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
Online Learning to Rank for Sequential Music Re...
Search
zr_4
October 05, 2019
Research
0
810
Online Learning to Rank for Sequential Music Recommendation
zr_4
October 05, 2019
Tweet
Share
More Decks by zr_4
See All by zr_4
ニュース記事の品質が広告消費行動に与える影響の調査
zr_4
0
820
パーソナライゼーションのためのマルチリービング
zr_4
3
1k
consumption.pdf
zr_4
0
290
Other Decks in Research
See All in Research
LINEチャットボット「全力肯定彼氏くん(LuC4)」の 1年を振り返る
o_ob
0
2.5k
Physics of Language Models: Part 3.1, Knowledge Storage and Extraction
sosk
1
930
[2024.08.30] Gemma-Ko, 오픈 언어모델에 한국어 입히기 @ 머신러닝부트캠프2024
beomi
0
650
KDD論文読み会2024: False Positive in A/B Tests
ryotoitoi
0
170
「Goトレ」のご紹介
smartfukushilab1
0
750
ICLR2024: Reading "Training Unbiased Diffusion Models From Biased Dataset"
hotekagi
0
100
データサイエンティストをめぐる環境の違い 2024年版〈一般ビジネスパーソン調査の国際比較〉
datascientistsociety
PRO
0
450
Weekly AI Agents News! 8月号 プロダクト/ニュースのアーカイブ
masatoto
1
180
Introducing Research Units of Matsuo-Iwasawa Laboratory
matsuolab
0
650
Generative Predictive Model for Autonomous Driving 第61回 コンピュータビジョン勉強会@関東 (後編)
kentosasaki
0
200
機械学習でヒトの行動を変える
hiromu1996
1
210
ミニ四駆AI用制御装置の事例紹介
aks3g
0
160
Featured
See All Featured
The World Runs on Bad Software
bkeepers
PRO
65
11k
"I'm Feeling Lucky" - Building Great Search Experiences for Today's Users (#IAC19)
danielanewman
226
22k
Statistics for Hackers
jakevdp
796
220k
実際に使うSQLの書き方 徹底解説 / pgcon21j-tutorial
soudai
167
49k
Build The Right Thing And Hit Your Dates
maggiecrowley
32
2.4k
Performance Is Good for Brains [We Love Speed 2024]
tammyeverts
3
370
Scaling GitHub
holman
458
140k
How to train your dragon (web standard)
notwaldorf
88
5.7k
ReactJS: Keep Simple. Everything can be a component!
pedronauck
664
120k
jQuery: Nuts, Bolts and Bling
dougneiner
61
7.5k
Bootstrapping a Software Product
garrettdimon
PRO
305
110k
Practical Tips for Bootstrapping Information Extraction Pipelines
honnibal
PRO
9
680
Transcript
0OMJOF-FBSOJOHUP3BOLGPS 4FRVFOUJBM.VTJD3FDPNNFOEBUJPO #-1FSFJSB "6FEB (1FOIB 3-54BOUPT /;JWJBOJ 3FD4ZTจಡΈձࢿྉ ൧௩ᔨೋ![S@
ஶऀհ w ॴଐ w 'FEFSBM6OJWFSTJUZPG.JOBT(FSBJT w ஶऀ w #SVOP-1FSFJSB w
"MCFSUP6FEB w (VTUBWP1FOIB w 3PESZHP-54BOUPT w /JWJP;JWJBOJ
എܠ w ԻָετϦʔϛϯάαʔϏεͷོ w ਓྗϓϨΠϦετͷݶք w Ϣʔβʔͷझଊ͍͑ͯͳ͍ w ͯ͢ͷۂΛΧόʔ͢Δͷແཧ w
աڈͷϩά zݱࡏͷηογϣϯz͔Βֶश͍ͨ͠ w ؾɺΞΫςΟϏςΟʹΑͬͯฉ͖͍ͨۂมΘΔ
ߩݙ w ϢʔβʔϑΟʔυόοΫΛ࿈ଓతʹֶश͢ΔΦϯϥΠϯϥϯ Ϋֶशख๏$PVOUFSGBDUVBM%VFMJOH#BOEJUT $%# ΛఏҊ w *NQMJDJUGFFECBDL QMBZPSBDUJPO ͷΈΛ͏
w ࣍ݩಛϕΫτϧΛείΞϦϯάʹ͏ w Ϟσϧ୳ࡧ࣌ʹҼՌਪ͢Δ w ࠷৽ͷΦϯϥΠϯֶशख๏ΑΓֶश͕ૣ͘ɺߴਫ਼Λୡ
ؔ࿈ݚڀ w "VUPNBUJD1MBZMJTU(FOFSBUJPO w 4UBUJD w .BSLPWSBOEPNpFMEBQQSPBDI 4*(.. 3BHOPFUBM
w %ZOBNJD w JNQMJDJUGFFECBDLΛͬͨώϡʔϦεςΟΫε *4.*3 1BNQBMLFUBM w BVEJPCBTFESFQSFTFOUBUJPOʹجͮ͘ۂΫϥελۭؒͷڧԽ ֶश 3FD4ZT ,JOHBOE*NCSBTBJUF
ؔ࿈ݚڀ w $POUFYUVBMCBOEJUT w /FXTDPOUFYUVBMCBOEJU 84%. -JFUBM w ۂΛBSNʹ͢ΔͱϢʔβʔମݧతʹΑ͘ͳ͍͔
w 0OMJOFMFBSOJOHUPSBOL w ΞΠςϜۭؒͷ࣍ݩಛྔΛBSNʹ͢Δ w ۂҰͭΛධՁ͢ΔͷʹϥϯΩϯάΛධՁ͢ΔJOUFSMFBWJOH ෆద
ઃఆ શۂσʔλ ਪનީิۂ
$%#"MHPSJUIN
$%#"MHPSJUIN VରϢʔβ TϢʔβ͕બΜͩॳظۂ NΞʔϜͷ Tͷϙδγϣϯ ॳظԽ
$%#"MHPSJUIN VରϢʔβ TϢʔβ͕બΜͩॳظۂ NΞʔϜͷ Tͷϙδγϣϯ ॳظԽ ϕΫτϧ୳ࡧ
$%#"MHPSJUIN
$%#"MHPSJUIN VରϢʔβ TϢʔβ͕બΜͩॳظۂ NΞʔϜͷ Tͷϙδγϣϯ ॳظԽ ϕΫτϧ୳ࡧ ධՁ
࣮ݧɿௐ߲ࠪ $%#ͲΕ΄Ͳֶ͘श͢Δ͔ $%#ऩଋ࣌ʹͲΕ΄ͲͷޮՌ͕͋Δ͔ EVFMͲͷΑ͏ʹӨڹ͢Δ͔
࣮ݧɿσʔληοτ w -BTUGN,EBUBTFU w ظؒɿ w ΧϥϜɿVTFS*% BSUJTU*% TPOH*% UJUMF
w 4QPUJGZ .VTJD#SBJO[ͷσʔλͱ݁߹ w VTFST TPOHT FWFOUT
࣮ݧɿධՁ
࣮ݧ݁Ռֶश ˠ-JO6$#ʹൺɺֶ͘शͰ͖͍ͯΔ
࣮ݧ݁ՌֶशޮՌ ˠ-JO6$#ʹൺɺޮՌతʹֶशͰ͖͍ͯΔ
࣮ݧ݁ՌEVFMͷӨڹ ˠNҎ߱Ͱ-JO6$#ʹൺɺใु͕ߴ͔ͬͨ
·ͱΊͱࠓޙͷ՝ w Իָͷਪનʹର͢ΔΦϯϥΠϯֶशख๏$%#ΛఏҊ w ύʔιφϥΠζ͔ͭগͳ͍GFFECBDLͷঢ়گطଘݚڀʹͳ͔ͬͨ w $POUFYUVBMCBOEJUͱͷൺֱ࣮ݧΛߦͬͨ w গͳ͍ϢʔβʔϑΟʔυόοΫͰޮՌతͳֶश w
ࠓޙͷ՝ w χϡʔεਪનͷద༻ w ࣮ੈքͷద༻
ॴײ w 4FRVFOUJBM3FDPNNFOEBUJPOͷΦϑϥΠϯධՁͷΓ ͔͕ͨΓ͔ͨͬͨͷͰࢀߟʹͳͬͨ w Իָਪનઃఆ͕໘ന͍ w ઃఆ͔Β͢Δͱ࣮ݧͷϕʔεϥΠϯଥ͔͠Εͳ ͍͕ɺ$'ͳͲ౷తͳख๏ͱͷਫ਼ͷҧ͍͕ؾʹͳΔ