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
ニーズ指向研究の活性化の観点からみたオープンサイエンスの可能性 / Possibility of open science from the viewpoint of revitalizing needs-oriented research
Search
kiyota-yoji
June 19, 2018
Science
2
160
ニーズ指向研究の活性化の観点からみたオープンサイエンスの可能性 / Possibility of open science from the viewpoint of revitalizing needs-oriented research
Japan Open Science Summit (JOSS 2018)
June 19, 2018
at National Center of Science, Tokyo, Japan
kiyota-yoji
June 19, 2018
Tweet
Share
More Decks by kiyota-yoji
See All by kiyota-yoji
JSAI 2023企画セッション「AI哲学マップ」企画の振り返り / JSAI 2023 Reflections on the AI Philosophy Map lecture series project
ykiyota
0
140
住まい探しの利便性向上にデータベース・情報アクセス技術が 果たした役割 / The Role of Database and Information Access Technology in Improving the Convenience of Housing Search
ykiyota
0
90
デジタル社会の行き着く先にライブラリアンが果たしうる役割を考える / Thinking about the role librarians can play in the destination of the digital society
ykiyota
0
1.6k
LIFULLアジェンダ -社会課題の発見と解決に向けた研究開発活動の紹介- / LIFULL Agenda -Research and development activities to discover and solve social issues-
ykiyota
0
210
JSAI 2022チュートリアル講演 AI哲学マップ / JSAI 2022 Tutorial "AI Philosophy Map"
ykiyota
1
750
不動産コンテンツ研究における異分野研究者との協働の取り組み / Collaborative efforts with researchers from different fields for real estate content studies
ykiyota
1
84
AI研究コミュニティのダイバーシティ&インクルージョン / Diversity and Inclusion in AI Research Communities
ykiyota
0
180
LIFULL HOME’Sデータセット提供がもたらす不動産業界のイノベーション / Innovation in the real estate industry brought about by the provision of the LIFULL HOME'S data set
ykiyota
0
110
LIFULL Poster of IDR User Forum 2020
ykiyota
0
93
Other Decks in Science
See All in Science
(neuro)science with AI: Machine learning as scientific modeling
gaelvaroquaux
0
680
Direct Preference Optimization
zchenry
0
110
5色定理
techmathproject
0
190
早わかり W3C Community Group
takanorip
0
250
SHAPで因果関係を説明できる?
s1ok69oo
7
4k
Leveraging conformal prediction for calibrated probabilistic time series forecasts to accelerate the renewable energy transition
ingevandenende
2
230
2023-10-03-FOGBoston
lcolladotor
0
150
Machine Learning for Materials (Lecture 5)
aronwalsh
0
530
Yasuke
drawsbygba
0
600
Spark_Task_Optimization_Journey_How_I_Increased_10x_Speed_by_Performance_Tuning
tlyu0419
0
190
HIBINO Aiko
genomethica
0
320
20240127_OpenRadiossエアバッグ解析
kamakiri1225
0
130
Featured
See All Featured
How to train your dragon (web standard)
notwaldorf
71
5.1k
A designer walks into a library…
pauljervisheath
199
23k
Principles of Awesome APIs and How to Build Them.
keavy
119
16k
What’s in a name? Adding method to the madness
productmarketing
PRO
14
2.6k
Making the Leap to Tech Lead
cromwellryan
122
8.4k
Refactoring Trust on Your Teams (GOTO; Chicago 2020)
rmw
24
2.2k
Templates, Plugins, & Blocks: Oh My! Creating the theme that thinks of everything
marktimemedia
18
1.7k
The Illustrated Children's Guide to Kubernetes
chrisshort
28
46k
The Pragmatic Product Professional
lauravandoore
24
5.7k
Rails Girls Zürich Keynote
gr2m
91
13k
GraphQLとの向き合い方2022年版
quramy
28
12k
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
28
5.9k
Transcript
χʔζࢦݚڀͷ׆ੑԽͷ؍͔ΒΈͨ ΦʔϓϯαΠΤϯεͷՄೳੑ L I F UL L L a bɹओ੮ݚڀһ
ਗ਼ాɹཅ࢘ 2018.06.19 JOSS 2018 ηογϣϯB4 ʮຽؒاۀσʔλʹΑΔΦʔϓϯαΠΤϯεͷՄೳੑʯ Copyright© LIFULL All Rights Reserved.
ਗ਼ా ཅ࢘ LIFULL Lab ओ੮ݚڀһ Ԭݝੜ·Ε→େֶ(Ӄ)ˏژ→౦ژ ؔ৺: ࣗવݴޠॲཧԠ༻ → ݕࡧɾਪન
→ ใϦςϥγʔ (ਤॻؗ) → ੜ׆ྖҬ (ෆಈ࢈ɺհޢ etc.) ͰͷϝσΟΞٕज़׆༻ (ը૾ղੳؚΉ) ܦྺ: େֶڭһ → ݉ۀͰىۀ → ͦͪΒ͕ຊۀʹ → ങऩ ର֎త׆ಈ • ਓೳֶձ ฤूҕһձ ෭ҕһ (2018-) • ใॲཧֶձσʔλϕʔεγεςϜݚڀձ װࣄ (2017-) • WebDB Forum ࢈ֶ࿈ܞ୲װࣄ (2015-) • Code4Lib JAPANڞಉද (2011-) etc. 2
૯ܝࡌ݅No.1ͷ ෆಈ࢈ɾॅใαΠτ Β͠ʹີணͨ͠͞·͟·ͳ ใαʔϏε (Ҿӽ, อݥ, հޢ, ࢠҭͯ, etc.) ੈք50Χࠃ͚ʹల։͢Δ
ॅɾதݹंɾٻ৬ͳͲͷ ΞάϦήʔγϣϯαΠτ (本社: スペイン バルセロナ) LIFULLάϧʔϓͷαʔϏε܈
None
ग़య: Manyika, James et al. Digital America: A tale of
the haves and have-mores. McKinsey Global Institute. 2015 http://www.mckinsey.com/industries/high-tech/our-insights/digital-america-a-tale-of-the-haves-and-have-mores ෆಈ࢈ςοΫͷҐஔ͚ͮ ʢMcKinseyͷϨϙʔτʣ • ଞͷʮXςοΫʯͱൺֱ ͢ΔͱσδλϧԽͷ߹ ͍தؒతͳҐஔ͚ͮ • λʔήςΟϯάࠂͳͲɺ ͢Ͱʹߴʹσʔλ׆༻ • ҰํͰɺ٬ݟͳͲ ਓʹཔΔ෦͕େ͖͍ • ࠓޙͷσδλϧԽਁಁʹ ͱͳ͏มֵ͕ظ͞Ε Δ
Agenda • LIFULL HOME’SσʔληοτఏڙͷऔΓΈ • ຊͷݚڀίϛϡχςΟͷ՝ ʙγʔζࢦͱ χʔζࢦʙ • ຽؒاۀ͔ΒΈͨΦʔϓϯαΠΤϯε࣮ફͷ՝
ͱҙٛ • ͓ΘΓʹ 6
LIFULL HOME’Sσʔληοτ ఏڙͷऔΓΈ 7
ࠃͷିෆಈ࢈݅σʔλ 530ສ݅ • ॴࡏ (༣ศ൪߸ɺ࠷دΓӺͳͲ) • ྉɺ໘ੵɺஙɺ෦λΠϓ etc. • ݐߏ
(ɺమࠎɺమےίϯΫϦʔτ etc.) • ֤छͩ͜ΘΓ݅ (ϖοτՄɺָثɺΧ ϯλʔΩονϯɺόεɾτΠϨผ etc.) 物件画像 約8300万点 間取り図 約510万点 重厚な感じの エントランス 日当たりの 良いリビング • 2015年11月より提供開始 • 国内外の50を超える研究 組織への提供実績 LIFULL HOME’Sσʔληοτ 8
ෆಈ࢈݅ը૾ɾؒऔΓਤσʔλΛ ར༻ͨ͠ݚڀʹΑΔΠϊϕʔγϣϯग़ 9 不動産会社が⼊稿する画像の不整合検出 ユーザーに提供する 不動産情報品質の向上 間取り図からの3Dモデル⽣成 古川康隆准教授(Simon Fraser Univ.)による
LIFULL HOMEʼSデータセット利⽤研究 ICCV 2017に採択 VRコンテンツなどの 新たなUXの提供
σʔλαΠΤϯεΞϫʔυ2017 ϑΝΠφϦετબग़ 10
ਓೳֶձࢽ্Ͱͷ ಛूاըʮෆಈ࢈ͱAIʯ 20177݄߸ʹܝࡌɺهࣄ11ຊɾ61ϖʔδ 11
ෆಈ࢈ςοΫݚڀίϛϡχςΟ ͷग़ͱ׆ੑԽ ਓೳֶձશࠃେձOS ʮෆಈ࢈ͱAIʯ(2017, 2018) ICMR 2018ซઃࠃࡍϫʔΫγϣοϓ Multimedia for Real
Estate Tech (2018) 12
ຊͷݚڀίϛϡχςΟͷ՝ ʙγʔζࢦͱχʔζࢦʙ 13
ຊͷݚڀίϛϡχςΟͷ՝ • ࢈ֶؒͷ૬ޓཧղෆʁ • ϦεΫςΠΫͷෆʁ • γʔζࢦͷۃͳภΓʁ 14
ຊͷݚڀίϛϡχςΟͷ՝ • ࢈ֶؒͷ૬ޓཧղෆʁ • ϦεΫςΠΫͷෆʁ • γʔζࢦͷۃͳภΓʁ 15
ҏ౻و೭ઌੜˏ͓ͷਫঁࢠେ Visual Analyticsͱ͍͏ݚڀʹ͍ͭͯ ...ҰछͷχʔζࢦͰੜ·Εֶͨज़ίϛϡχςΟͱ ͍͑Δɻͱ͜Ζ͕ஶऀͷܦݧͱͯ͠ɺ͜ͷֶज़ ΛຊͰհ͢Δͱ͕טΈ߹Θͳ͍͜ͱ͕͋Δɻ ݚڀશମͷϑϨʔϜϫʔΫ࣮༻ࣄྫʹؔ৺Λ ࣋ͨͣɺػցֶशΠϯλϥΫγϣϯͳͲͷ֤ཁૉ ٕज़ͷཧత৽نੑ͚ͩΛࣥ፠ʹ࣭ͯ͘͠Δਓ͕ গͳ͔Βͣݟड͚ΒΕΔͷͰ͋Δɻʢதུʣཧ
ج൫ٕज़ͷ৽نੑ͔Βग़ൃͨ͠ϘτϜΞοϓతͳ γʔζࢦʹͩ͜Θͬͨڱ͍ํ๏͕ɺஶऀͷपғ ʹ͓͚Δຊͷֶज़քͰࠜڧ͗͢ΔͷͰɺͱ͍ ͏ײΛͨ͟ΔΛಘͳ͍ɻ ҏ౻و೭. 2015. ΠϊϕʔγϣϯͷͨΊͷ࢈ֶ࿈ܞͱجૅڭҭʹؔ͢ΔҰߟ. ಛूʮΠϊϕʔγϣϯͱAIݚڀʯ. ਓೳ, Vol. 30, No. 3, pp. 337-343. 16
χʔζࢦݚڀͷ׆ੑԽʹ ͚ͯ • ʮχʔζͷϑΥʔΧεͰ͔͠ൃݟ͞Εͳ͍ຊ࣭ తͳݚڀ՝ͨ͘͞Μ͋Δʯͱ͍͏ೝ͕ࣝί ϛϡχςΟશମͰڞ༗͞ΕΔ͜ͱ͕େ • େֶʹ͓͚Δجૅڭҭͷॏཁੑͷ࠶ೝࣝ • ઢܗɺ౷ܭֶɺΞϧΰϦζϜɺਓจࣾձՊֶɺetc.
ͷਂ͍ཧղ͕ɺχʔζࢦݚڀͷԼࢧ͑ͱͳΔ • ݚڀۀͱಉ༷ʹڭҭۀߴ͘ධՁ͞ΕΔʹʁ • ࢈ֶ࿈ܞͷதͷॆ࣮ • ͱ͘ʹɺ࢈ֶͷ૬ޓཧղΛਐΊΔࢪࡦʢਓࡐަྲྀɾΠ ϯλʔϯγοϓͳͲʣ 17
ຽؒاۀ͔ΒΈͨ ΦʔϓϯαΠΤϯε࣮ફͷ ՝ͱҙٛ 18
ຽؒاۀʹͱͬͯͷ ΦʔϓϯαΠΤϯε࣮ફͷ՝ • εςʔΫϗϧμʔ͔Βͷཧղ • ސ٬ɺܦӦਞɺΤϯυϢʔβʔɺࣾձɺetc. • EUҰൠσʔλอޢنଇʢGDPRʣͷࢪߦʹΑΓɺΤϯυ Ϣʔβʔ͔ΒͷཧղΑΓॏཁʹ •
σʔλఏڙͷίετ • ఏڙͷޮՌίετΛਖ਼Խ͠͏Δ͔ʁ • ϦεΫͷίϯτϩʔϧ • ͱ͘ʹɺڝ߹اۀΛར͢ΔϦεΫʹ৻ॏʹͳΔ 19
ͦΕͰຽؒاۀ͕ ΦʔϓϯαΠΤϯεΛ࣮ફ͢Δҙٛ • ಉҰͷݚڀ՝Λڞ༗͢ΔίϛϡχςΟͷܗ • େֶ͕ͦͷݚڀ՝Λબ͢Δಈػ͚ͮͱͳΔ • ʮاۀ͔Βͷݸผ૬ஊͷରԠ͏ݶք…ʯ • ڝ߹Ͳ͏͠Ͱ͋ͬͯڠྗͰ͖Δ෦ଟ͍
• ݚڀ։ൃඅͷ༗ޮ׆༻ʹͭͳ͕Δ • χʔζࢦݚڀΛ૿͍ͨ͠ͷͰ͋Εɺاۀଆ ͔ΒͷΑΓੵۃతͳಇ͖͔͚͕ඞཁ 20
͓ΘΓʹ 21
͓ΘΓʹ • ʮΦʔϓϯʯͱʮΫϩʔζυʯͷؒʹάϨʔ κʔϯ͕ଘࡏ • اۀʹͱͬͯͷσʔλఏڙͷऔΓΈʹεςʔΫϗ ϧμʔͷઆಘ͕ෆՄܽ • ݱ࣮తͳσʔλެ։Ϩϕϧͷઃఆ͕ඞཁ •
ʮΦʔϓϯʯͷՌ࣮ΛಘΔͨΊͷઓུ • ڭओٛʹؕΒͳ͍Α͏ʹ 22