Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Sign up for free
Menu
Search
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Pricing
Search
Sign in
Sign up for free
東海高校OBが語るマスコミでのデータサイエンティストの仕事 / data scientist ...
Search
Shotaro Ishihara
February 24, 2021
Education
1.9k
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
東海高校OBが語るマスコミでのデータサイエンティストの仕事 / data scientist in news media at satprogram38
中止となった「第38回サタデープログラム」で発表予定だった資料を、一般公開用に編集しました。
https://www.satprogram.net/list.html
Shotaro Ishihara
February 24, 2021
More Decks by Shotaro Ishihara
See All by Shotaro Ishihara
Agent 時代の Kaggle 展望 / kaggle-in-the-agentic-era
upura
1
1.1k
大規模言語モデルは誰を覚えているか / Who Do Large Language Models Memorize?
upura
0
160
[ACL 2026 Demo] Fast-MIA: Efficient and Scalable Membership Inference for LLMs
upura
0
120
Fast-MIA: Efficient and Scalable Membership Inference for LLMs
upura
0
79
JAPAN AI CUP Prediction Tutorial
upura
2
1.4k
情報技術の社会実装に向けた応用と課題:ニュースメディアの事例から / appmech-jsce 2025
upura
0
440
日本語新聞記事を用いた大規模言語モデルの暗記定量化 / LLMC2025
upura
0
800
Quantifying Memorization in Continual Pre-training with Japanese General or Industry-Specific Corpora
upura
1
140
JOAI2025講評 / joai2025-review
upura
0
1.8k
Other Decks in Education
See All in Education
INTRODUCTION TO THE CELL
pawan5505
0
220
Data Management and Analytics Specialisation
signer
PRO
0
2k
DecisionesEjecutivasCrisis
mariooc
0
270
Comentario del plano urbano de Madrid hasta 1860 (1ª parte )
juanmartin2026
1
61k
Fundamentos, Caracteristicas y Aplicaciones de los Modulos NumPy , Matplotlib y Pandas
robintux
0
250
[2026前期火5] 論理学(京都大学文学部 前期 第6回)「かつとまたはの規則」
yatabe
0
510
Center for Entrepreneurship Education | Science Tokyo (Institute of Science Tokyo)
sciencetokyo
PRO
0
370
1人 × AI、1か月でここまで作れる ー 数年前の外注換算3.8〜7.4億円・241〜379人月分の作業を、AI費用 約10万円・31日で
frievea
0
470
[2026前期火5] 論理学(京都大学文学部 前期 第5回)「 ならばの問題演習・proof net・かつの規則」
yatabe
0
420
Antigravityを使ってGeminiAPI(NanobananaPro)と連携して挿絵メーカーを作った
yoshimura_datam
0
210
Leveraging LLMs for student feedback in introductory data science courses
minecr
0
110
Examen de Selectividad. Geografía julio 2026 (Convocatoria Extraordinaria). UCLM
juanmartin2026
1
13k
Featured
See All Featured
Dealing with People You Can't Stand - Big Design 2015
cassininazir
367
27k
Lightning Talk: Beautiful Slides for Beginners
inesmontani
PRO
2
680
B2B Lead Gen: Tactics, Traps & Triumph
marketingsoph
0
230
StorybookのUI Testing Handbookを読んだ
zakiyama
31
6.9k
Lessons Learnt from Crawling 1000+ Websites
charlesmeaden
PRO
1
1.6k
Unsuck your backbone
ammeep
672
58k
Building a A Zero-Code AI SEO Workflow
portentint
PRO
0
710
DevOps and Value Stream Thinking: Enabling flow, efficiency and business value
helenjbeal
1
380
Believing is Seeing
oripsolob
1
210
The Straight Up "How To Draw Better" Workshop
denniskardys
239
140k
Scaling GitHub
holman
464
140k
Paper Plane (Part 1)
katiecoart
PRO
1
11k
Transcript
౦ւߴߍ0#͕ޠΔ ϚείϛͰͷ σʔλαΠΤϯςΟετͷࣄ ੴݪↅଠ ౦ւߴߍճଔʢʣ ୈճαλσʔϓϩάϥϜʢதࢭʣ ˞ൃද༧ఆͩͬͨࢿྉΛɺҰൠެ։༻ʹฤू ݄
ٕज़ֵ৽ͰมΘΔϚείϛ σʔλαΠΤϯςΟετ l"*z͕ͨΒ͢՝ͱઓ z"*z࣌Λੜ͖͍ͯ͘
χϡʔεϝσΟΞ º σʔλαΠΤϯε
໊ݹͰੜ·ΕΔ த৽ฉ ಡച৽ฉ ΛಡΜͰҭͭ ౦ւߴߍʹೖֶ ౦େ ཧᶗ ʹೖֶˍ ެӹࡒஂ๏ਓ౦ژେֶ৽ฉࣾʹೖࣾ
౦ژେֶ৽ฉࣾʹͯ • هऀɺ൛ɺฤू • σδλϧ൛ͷ্ཱͪ͛ • Πϕϯτओ࠵ • ࠂӦۀ
౦େͰֶ෦ʹਐֶ͠ɺσʔλੳ ΛςʔϚʹݚڀ ଔɿίϛϡχςΟͷಛΛߟྀͨ͠ ݟकΓαʔϏεઃܭख๏ͷ։ൃ ˞ֶ෦ͷଔ༏लΛड
⚔χϡʔεϝσΟΞºσʔλαΠΤϯε • ࠃࡍχϡʔεϝσΟΞڠձʹΑΔ ʮੈքͷࡀҎԼਓʯʹબग़ • ੳͷੈքେձͰ༏উ • ʰ,BHHMFελʔτϒοΫʱग़൛
ٕज़ֵ৽ͰมΘΔ χϡʔεϝσΟΞ
ຊͷ৽ฉͷൃߦ෦ͷਪҠ IUUQTXXXQSFTTOFUPSKQEBUBDJSDVMBUJPODJSDVMBUJPOQIQ ੈଳ୯Ґ
എܠʹٕज़ֵ৽ • ΠϯλʔωοτɾεϚϗͷීٴ • 4/4ొʢ୭͕ൃ৴Ͱ͖Δ࣌ʣ • ʮϚείϛʯͷ่յ • ൃ৴ ༰ݕূ
ݸਓ࠷దԽ
ຖ͍ͬͯΔɾαʔϏεʁ✋ • εϚϗɾλϒϨοτɾύιίϯ ͳͲ • 4/4ɾಈը৴ ͳͲ • χϡʔεαΠτ •
ʢࢴͷʣ৽ฉ
ओઓΠϯλʔωοτ • Մॲ࣌ؒͷୣ͍߹͍ • ڝ߹4/4ɾಈը৴ ͳͲ • ৽ฉΠϯλʔωοτਐग़ • ΩʔϫʔυʮσδλϧԽɾࠃࡍԽʯ
৽ฉࣾͷࡏΓํ͕มΘͬͨ • ࢴͷ৽ฉΛ࡞Δਓ • 🆕 ిࢠ൛Λ࡞Δਓ • 🆕 ϢʔβͷԠΛݟΔਓ •
🆕 ϢʔβͷԠΛݟͯվળ͢Δਓ
σʔλαΠΤϯςΟετ ͱԿऀ͔
ࢴͷ৽ฉ͚ͩͩͬͨ࣌
Πϯλʔωοτ͕ීٴͨ࣌͠ 🆕 ిࢠ൛Λ࡞Δਓ 🆕 ϢʔβͷԠΛ ݟΔਓ 🆕 ϢʔβͷԠΛ ݟͯվળ͢Δਓ
ଟ͘ͷۀքͰى͖͍ͯΔྲྀΕ • ϢϏΩλε *OUFSOFUPG5IJOHT %JHJUBM5SBOTGPSNBUJPO • ͋ΒΏΔͷΛܭଌ͠׆༻͢Δ • ڭҭɾεϙʔπɾྲྀ௨ɾ ͳͲ
ʮσʔλʯ͕ՁΛ࢈Ή σʔλʢӳEBUBʣͱɺࣄ࣮ࢿྉΛ ͢͞ݴ༿ɻ ݴޠతʹෳܗͰ͋ΔͨΊɺ ݫີʹෳͷࣄͷू·Γͷ͜ͱ Λࢦ͠ɺ୯ܗ EBUVNʢσʔλϜʣͰ ͋Δɻ ຊޠXJLJQFEJBΑΓ
σʔλαΠΤϯςΟετͱ ʮσʔλ͔ΒՁΛग़͠ɺ Ϗδωε՝ʹ͑Λग़͢ ϓϩϑΣογϣφϧʯ σʔλαΠΤϯςΟετڠձ IUUQXXXEBUBTDJFOUJTUPSKQGJMFTOFXTQEG
खஈ • ͷઐࣝ • ֶɾ౷ܭͷࣝʢ࣌ʹʮ"*ʯʣ • ϓϩάϥϛϯάͷٕज़ ϓϩάϥϚɾ*5ΤϯδχΞͷҰछ
IUUQTICSPSHEBUBTDJFOUJTUUIFTFYJFTUKPCPGUIFTUDFOUVSZ ੈلɺ࠷ηΫγʔͳࣄ
IUUQTCMPHPTDPN BSUJDMF
۩ମతʹ͍ͬͯΔ͜ͱͷൈਮ • Ϣʔβͷߦಈੳ • هࣄاۀͷਪન • ۀͷࣗಈԽɾޮԽ
• σʔλऩूج൫ͷߏங • ར༻ಈͷੳ • ࢪࡦͷʮ"#ςετʯ • ݟग़͠ͷग़͚͠ʢڧԽֶशʣ
ࢪࡦͷʮ"#ςετʯ " ΫϦοΫ ΫϦοΫ
۩ମతʹ͍ͬͯΔ͜ͱͷൈਮ • Ϣʔβͷߦಈੳ • هࣄاۀͷਪન • ۀͷࣗಈԽɾޮԽ
هࣄاۀͷਪન աڈʹಡΜͩهࣄ ৽͍͠هࣄʢຊʣ Λֶश ਪન
จষΛʮϕΫτϧʯʹม աڈʹಡΜͩهࣄ Λֶश (𝒙, 𝒚) = (𝟕, 𝟐)
ϢʔβͷʮΈʯϕΫτϧΛࢉग़ աڈʹಡΜͩهࣄ Λֶश (𝒙, 𝒚) = (𝟕, 𝟐) ✗
ʮΈʯʹ͍ۙهࣄΛਪન աڈʹಡΜͩهࣄ Λֶश (𝒙, 𝒚) = (𝟕, 𝟐) ✗ (𝒙,
𝒚) = (𝟕, −𝟐) (𝒙, 𝒚) = (𝟑, 𝟔) ڑΛܭࢉ
จষΛʮϕΫτϧʯʹม ౦ւߴߍ0#͕ޠΔʂϚείϛʹ͓͚ΔσʔλαΠΤϯςΟ ετͷࣄ 𝑨 = (𝟏, 𝟏, 𝟏) ςϨϏہ͕౦ւߴߍʹऔࡐʹདྷͨΒ͍͠ 𝑨
= (𝟏, 𝟎, 𝟎) σʔλαΠΤϯςΟετʹͳΔͨΊʹ 𝑨 = (𝟎, 𝟎, 𝟏) ໊ݹʹདྷͨΒखӋઌͱϥʔϝϯͩ 𝑨 = (𝟎, 𝟎, 𝟎)
ϧʔϧ࡞Γқ͕ߴ͍ • Ͳͷ୯ޠΛ࠾༻͢Δʁ • Կݸͷ୯ޠΛج४ʹ͢Δʁ • ʮϚείϛʯʮςϨϏہʯҧ͏ʁ • ͲΕ͚͚ͩۙΕਪન͢Δʁ
ػցֶशʢڭࢣ͋Γֶशʣ
ʮܾఆڥքʯΛֶश աڈʹಡΜͩهࣄ Λֶश
ڭࢣ͋Γֶशͷ۩ମྫ • ը૾ೝࣝʢ(PPHMFը૾ݕࡧʣ • Իೝࣝʢ4JSJʣ • ໎ϝʔϧͷࣗಈྨʢ(NBJMʣ
σʔλੳͷੈքେձʢࢲͷ࣮ʣ • ϖοτͷҾ͖औΓ༧ଌʢ,BHHMF Ґʣ • ѱ࣭ͳίϝϯτͷྨʢ,BHHMF Ґʣ • $07*%ͷ3/"ͷ׆ੑ༧ଌʢ,BHHMF Ґʣ
• පͷ༧ଌʢ4*(/"5& Ґʣ • ٿͷདྷ༧ଌʢύɾϦʔά Ґʣ
۩ମతʹ͍ͬͯΔ͜ͱͷൈਮ • Ϣʔβͷߦಈੳ • هࣄاۀͷਪન • ۀͷࣗಈԽɾޮԽ
• هࣄͷࣗಈཁɾࣗಈੜ • จষͷߍӾɾߍਖ਼ • όΠΞεͷݕ
l"*z͕ͨΒ͢ ՝ͱઓ
• ѱҙΛ͍࣋ͬͨํ • ݸਓ࠷దԽͷฐ • "*ͷެฏੑɾղऍੑ
ѱҙΛ࣋ͬͨίϯςϯπͷੜ • ϑΣΠΫχϡʔεͷ֦ࢄ • ੜٕज़ͷߴԽʼݕূ IUUQTZPVUVCFD2(%NF- IUUQTZPVUVCFG+3O&@)N"
("/ɿఢରతੜωοτϫʔΫͱԿ͔ ʙʮڭࢣͳֶ͠शʯʹΑΔը૾ੜ
ِ͔Λఆ͢Δίϯςετ͕։࠵ ༏উۚສԁ IUUQTXXXLBHHMFDPNDEFFQGBLFEFUFDUJPODIBMMFOHF
ϑΟϧλʔόϒϧ ࣗͷΈͷ༰͚͕ͩ৴͞Εͯࢹ͕ڱ͘ͳΔ ΤίʔνΣϯόʔ ࣗͱಉ͡ҙݟ͔Γʹ͢Δ͜ͱͰɺࣗͷҙݟ ͕ઈରతͩͱޡղͯ͠͠·͏
ηϨϯσΟϐςΟ ૉఢͳۮવɻࢴͷ৽ฉʹ٭ޫʁ 6*69Ͱͷ 4NBSU/FXTͷྫ IUUQTUFDIDSVODIDPNTNBSUOFXTMBUFTU OFXTEJTDPWFSZGFBUVSFTIPXTVTFSTBSUJDMFTGSPNBDSPTT UIFQPMJUJDBMTQFDUSVN
"*ͷஅͰਓੜΛࠨӈ͞ΕΔࣄྫ • "*ͳͥͦͷஅΛԼ͔ͨ͠ʁ • அྙཧతɾಓಙతʹଥ͔ • ઃܭࣗମʹͳ͍ͷ͔ʁ
ۙͰٞΛݺΜͩྫ • ࠾༻ʹؔ͢Δ"* • ۚ༥ʹؔ͢Δ"* • ਓछʹؔ͢Δ"*
l"*z࣌Λ ੜ͖͍ͯ͘
ຊͷ·ͱΊ • ٕज़ֵ৽ͰۀքʹมԽ͕ى͖ͨ • σʔλ͔ΒՁΛग़͢Δ࣌ • େͳͷ͍ํ
དྷऀͷϝοηʔδ ʮ"*ʯͷൃల͕͞·͟·ʹͳΔதɺਓؒʹ͔͠Ͱ͖ͳ͍ͷ ʮఆٛʯ͢Δ͜ͱͩͱݸਓతʹࢥ͍ͬͯ·͢ɻٕज़ͰԿͰՄೳʹ ͳͬͨͱ͖ɺԿΛͬͯΑ͍͔ܾΊΔͷ͕େͰ͢ɻʮཧܥɾจܥʯ ͱ͍͏ΈʹनΘΕͳ͍ɺΑΓྖҬԣஅతͳ͕ٞඞཁͰɺதߴੜ ͷօ͞Μʹͥͻ෯͘ઓͯ͠Έͯ΄͍͠Ͱ͢ɻࣗʮχϡʔε ϝσΟΞºσʔλαΠΤϯεʯͷֻ͚߹ΘͤͰࣄΛ͍ͯ͠·͢ɻ