͯ͠ৼ͏͜ͱɻ৬ۀ͚ͩͰͳ͘ɺͯ͢ ͷΞΠσϯςΟςΟͷຊ࣭ԋٕͰ͋ΔɻͦͷͨΊʹɺԋٕΛݟΔͷΛඞཁͱ ͢Δ ▸ ׂڑ ▸ ීஈͷࣗͱׂͱͷڑɻ࠷ॳΕ͍ͯΔ͕ɺԋ͍ͯ͡Δ͏ͪʹ͍͍ۙͮͯ͘
ਫ਼ਆੳ
ਫ਼ਆੳ14:$)0"/"-:4*4 ਫ਼ਆੳͱ ▸ ਫ਼ਆҩֶͱҟͳΔɻױऀͱͷରʹΑͬͯɺױऀͷ৺ͷ Λੳ͠ɺױऀࣗʹΑΔղܾʹಋ͜͏ͱ͢Δ ▸ ਫ਼ਆੳͷױऀʹେ͖͚ͯ͘ਆܦͱਫ਼ਆපͷೋ͕ͭ ͋Δ ▸ ҩֶ͔ΒΘΓͱന͍ͰݟΒΕ͍ͯΔΒ͍͕͠ɺਫ਼ਆੳ
·ͨਫ਼ਆҩֶΛܝ͛Δ͜ͱ͕͋ΔʢಛʹϨΠϯʣ
ਫ਼ਆੳ14:$)0"/"-:4*4 ϑϩΠτ ▸ உࠜͷϝλϑΝʔ ▸ ਫ਼ਆੳͷ࢝ऀɻਆܦʢώεςϦʔɾڧഭਆܦʣΛੳͷରͱ͠ɺਫ਼ਆ පੳෆՄೳͱͨ͠ɻਆܦɺͦͷͯ͢ΛऔΓআ͘͜ͱෆՄೳͱͷཱ ▸ 'PSU%Bʢ͍ͳ͍͍ͳ͍͋ʣ༡ͼ ▸
ࢳר͖ΛΧʔςϯ͖ͷϕουʹ͛ࠐΜͩΓʢʮΦʔΦʔΦʔΦʔʯʣɺ Ҿͬுͬͯखݩʹͨ͠Γʢʮμʔʯʣ͢Δ༡ͼΛ܁Γฦ͢ɻࢳר͖Λʹ ݟཱ͍ͯͯΔ ▸ ڸʹөͬͨ࢟Λݟ͑ͳ͘͢Δ͜ͱʹΑͬͯɺ൴ࣗࣗΛফ͢͜ͱ͕Ͱ͖ͨ ▸ চ·Ͱಧ͔ͳ͍େ͖ͳڸʹөͬͨࣗͷ࢟Λɺ͔͕Έ͜ΜͰফ͠ɺʮϕΠ ϏʔɺΦʔΦʔΦʔʂʯͱݴ͍ͬͯͨ
ਫ਼ਆੳ14:$)0"/"-:4*4 ϥΧϯ ▸ ϑϩΠτʹ͙࣍ਫ਼ਆੳͷେՈɻਫ਼ਆපΛਫ਼ਆੳͷൣᙝʹೖΕͨɻ ͨͩ͋͘͠·Ͱਆܦͱਫ਼ਆපͱΛݫີʹ۠ผ͠ɺରԠΛ͚Δɻϑ ϩΠτಉ༷ɺΛͯ͢औΓআ͘͜ͱෆՄೳɻʮঢ়ͷͳ͍ओମ ͳ͍ʯ ▸ ʮཉଞऀͷཉͰ͋Δʯ ▸
͕تͿ͜ͱΛͯ͠تͿࢠڙɻͷ༪Ίݴ༿ͱ͍͏ݴޠͷޮՌ ▸ ڸ૾ஈ֊ ▸ ओମଞऀͷ࢟ʹಉҰԽ͢Δ͜ͱͰࣗͷମΠϝʔδʢࣗզʣΛ ࡞Γ্͛Δɻࣗզଞऀʹ༩͑ΒΕͨͷ
ਫ਼ਆੳ14:$)0"/"-:4*4 ϨΠϯ ▸ ਫ਼ਆҩֶͷҰ֯ɻݴ͏͜ͱ͕͍͍ͪͪΤϞ͍ ▸ ʮ͕ࣗʪͨͩͷʫͰ͋Δ;ΓΛ͠Α͏ͱࣦͯ͠ഊ͢Δ ࡾࡀࣇɺඞͣোΛ͜͏ΉΔʯ ▸ ଘࡏͱ֮͞ΕΔ͜ͱɻ͕෦͔Β͍ͳ͘ͳΔͱٽ ͖ग़͢ࢠڙɻʹిؾΛ͚͕ͭͨΓɺΔ·Ͱ྆ʹͦ
ʹ͍ͯΒ͍͕ͨΔ
ψʔιϩδʔ
ψʔιϩδʔ/0040-0(: ψʔιϩδʔͱ ▸ ਫ਼ਆͱ࣭ͷ߹ΛਤΔӉɻాએͷఏএʹΑΔɻ ඇৗʹൣғʹΘͨΔٞͰ͋Γɺ֓આࠔ ▸ ओͳςʔθʮૉཻࢠͱਓؒͷҙࣝͰ͋Δʯʮۭؒͷߏ ਓؒͷແҙࣝͷߏͰ͋Δʯʮ࣭ࣗݾͱଞऀͷ݁߹ ʹΑͬͯੜ·Ε͍ͯΔʯʮަ؍Λҙຯ͢Δʯ ▸
औΓ্͛ΔͷɺओମͷҐஔɺۭؒʹ͍ͭͯ
ψʔιϩδʔ/0040-0(: ओମͷҐஔ ▸ ψʔιϩδʔɺࣗݾͷओମͷҐஔલʹ͋Δͱ͍͏ɻࣗ ͕ݟ͍ͯΔํͷແݶԕɻ͘·͙ͬ͢ͷ࿓ԼΛา͍ͯ ͍Δͱ͖ͷফࣦͱߟ͑ΔͱΘ͔Γ͍͢ ▸ ࢹઢʢޫʣແݶେͱແݶখͱΛԽ͍ͯ͠Δ
ψʔιϩδʔ/0040-0(: ෯ͱԞߦ͖ͷઈରతࠩҟ ▸ ڑۭؒͰӉΛ෴͍ͬͯΔ͕ɺ͜Ε٬؍తͳ෯ͷۭؒ ʢଞऀʣͰ͋ͬͯɺ୭ͦΜͳݟํͰੈքΛݟΔ͜ͱͰ͖ ͳ͍ɻਓؒͷओ؍తͳԞߦ͖ʢࣗݾʣͦ͜ʹͳ͍ ▸ ෯ͱԞߦ͖ɺࣗݾͱଞऀͰߏ͕స͍ͯ͠Δɻࣗݾʹ ͱͬͯͷԞߦ͖ଞऀʹͱͬͯͷ෯ɺଞऀʹͱͬͯͷԞߦ͖ ࣗݾʹͱͬͯͷ෯
▸ ݱࡏͷਓؒͷۭؒೝࣝస͠ɺଞऀଆ͕ઌखʹͳ͍ͬͯΔ
ऀͷࣗ༝
ऀͷࣗ༝5)&"$503`4'3&&%0. ԋܶͷੳ ▸ ΰʔϧυϚϯɺԋܶཧͷ΄ͱΜͲ͕ɺٔۂΛ୯ͳΔจֶ ࡞Ҏ্ͷͷͰͳ͍͔ͷΑ͏ʹੳ͍ͯ͠Δ͜ͱΛ൷ ͢ΔɻͦͷΑ͏ͳੳԋܶʹݻ༗ͷʮऀʯʮ؍٬ʯ ͱ͍͏ཁૉΛແࢹͯ͠͠·͏ ▸ ࠓ·Ͱݟམͱ͞Ε͖ͯͨʮऀʯͱͲ͏͍͏ͷͳͷ͔
ऀͷࣗ༝5)&"$503`4'3&&%0. ੌ͍ྗ ▸ ऀͷٕΛఆٛ͢ΔͷΛੌ͍UFSSJpDྗͱ໊͚Δ ▸ ʮҞܟͷ೦ʯͱ͍͏Α͏ʹɺڪΕͱײಈͱҰମͰ͋Δͱ͍͏ؚҙ ͕͋ΔɻϑϩΠτͷ͍͏ෆؾຯͳͷʢʹೃછΈͷͷʣ྆Ձੑ ͱ͍ͬͨ֓೦ͦͷͷɻѪೖΓࠞͬͨରʹਓऒ͖͚ͭΒΕΔ ▸ ݪ࢝จԽͰࢮऀͷྶʹก͢Δّ͕ࣜීวతʹݟΒΕΔɻͪΖΜ
ԋܶʹଟ͍ɻ؍٬ʹঝͰɺԽ͚ҟੑʹก͢Δ͜ͱ͋Δ͠ɺ Ұਓͷऀ͕ෳͷΛԋ͡Δ͜ͱ͋ΔʢөըʹݟΒΕͳ͍ʣ ▸ ૬͢ΔແݶͷՄೳੑΛҰʹҾ͖ड͚ΔɺΩϟϥΫλʔͱͯ͠ͷ ऀBDUPSBTDIBSBDUFSͷੌ͍ྗͦ͜ɺԋܶΛఆ͚͍ٛͮͯΔ
ݟ͑ͯ͘Δͷ ▸ ෆՄೳੑɻໃ६ٯઆɺܽɻཉܾͯ͠ຬ͞Εͳ͍ ͕ɺ͔ͩΒͦ͜ཉ͢Δ ▸ डಈੑɻଞऀͷ༏Ґੑɺಛݖੑɻଞऀ͔ΒͨΒ͞Εͨࣗ ݾΠϝʔδݴޠΛड͚ೖΕΔ͔͠ͳ͍ ▸ ্هͷೋ͕ͭਓؒଘࡏͷࠜຊతͳੑ࣭ͱͯ͠ɺΘΕΘΕͷΞ ΠσϯςΟςΟࣾձੜ׆Λࢧ͍͑ͯΔ
ΘΕΘΕʹͱͬͯͷࣾձɺ͔ΕΒʹͱͬͯͷࣾձ ▸ ΘΕΘΕਓؒͷӦҝʹ͍ͭͯલड़ͷͱ͓ΓͰ͋Δ ▸ Ͱɺօ͞Μͷ࡞ੈքจԽʹ͓͚Δɺਓؒ૬ͷੜ͖ ͷӦҝʹ͍ͭͯͲ͏ͩΖ͏͔ʁ ▸ ͔ΕΒࣗࣗʹ͍ͭͯԿΛ͍ͬͯͯɺԿΛΒͳ͍ʁ ͔ΕΒଞऀʹ͍ͭͯԿΛ͍ͬͯͯɺԿΛΒͳ͍ʁ ▸
ࣗݾͱଞऀʹؔΘΔ͔ΕΒࣗͷೝࣝʹແணͳ··ɺ ਓশ๏ੑɺଶɺ͋Δ͍ൃߦҝΛߏங͍ͯ͠ͳ͍͔ʁ