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
KaggleとRコンペの紹介 / Kaggle R Competition
Search
Shotaro Ishihara
December 06, 2019
Technology
1.7k
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
KaggleとRコンペの紹介 / Kaggle R Competition
Japan.R 2019での発表資料
https://japanr.connpass.com/event/154070/
Shotaro Ishihara
December 06, 2019
More Decks by Shotaro Ishihara
See All by Shotaro Ishihara
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
88
[ACL 2026 Demo] Fast-MIA: Efficient and Scalable Membership Inference for LLMs
upura
0
88
Fast-MIA: Efficient and Scalable Membership Inference for LLMs
upura
0
51
JAPAN AI CUP Prediction Tutorial
upura
2
1.2k
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
410
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
760
Quantifying Memorization in Continual Pre-training with Japanese General or Industry-Specific Corpora
upura
1
130
JOAI2025講評 / joai2025-review
upura
0
1.7k
AI エージェントを活用した研究再現性の自動定量評価 / scisci2025
upura
1
270
Other Decks in Technology
See All in Technology
OPENLOGI Company Profile for engineer
hr01
1
74k
企業でAWS Organizationsを動かすための組織設計の考え方
nrinetcom
PRO
1
110
変更し続けられるシステムをどう保つか — AI時代のSSoTという設計原則
kawauso
1
260
Multicaで30個のミニプロジェクトをAIエージェント運用して見えてきたこと
eiei114
0
290
公式ドキュメントの歩き方etc
coco_se
1
120
Oracle Base Database Service 技術詳細
oracle4engineer
PRO
15
110k
kaonavi Tech Night#1
kaonavi
0
120
SREとQA 二人三脚で進めるSLO運用/sre-qa-slo
sugitak
0
1k
インシデント事例と パッケージの全量解析に学ぶ ソフトウェアサプライチェーンの守り方 / supply-chain-attack-defense
flatt_security
0
140
Data + AI Summit 2026 イベントレポート: 「AIがビジネスで意思決定するデータ基盤」へ
nek0128
0
290
誤解だらけの開発生産性 / Myths and Misconceptions about Developer Productivity
i35_267
2
830
脱金融のフューチャー・デザイン / Future Design Beyond Finance
ks91
PRO
0
160
Featured
See All Featured
Paper Plane (Part 1)
katiecoart
PRO
1
9.7k
Technical Leadership for Architectural Decision Making
baasie
3
440
4 Signs Your Business is Dying
shpigford
187
22k
How to train your dragon (web standard)
notwaldorf
97
6.7k
Designing for humans not robots
tammielis
254
26k
Why Your Marketing Sucks and What You Can Do About It - Sophie Logan
marketingsoph
0
280
Designing for Timeless Needs
cassininazir
1
370
Max Prin - Stacking Signals: How International SEO Comes Together (And Falls Apart)
techseoconnect
PRO
0
200
Large-scale JavaScript Application Architecture
addyosmani
515
110k
Intergalactic Javascript Robots from Outer Space
tanoku
273
27k
Understanding Cognitive Biases in Performance Measurement
bluesmoon
32
3k
The B2B funnel & how to create a winning content strategy
katarinadahlin
PRO
1
420
Transcript
,BHHMFͱ3ίϯϖͷհ V !VQVSB +BQBO3 %FDUI
ຊͷ • ػցֶशίϯϖʮ,BHHMFʯͷհ • ʹऴྃͨ͠ʮ3ίϯϖʯͷհ • 3Ϣʔβʹ,BHHMFΛ͓קΊ͢Δཧ༝
ࣗݾհʛV • ࣄۀձࣾͰσʔλαΠΤϯςΟετ • ,BHHMF.BTUFS • ࠷ߴϥϯΫҐɺҐೖܦݧ͋Γ • ࡢͷ+BQBO3ܖػͰൃͨ͠
TQPBOB ӡӦ • ࢀՃͯ͠ͳ͔ͬͨʢཪͷʮ,BHHMF 5PLZP .FFUVQʯͰൃදͯ͠·ͨ͠ʣ
ຊͷ • ػցֶशίϯϖʮ,BHHMFʯͷհ • ʹऴྃͨ͠ʮ3ίϯϖʯͷհ • 3Ϣʔβʹ,BHHMFΛ͓קΊ͢Δཧ༝
,BHHMFͱ • ओʹڭࢣ͋Γֶश͕ରͷػցֶशϞσϧͷ ੑೳΛڝ͏ίϯϖςΟγϣϯ • (PPHMFࡿԼͷ,BHHMF͕ओ࠵ • ੈքதͷσʔλαΠΤϯςΟετ͕ू͏
ڭࢣ͋Γֶशͷ֓ཁ
(PPHMFτϨϯυ • ۙɺຊͰਓؾ͕ߴ·͍ͬͯΔ
ຊͷ • ػցֶशίϯϖʮ,BHHMFʯͷհ • ʹऴྃͨ͠ʮ3ίϯϖʯͷհ • 3Ϣʔβʹ,BHHMFΛ͓קΊ͢Δཧ༝
34UVEJP͕ࣾίϯϖ։࠵ • ݄։࢝ɺ݄ऴྃ • ্ҐνʔϜˍ্Ґ3ϢʔβνʔϜʹۚ • ʮ(PPHMF.FSDIBOEJTF4UPSFʯͷސ٬͝ͱ ͷऩӹΛ༧ଌ
σʔλˍධՁࢦඪ • (PPHMF"OBMZUJDTͰऔಘͨ͠ใͷҰ෦ • Ϣʔβ*% • ηογϣϯ*% • ηογϣϯ
• σόΠεใ ͳͲͳͲ • ධՁࢦඪɿ
σʔλ֓ཁ • ࣌ܥྻͰͷσʔλׂ • ࣮Ұண͋ͬͨʢ("ͷσϞΞΧϯτͰ ͕͑ೖखͰ͖ͨʢAʣŇP0ʁʁʁʁʁʣ IUUQTXXXTMJEFTIBSFOFU4IVIFJ(PEBLBHHMFUPLZPNFFUVQ S
Ґͷղ๏ • ୯७ͳಛྔΤϯδχΞϦϯά • Ϣʔβ୯Ґͷ౷ܭྔɺ࣌ܥྻใͳͲ • ճؼͰͳ͘ɺྨˍճؼͰղ͘ • ʮऩӹΛಘΔ֬ͷ༧ଌʯ✕ʮऩӹͷ༧ଌʯ
• ҐˠҐʹ • ίʔυެ։ʢ3Ͱߦʣ IUUQTXXXLBHHMFDPNDHBDVTUPNFSSFWFOVFQSFEJDUJPOEJTDVTTJPO
ຊͷ • ػցֶशίϯϖʮ,BHHMFʯͷհ • ʹऴྃͨ͠ʮ3ίϯϖʯͷհ • 3Ϣʔβʹ,BHHMFΛ͓קΊ͢Δཧ༝
࠷ۙͷςʔϒϧίϯϖ • σʔλΛΑ͘ݟΔॏཁੑ͕૿͍ͯ͠Δ • ཧ༝ɿ • σʔλαΠΤϯεͷύοέʔδٕज़ͳͲ͕ঃʑʹ ීٴ͖ͯͯ͠ɺҰൠతͳճؼɾྨࣄۀձࣾ Ͱղ͚ͯ͠·͏
• ݁Ռͱͯ͠ɺ,BHHMFʹग़͞ΕΔςʔϒϧίϯϖ σʔληοτ՝ઃఆʹҰบ͋Δ͕૿Ճ
ίϯϖʢʙ݄ʣ • ͕͍ͭى͖Δ͔Λ༧ଌ͢Δ • ఆείΞ༻࠷ऴείΞ༻ͷσʔλ͕େ͖͘ ဃ͍ͯͨ͠ • ࢲͷνʔϜҐ͔ΒҐʹసམ IUUQTXXXLBHHMFDPND-"/-&BSUIRVBLF1SFEJDUJPO
,BHHMFY3 • 3ݴޠͷརͷҰͭσʔλॲཧɾՄࢹԽ • σʔλΛஸೡ͔ͭਝʹோΊΒΕΔਓڧ͍ • ˞ ػցֶशϞσϦϯάͷ߹Ͱɺ1ZUIPO͕ ඞཁʹͳΔ໘͋Δ͔
,BHHMFY1ZUIPO ˞ ʹߨஊ͔ࣾΒॻ੶Խ༧ఆ IUUQTRJJUBDPNVQVSBJUFNTDGGGFEFDEG
·ͱΊ • ػցֶशίϯϖʮ,BHHMFʯͷհ • ʹऴྃͨ͠ʮ3ίϯϖʯͷհ • 3Ϣʔβʹ,BHHMFΛ͓קΊ͢Δཧ༝