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
MIRU 2019 Lunch on Seminar
Search
Sponsored
·
Your Podcast. Everywhere. Effortlessly.
Share. Educate. Inspire. Entertain. You do you. We'll handle the rest.
→
Hayato Maki
July 31, 2019
Research
300
1
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
MIRU 2019 Lunch on Seminar
Event URL:
https://sites.google.com/zozo.com/miru2019/
Hayato Maki
July 31, 2019
More Decks by Hayato Maki
See All by Hayato Maki
Billion-scale Embedding for E-commerce Recommendation in Alibaba
hamaki
0
120
Today was a Good Day: The Daily Life of Software Developers
hamaki
0
130
論文紹介:Relaxed Softmax for PU Learning
hamaki
3
1.1k
コーディネート整合性を考慮したカテゴリ間推薦
hamaki
0
1.2k
Regularization_The Element of Statical Learning
hamaki
0
220
Neural Activity During Sentence Processing as Reflected in Theta, Alpha, Beta, and Gamma Oscillations
hamaki
0
260
【ICML読み会】Unsupervised Deep Embedding for Clustering Analysis
hamaki
0
1.4k
Other Decks in Research
See All in Research
2026年度 生成AI を活用した論文執筆ガイド/ワークショップ / 2026 Academic Year Guide to Writing Papers Using Generative AI - Workshop
ks91
PRO
0
190
SOTAのさらに先へ:厳しい推論制約下での高性能モデルのPost-Training
analokmaus
0
1.4k
オーストリア流 都市の公共交通サービス水準評価@公共交通オープンデータ最前線2026
trafficbrain
0
210
適応的スパムフィルタのための軽量な類似メッセージカウンタ / jsai2026-adaptive-spam-filter
monochromegane
0
4.4k
Apache Gravitinoで実現する Icebergカタログ統合とアクセスの一元化
matsumooon
0
350
typst の使い方:言語学を研究する学生のために
gitomochang
0
520
The Landscape of Agentic Reinforcement Learning for LLMs: A Survey
shunk031
4
1.1k
Fukui Shibiten 39 - AI Art
butchi
0
150
YOLO26_ Key Architectural Enhancements and Performance Benchmarking for Real-Time Object Detection
satai
3
890
第12回人と環境にやさしい交通をめざす全国大会/熊本都市圏「車1割削減、渋滞半減、公共交通2倍」をめざして
trafficbrain
0
140
NII S. Koyama's Lab Research Overview AY2026
skoyamalab
0
430
SLAMはどこまで解決されたのか?
tomonom
0
840
Featured
See All Featured
A Tale of Four Properties
chriscoyier
163
24k
Designing for Performance
lara
611
70k
The Illustrated Children's Guide to Kubernetes
chrisshort
51
53k
Paper Plane
katiecoart
PRO
2
52k
Building Applications with DynamoDB
mza
96
7.1k
Color Theory Basics | Prateek | Gurzu
gurzu
0
390
How Software Deployment tools have changed in the past 20 years
geshan
0
34k
Pawsitive SEO: Lessons from My Dog (and Many Mistakes) on Thriving as a Consultant in the Age of AI
davidcarrasco
0
190
Into the Great Unknown - MozCon
thekraken
41
2.6k
Visualization
eitanlees
152
17k
Primal Persuasion: How to Engage the Brain for Learning That Lasts
tmiket
0
390
Why Mistakes Are the Best Teachers: Turning Failure into a Pathway for Growth
auna
0
180
Transcript
ϑΝογϣϯΛՊֶ͢ΔऔΓΈ ਅ༐ਓ ϦαʔναΠΤϯςΟετ גࣜձࣾ;0;0ςΫϊϩδʔζ
!2 ࣗݾհ גࣜձࣾ;0;0ςΫϊϩδʔζ ϦαʔναΠΤϯςΟετ ਅ༐ਓ ·͖ɹͱ dݱ৬ /"*45ใֶՊത࢜՝ఔमྃ ਪનγεςϜɺը૾ೝࣝͷݚڀ։ൃʹैࣄ
!3 ;0;0ͷ3%ମ੍ w 3%ͷઐ෦ॺʹਓ͕ॴଐʢ੨ࢁɿਓɺԬɿਓʣ w ݚڀΛ͍ͯ͠Δਓɺ։ൃΛ͍ͯ͠Δਓɺ྆ํ͕͍Δ w ݄ʹൃ ੨ࢁ Ԭ
!4 ͳͥϑΝογϣϯΛݚڀ͢Δͷ͔ w ୭͕ΛબͼɺΛணΔ ਓྨʹͱͬͯීวతͳςʔϚ w ϑΝογϣϯ࢈ۀڊେ ੈքͰஹԁͷࢢن
ήʔϜࢢɿஹԁɺөըࢢɿஹԁ
!5 ֶज़ݚڀʹ͓͚ΔϑΝογϣϯ w ༗ྗࠃࡍձٞͰϑΝογϣϯͷϫʔΫγϣοϓ͕։࠵ *$$7&$$7ɺ$713ɿը૾ೝࣝ ,%%ɿσʔλϚΠχϯά 3FD4ZTɿਪનγεςϜ
w ςʔϚଟ༷Խ͍ͯ͠Δ $713`ɿϑΝογϣϯʹ͓͚Δݕࡧɾಛදݱ &$$7`ɿࣗવݴޠ͔Βͷը૾Λੜ <&$$7`>
ࢲͨͪͷσʔλࢿ࢈
!7 ͷσʔλ w ԯຕҎ্ͷը૾ ΄΅ͯࣗࣾ͢ݿͰࡱӨ ౷Ұ͞ΕͨࡱӨ݅ • આ໌จɺૉࡐใɺֹۚͳͲɺ
ϒϥϯυͷखೖྗΑΔৄࡉͳ Ξϊςʔγϣϯ • ߪങϩά ͷʮങ͍ํʯʹؔ͢Δߴ࣭େྔͳσʔλ ϖʔδͷྫ
!8 ͷσʔλ • ຊ࠷େڃͷϑΝογϣϯίʔ σΟωʔτڞ༗αΠτը૾ • ༻ΞΠςϜͷλά͚ Έ߹Θͤ • ϋογϡλάɺίϝϯτ
ࣗવݴޠ • ϑΥϩʔɾϑΥϩϫʔͷωοτ ϫʔΫάϥϑ ͷʮண͜ͳ͠ʯʹؔ͢ΔϚϧνϞʔμϧͳσʔλ
!9 ͷσʔλ • ຊ࠷େڃͷϑΝογϣϯίʔ σΟωʔτڞ༗αΠτը૾ • ༻ΞΠςϜͷλά͚ Έ߹Θͤ • ϋογϡλάɺίϝϯτ
ࣗવݴޠ • ϑΥϩʔɾϑΥϩϫʔͷωοτ ϫʔΫάϥϑ ͷʮண͜ͳ͠ʯʹؔ͢ΔϚϧνϞʔμϧͳσʔλ
!10 ͷσʔλ • ຊ࠷େڃͷϑΝογϣϯίʔ σΟωʔτڞ༗αΠτը૾ • ༻ΞΠςϜͷλά͚ Έ߹Θͤ • ϋογϡλάɺίϝϯτ
ࣗવݴޠ • ϑΥϩʔɾϑΥϩϫʔͷωοτ ϫʔΫάϥϑ ͷʮண͜ͳ͠ʯʹؔ͢ΔϚϧνϞʔμϧͳσʔλ
!11 ͷσʔλ • ຊ࠷େڃͷϑΝογϣϯίʔ σΟωʔτڞ༗αΠτը૾ • ༻ΞΠςϜͷλά͚ Έ߹Θͤ • ϋογϡλάɺίϝϯτ
ࣗવݴޠ • ϑΥϩʔɾϑΥϩϫʔͷωοτ ϫʔΫάϥϑ ͷʮண͜ͳ͠ʯʹؔ͢ΔϚϧνϞʔμϧͳσʔλ
!12 ͜Ε·Ͱͷݚڀ w ίʔσΟωʔτͷఏҊ<*#*4.-`> ୯ҰͷΞΠςϜͰͳ͘ɺΞΠςϜͷू߹Λਪન͢Δ w ͷ$(දݱ w ܕ͔ΒίʔσΟωʔτΛݕࡧ
࢈ֶ࿈ܞͷ
!14 ࢈ֶ࿈ܞ ڞಉݚڀύʔτφʔ and more… େֶɾͦͷଞݚڀػؔͱͷڞಉݚڀΛਪਐ
!15 ͳͥ࢈ֶ࿈ܞͳͷ͔ w ଟ༷ͳσʔλɺଟ༷ͳधཁ͕͋Δ ը૾ɺࣗવݴޠɺάϥϑɺ࣌ܥྻɺιʔγϟϧλάɺϥϯΩ ϯάɺΞΫηεϩάʜ ݕࡧɾਪનɺࣗಈλά͚ɺधཁ༧ଌɺ$(දݱɺҟৗݕɺ
%Ϗδϣϯ w 3%෦ॺઃཱॳɺओʹը૾ͷݚڀΛ͍ͬͯͨ ଞͷઐՈͱڠྗ͢Δ͜ͱͰɺݚڀΛ͛Δ
!16 ͳͥ࢈ֶ࿈ܞͳͷ͔ w ޮՌతͳใ େֶͷݚڀऀͱ͕ٞͰ͖ΔϨϕϧͷݟɾٕज़ྗ Λ࣋ͭ͜ͱΛࣔ͢ ൃදΛ௨ֶͯ͡ձͷ࿐ग़Λ૿͠ɺ༏लͳֶੜʹ
Ϧʔν͢Δ w தظతͳࢹ࠲ʹཱͬͨݚڀ͕Ͱ͖Δ اۀͰͷݚڀظརӹʹͱΒΘΕɺࢹ͕ڱ͘ͳ Γ͍͢
!17 ڞಉݚڀΛ࣮ݱ͢Δ·Ͱ ͚ࣾͷίϛϡχέʔγϣϯ ͳͥ֎෦ػؔͱڠྗ͢Δͷ͔ ϦεΫʹݟ߹͏ϕωϑΟοτ͕͋Δ͜ͱΛઆ໌ ݸਓใͳͲɺ๏্ͷͷ֬ೝ ւ֎ͷ๏ͰอޢରͱͳΔՄೳੑͷ͋ΔσʔλΛআ֎
େྔͷσʔλΛɺޮΑ͘ɺ҆શʹ͢ (PPHMF#JH2VFSZͰσʔλΛऩूɾલॲཧ$MPVE4USBHFసૹ ఏܞઌͷେֶυϝΠϯʹݶͬͯΞΫηεΛڐՄ
!18 ·ͱΊ ϑΝογϣϯͷɺϏδωεɾݚڀͱʹ Γ্͕͍ͬͯΔɻ ࣾʹཷ·͍ͬͯΔଟ༷ͳσʔλΛ׆͔ͨ͢Ίɺ 3%ͷ෦ॺΛ্ཱͪ͛ͨɻ ݚڀྖҬ͕ଟذʹΔͨΊɺେֶͱͷ࿈ܞΛਪਐ ͍ͯ͠Δɻ