オンライン広告関連の論文を50本くらい雑に紹介する AdKDD編 / adkdd-all

Slide 1

Slide 1 text

ΦϯϥΠϯ޿ࠂؔ࿈ͷ࿦จΛ 50ຊ͘Β͍ࡶʹ঺հ͢Δ AdKDDฤ @todesking

Slide 2

Slide 2 text

ࣗݾ঺հ • @todesking • DSP/ΞυωοτϫʔΫͷձࣾ ͰػցֶशΛ΍͍ͬͯΔ • ೔ࠒͷٕज़ௐ͕ࠪ଍Γͳ͍ͳ ͱࢥͬͨͷͰAdKDDͷ࿦จΛ ோΊͯΈ·ͨ͠

Slide 3

Slide 3 text

AdKDD/TargetADͱ͸ • KDD: Knowledge Discovery and Data Mining • σʔλϚΠχϯά΍ػցֶशʹؔ͢ΔԠ༻తͳ࿩୊Λѻ͏ࠃࡍ ձٞ • AdKDD: ΦϯϥΠϯ޿ࠂؔ࿈ٕज़Λѻ͏෼Պձ • TargetAd: WSDMͱ͍͏ࠃࡍձٞʹ͓͍ͯΦϯϥΠϯ޿ࠂؔ࿈ٕज़ Λѻ͏෼Պձ • ࠷ۙ͸AdKDDʹٵऩ͞Εͨ? • ࠾୒࿦จ਺͸ຖ೥10ຊऑ • KDDຊମʹ΋޿ࠂܥ࿦จ͕͋Δ(੗Έ෼͚͸Ṗ)

Slide 4

Slide 4 text

஫ҙ • ཁ໿͸΄΅AbstractோΊ͚ͨͩɺਖ਼֬͞͸อূ͠·ͤΜ • ؒҧ͍ͷࢦఠ͸׻ܴ • ۙ೥ͷ࿦จ͸ެࣜαΠτ͔ΒPDFͷ௚ϦϯΫ͕͋Δ • Keynote/Invited talk͸লུ

Slide 5

Slide 5 text

AdKDD 2014 https://dblp.org/db/conf/kdd/adkdd2014

Slide 6

Slide 6 text

A Dynamic Pricing Model for Unifying Programmatic Guarantee and Real-Time Bidding in Display Advertising • publisher͕impΛചΔʹ͸RTBͱguranteed contractͱ͍ ͏ೋͭͷํ๏͕͋Δ • publisherͷऩӹΛදݱ͢Δ਺ཧϞσϧΛ࡞Γɺೋͭͷํ ๏΁ͷ༧ࢉ഑෼Λ࠷దԽ͢Δ

Slide 7

Slide 7 text

Multi-Touch Attribution Based Budget Allocation in Online Advertising • Ωϟϯϖʔϯશମͷ༧ࢉ͕ܾ·͓ͬͯΓɺԼҐͷΩϟϯ ϖʔϯʹ഑෼͍ͨ͠ • ໨ඪ΁ͷߩݙʹԠͯ͡഑෼͍ͨ͠ˠlast-touch/multi- touch attributionͷ৘ใΛݩʹׂͯ͠Γ౰ͯΔํ๏Λ࣮૷ ͨ͠

Slide 8

Slide 8 text

Pleasing the advertising oracle: Probabilistic prediction from sampled, aggregated ground truth • ޿ࠂΩϟϯϖʔϯͷηάϝϯτਫ਼౓Λୈࡾऀ͕ධՁ͢Δ έʔε • ධՁͱͯ͠ɺ޿ࠂΛݟͨϢʔβͷσϞάϥଐੑͷׂ߹͕ ϑΟʔυόοΫ͞ΕΔ • ݸผϢʔβͷଐੑʹ͍ͭͯͷ৘ใ͸ಘΒΕͳ͍ • ͜ͷ৘ใΛ࢖ͬͯηάϝϯτͷਫ਼౓Λվળ͢Δ

Slide 9

Slide 9 text

The PlaceIQ Analytic Platform: Location Oriented Approaches to Mobile Audiences • PlaceIQͱ͍͏Ґஔϕʔεͷ޿ࠂ෼ੳπʔϧͷ঺հ

Slide 10

Slide 10 text

Practical Lessons from Predicting Clicks on Ads at Facebook • ஶ໊ͳΦϯϥΠϯ޿ࠂ࿦จ • FBͷCTR༧ଌϞσϧͷ࣮૷ʹ͍ͭͯ

Slide 11

Slide 11 text

iPinYou Global RTB Bidding Algorithm Competition Dataset • iPinYouͱ͍͏DSP͕ओ࠵ͨ͠RTBΞϧΰϦζϜͷίϯς ετͰ࢖༻͞Εͨσʔληοτʹ͍ͭͯ • ͜ͷσʔληοτ͸RTBܥϩδοΫͷ࣮ݧʹΑ͘࢖ΘΕͯ ͍Δ

Slide 12

Slide 12 text

Can Television Advertising Impact Be Measured on the Web? Web Spike Response as a Possible Conversion Tracking System for Television. • ςϨϏࢹௌऀ͸ಉ࣌ʹΠϯλʔωοτ΋΍ΔͷͰɺςϨϏ CMͷࢹௌ͕े਺ඵޙͷωοτ׆ಈʹӨڹΛ༩͑Δ • ͜ͷΑ͏ͳӨڹΛܭଌ͢ΔͨΊͷwebϕʔεγεςϜΛ։ ൃͨ͠

Slide 13

Slide 13 text

TargetAd 2016 https://sites.google.com/site/targetad2016/program

Slide 14

Slide 14 text

TargetAd 2016 https://sites.google.com/site/targetad2016/program

Slide 15

Slide 15 text

TargetAd 2016 https://sites.google.com/site/targetad2016/program

Slide 16

Slide 16 text

Distributed Representations of Web Browsing Sequences for Ad Targeting • Ϣʔβͷϒϥ΢δϯάཤྺΛಛ௃ྔԽ͍ͨ͠ • ϢʔβΛparagraphɺURLΛwordʹݟཱͯͯParagraph Vector(W2Vͷ೿ੜɺύϥάϥϑͷembeddingΛಘΒΕΔ) Λߏங͢Δͱ͍͏ͷΛલճ΍ͬͨ • Backward PV-DMͱ͍͏վྑ൛ͷఏҊ

Slide 17

Slide 17 text

Sponsored Video Advertising: Filling the Void between Online and TV Advertising • Ϣʔβ͕ಈըΛࢹௌ͍ͯ͠Δͱ͖ʹ޿ࠂΛݟ͍ͤͨɺݟͤ ·͘Γ͍ͨͱ͍͏ڧ͍ؾ͕࣋ͪ͋Δ • ಈըʹؔ࿈͢Δ޿ࠂͷ՝ۚମܥʹ͍ͭͯͷఏҊ

Slide 18

Slide 18 text

On Spatial-Aware Viral Marketing for Location-based Advertisements, • ৽͍͠ళΛ։͖͍ͨ৔߹ɺళͷ৔ॴɺޱίϛͷޮՌɺ෺ ཧ޿ࠂͷ৔ॴͳͲͷཁૉ͕ച্͛ʹӨڹΛ༩͑Δɻ • ͦͷΑ͏ͳ੍໿ͷԼͰ࠷దͳީิΛબͿํ๏ΛఏҊ

Slide 19

Slide 19 text

Keynote: Come back soon! Estimating return times for users • ߦಈཤྺΛݩʹɺϢʔβ͕࣍ʹ͜ͷαʔϏεΛ࢖͏ͷ͸ ͍͔ͭਪఆ͍ͨ͠ • ੜଘ෼ੳɺ఺աఔɺLSTM

Slide 20

Slide 20 text

Building a Bi-directed Recommendation System for Mobile Users and App-install Ad Campaigns • ϞόΠϧΞϓϦʹؔ͢Δ૒ํ޲Ϩίϝϯσʔγϣϯγες Ϝ • ΞϓϦͷ޿ࠂΩϟϯϖʔϯʹରͯ͠ɺޮՌతͳϢʔβηά ϝϯτΛਪન • Ϣʔβʹରͯ͠ɺڵຯΛ࣋ͪͦ͏ͳΞϓϦΛਪન

Slide 21

Slide 21 text

Finding Needle in a Million Metrics: Anomaly Detection in a Large-scale Computational Advertising Platform • ޿ࠂϓϥοτϑΥʔϜʹ͓͚Δҟৗݕ஌ • Ξυϑϥ΢υͷจ຺Ͱ͸ͳ͘ɺγεςϜͷҟৗΛݕग़͢Δ ͷ͕໨త • ҆ఆͨ͠ύϑΥʔϚϯεΛग़͍ͯ͠ΔΩϟϯϖʔϯΛ؍ଌ • ύϑΥʔϚϯεʹҟৗ͕͋ͬͨΒɺγεςϜଆʹݪҼ͕ ͋ΔՄೳੑ͕ߴ͍

Slide 22

Slide 22 text

Preserving Privacy in Geo- Targeted Advertising • ஍ҬͱऩೖΛ࢖ͬͨλʔήςΟϯά޿ࠂΛߟ͑Δ • ߈ܸऀ͸τϐοΫϞσϦϯά΍ػցֶशΛ࢖͏͜ͱͰɺ͋ Δ޿ࠂ͕දࣔ͞ΕͨϢʔβͷ஍Ҭ΍ऩೖΛਪఆͰ͖Δ • ഑৴͞ΕΔ޿ࠂʹϊΠζΛࠞͥΔ͜ͱͰ͜ͷΑ͏ͳਪఆΛ ๦͛Δ

Slide 23

Slide 23 text

Invited talk: User Recommendation in Instagram • Instagramʹ͓͚Δ͓͢͢ΊϢʔβͷ࢓૊Έ • ιʔγϟϧάϥϑͷߏ଄΍ڵຯͷ͋ΔτϐοΫΛ࢖ͬͨਪ ન

Slide 24

Slide 24 text

Demographic Prediction of Web Requests from Labeled Aggregate Data • Ϣʔβू߹ʹରͯ͠ɺσϞάϥଐੑͷׂ߹Λ஌Δ͜ͱ͕Ͱ ͖Δͱ͢Δ • ͜ͷ৘ใΛ࢖ͬͯϢʔβ୯ҐͰͷଐੑਪఆΛ͢Δ • Williams+, "Pleasing the advertising oracle: Probabilistic prediction from sampled, aggregated ground truth", AdKDD2014 ͷൃల

Slide 25

Slide 25 text

Toward Personalized Product Search for eCommerce Sites: A Case Study in Yahoo! Taiwan • Y!୆࿷ͷeίϚʔεαΠτʹ͓͚Δݕࡧ • ଟ͘ͷݕࡧΫΤϦ͸ҰൠతͩͬͨΓᐆດͰɺϢʔβ͸๲େ ͳ݁Ռͷத͔Β໨తͷ΋ͷΛ୳͞Ͷ͹ͳΒͳ͍ • ςΩετ৘ใ͚ͩͰͷϚονϯάͰ͸ෆे෼ͳͷͰɺϢʔ βͷڵຯʹԠͯ͡ύʔιφϥΠζͨ͠ϥϯΩϯάΛߦ͏

Slide 26

Slide 26 text

AdKDD & TargetAd 2017

Slide 27

Slide 27 text

AdKDD & TargetAd 2017 https://adkdd-targetad.wixsite.com/adkddtargetad2017/accepted-papers

Slide 28

Slide 28 text

AdKDD & TargetAd 2017 https://adkdd-targetad.wixsite.com/adkddtargetad2017/accepted-papers

Slide 29

Slide 29 text

AdKDD & TargetAd 2017 https://adkdd-targetad.wixsite.com/adkddtargetad2017/accepted-papers

Slide 30

Slide 30 text

Attribution Modeling Increases Efﬁciency of Bidding in Display Advertising • Criteo • ΞτϦϏϡʔγϣϯϞσϧΛ࢖ͬͨbidઓུ • 2nd price auctionʹ͓͚Δ࠷దઓུ͸ͦͷϦΫΤετͷՁ ஋Λೖࡳֹͱ͢Δ͜ͱ͕ͩɺਖ਼֬ͳՁ஋ΛٻΊΔͨΊʹ ͸ΞτϦϏϡʔγϣϯϞσϧΛߟྀ͢Δඞཁ͕͋Δ

Slide 31

Slide 31 text

Blacklisting the Blacklist in Online Advertising • Dstillery • RTBʹ͓͍ͯ͸publisherଆ͕ϒϥοΫϦετΛ͍࣋ͬͯͯ ಛఆͷ޿ࠂΛ͸͘͜͡ͱ͕͋Δ • ϒϥοΫϦετͷ಺༰͸ඇެ։ • ແବͳbidΛආ͚ΔͨΊϒϥοΫϦετͷ಺༰Λਪఆ͢Δ ػೳΛ࡞ͬͨ • γεςϜෛՙ͕ݮΓɺউ཰͕޲্ͨ͠

Slide 32

Slide 32 text

Anti-Ad Blocking Strategy: Measuring its True Impact • Adobe Research • AdBlockΛݕग़ͯ͠ϢʔβʹϝοηʔδΛग़͢ઓུ͕͋Δ (՝ۚίʔε΁ͷ༠ಋͳͲ) • ͜ͷઓུ͸Ͳͷఔ౓༗ޮͳͷ͔ௐ΂ΔͨΊͷ౷ܭతख๏ ΛఏҊ • ͳΜΒ͔ͷࣄ৘ʹΑΓɺී௨ʹA/Bςετ͢ΔΘ͚ʹ͸͍ ͔ͳ͍Β͍͠

Slide 33

Slide 33 text

MM2RTB: Bringing Multimedia Metrics to Real-Time Bidding • ޿ࠂͷ࣭ΛධՁ͢ΔͨΊͷmultimedia metricsͱ͍͏ͷ͕ ͋Δ • contextual relevance, visual saliency, ad memorability • ϖʔδ಺ʹෳ਺ͷ޿ࠂεϩοτ͕͋Δͱ͖ɺmultimedia metricsΛߟྀͯ͠޿ࠂͷ૊Έ߹ΘͤΛબ୒͢Δख๏ͷఏ Ҋ

Slide 34

Slide 34 text

Data-Driven Reserve Prices for Social Advertising Auctions at LinkedIn • ηΧϯυϓϥΠεΦʔΫγϣϯʹ͓͚Δ࠷దͳϦβʔϒϓ ϥΠεΛٻΊΔํ๏Λ2छྨ(Ϣʔβ/ηάϝϯτϨϕϧ)ఏ Ҋ • ࣮γεςϜʹ࣮૷ͯ͠ޮՌΛݕূͨ͠

Slide 35

Slide 35 text

Optimal Reserve Price for Online Ads Trading Based on Inventory Identiﬁcation • Alibaba+Yahoo • ͜Ε΋reserve priceͷܾఆઓུʹؔ͢Δ࿩

Slide 36

Slide 36 text

Ranking and Calibrating Click-Attributed Purchases in Performance Display Advertising • click-attributionͰCV՝ۚϞσϧͷ޿ࠂ഑৴ • bidͷࡍʹ͸ҎԼͷखॱ͕ඞཁ • CV͞Ε΍͍͢޿ࠂͷީิΛྻڍ(ΞτϦϏϡʔγϣϯͷ ੍໿ͰɺΫϦοΫ཰΋ߴ͍ඞཁ͕͋Δ) • CV཰ʹԠͯ͡ೖࡳֹۚΛܾఆ(ਖ਼֬ͳCV཰ͷਪఆ͸ࠔ ೉) • ϥϯΩϯά+ΩϟϦϒϨʔγϣϯʹΑͬͯ͜ͷ໰୊Λղ͘

Slide 37

Slide 37 text

Cost-sensitive Learning for Utility Optimization in Online Advertising Auctions • CTR/CVR༧ଌʹ͸ɺhighly non-uniform misprediction cost͕ଘࡏ͢Δ • UtilityͷΑ͏ͳࢦඪΛ࢖ͬͯੑೳධՁ͢Δख๏͕ఏҊ͞Ε ͍ͯΔ • ֶश࣌ʹ΋UtilityΛߟྀͯ͠࠷దԽ͍ͨ͠ • log lossʹॏΈΛ͚ͭΔ͜ͱͰੑೳվળ͢Δख๏ͷఏҊ

Slide 38

Slide 38 text

A Practical Framework of Conversion Rate Prediction for Online Display Advertising • CVR༧ଌʹؔ͢Δ༷ʑͳख๏ͷఏҊ • Over prediction΁ͷରԠ • delayed feedbackΛߟֶྀͨ͠श • ΞτϦϏϡʔγϣϯϞσϧΛߟྀͨ͠ೖࡳֹิਖ਼

Slide 39

Slide 39 text

An Ensemble-based Approach to Click- Through Rate Prediction for Promoted Listings at Etsy • Etsyͷlisting adͰ࢖ΘΕ͍ͯΔCTRਪఆγεςϜͷ঺հ

Slide 40

Slide 40 text

Proﬁt Maximization for Online Advertising Demand-Side Platform • CPC/CPAϞσϧʹ͓͍ͯɺDSPͷརӹΛ࠷େԽ͍ͨ͠ • ࠷దͳೖࡳઓུ(Ͳͷ޿ࠂΛ͍͘ΒͰೖࡳ͢Δ͔)ʹ͍ͭͯ ߟ࡯ • Ωϟϯϖʔϯ༧ࢉͱϖʔγϯάɺλʔήςΟϯάɺimpڙ څྔͷ੍໿͕͋Δ • ࣮༻తʹղ͚ΔϞσϧΛఏҊ

Slide 41

Slide 41 text

Deep & Cross Network for Ad Click Predictions • Google • Deep & Cross networkͱ͍͏DLͷϞσϧΛఏҊ • ަޓ࡞༻Λཅʹѻ͑Δ • CTRਪఆʹར༻ͯ͠ߴੑೳͩͬͨ

Slide 42

Slide 42 text

AdKDD & TargetAd 2018

Slide 43

Slide 43 text

AdKDD & TargetAd 2018 https://adkdd-targetad.wixsite.com/2018/accepted-papers

Slide 44

Slide 44 text

AdKDD & TargetAd 2018 https://adkdd-targetad.wixsite.com/2018/accepted-papers

Slide 45

Slide 45 text

A Large Scale Benchmark for Uplift Modeling • Criteo • Uplift modelingͷݕূʹ࢖͑Δେن໛σʔλΛެ։ͨ͠ • https://ailab.criteo.com/criteo-uplift-prediction-dataset/

Slide 46

Slide 46 text

Deep Density Networks and Uncertainty in Recommender Systems • Taboola(޿ࠂϓϥοτϑΥʔϜΛ΍ͬͯΔձࣾ) • Deep Density Networksͱ͍͏ϞσϧͷఏҊ • content based+ڠௐϑΟϧλϦϯάͷϋΠϒϦου • ༧ଌͷෆ࣮֬ੑΛදݱͰ͖Δ • ε-greedyͱUCBΛ૊Έ߹ΘͤͨΑ͏ͳϩδοΫͰ exploration/exploitation͢Δ

Slide 47

Slide 47 text

Deep Neural Net with Attention for Multi-channel Multi-touch Attribution • DNNΛ࢖ͬͨMulti-touch attributionϞσϧͷఏҊ

Slide 48

Slide 48 text

Deep Policy Optimization for E-commerce Sponsored Search Ranking Strategy • 2018͔ΒDNNΛ࢖ͬͨൃද͕໨ཱͪ·͢Ͷ • εϙϯαʔυαʔνͷϥϯΩϯά໰୊Λਂ૚ڧԽֶशͰ ղ͘

Slide 49

Slide 49 text

Designing Experiments to Measure Incrementality on Facebook • Facebookʹ͓͚Δincrementalityܭଌख๏ͷղઆ • upliftͱincrementalityͷҧ͍͕Ṗ • ୯ͳΔA/BςετΑΓॊೈͳ࣮ݧ͕Մೳ

Slide 50

Slide 50 text

Dynamic Hierarchical Empirical Bayes: A Predictive Model Applied to Online Advertising • Adobe • εϙϯαʔυαʔνʹ͓͚ΔCTR/CVRਪఆϞσϧ • ֊૚ϕΠζʹ͓͍ͯɺσʔλʹԠͯ͡ಈతʹ֊૚ߏ଄Λܾ ఆ͢ΔΑ͏ͳϞσϧͷఏҊ

Slide 51

Slide 51 text

Forecasting Granular Audience Size for Online Advertising • ಛఆͷηάϝϯτʹ޿ࠂ഑৴͢Δͱ͖ɺࣄલʹΦʔσΟΤ ϯεͷαΠζ͕஌Γ͍ͨ • Frequent Itemset Miningͱ͍͏ख๏ΛݩʹɺͦͷΑ͏ͳਪ ఆΛ͢Δख๏ΛఏҊ

Slide 52

Slide 52 text

Mini-Batch AUC Optimization • ֶश࣌ʹAUCΛ௚઀࠷దԽ͍͕ͨ͠ɺେن໛σʔληο τʹ͓͍ͯ͸೉͔ͬͨ͠ • ϛχόονΛ࢖͏͜ͱͰߴ଎ͳֶशΛՄೳͱ͢Δख๏Λ ఏҊ

Slide 53

Slide 53 text

Optimal Bidding, Allocation and Budget Spending for a Demand Side Platform Under Many Auction Types • ৽͍͠ೖࡳઓུϩδοΫͷఏҊ • 1st/2nd price auctionʹରԠɺ༧ࢉ੍໿΍ΫϥΠΞϯτ͝ ͱͷر๬ΛߟྀͰ͖Δ

Slide 54

Slide 54 text

AdKDD 2019

Slide 55

Slide 55 text

AdKDD 2019 https://www.adkdd.org/2019-papers-and-talks

Slide 56

Slide 56 text

AdKDD 2019 https://www.adkdd.org/2019-papers-and-talks

Slide 57

Slide 57 text

AdKDD 2019 https://www.adkdd.org/2019-papers-and-talks

Slide 58

Slide 58 text

Time-Aware Prospective Modeling of Users for Online Display Advertising • Yahoo • ࣌ܥྻΛߟྀͨ͠ϢʔβߦಈϞσϦϯάʹΑΓɺϢʔβͷ ҙਤΛߟྀͨ͠ਪఆΛߦ͏DNNϞσϧͷఏҊ

Slide 59

Slide 59 text

Graphing Crumbling Cookies • ΫϩεσόΠετϥοΩϯάΛ͍͕ͨ͠ɺۙ೥cookieͷ࢖ ༻੍ݶ͕ݫ͘͠ͳ͍ͬͯΔ(ITP౳) • Browser ﬁngerprintingͳͲͷΘΔ͍ٕज़Λ࢖Θͣ͜ͷঢ়گ ʹରԠ͢ΔͨΊͷख๏ΛఏҊ

Slide 60

Slide 60 text

Causally Driven Incremental Multi Touch Attribution Using a Recurrent Neural Network • JD.com(ژ౦঎৓): தࠃͷڊେECاۀ • multi-touch attributionΛRNNͰϞσϦϯά͢Δख๏Λఏ Ҋ • ͜ͷϞσϧ͸࣮ࡍ࢖ΘΕ͍ͯΔΒ͍͠

Slide 61

Slide 61 text

Combinatorial Keyword Recommendations for Sponsored Search with Deep Reinforcement Learning • Baidu • ਂ૚ڧԽֶशͰϨίϝϯμ࡞Δ΍ͭ

Slide 62

Slide 62 text

Feasible Bidding Strategies through Pure Exploration Bandits • ೖࡳઓུͷީิ͕େྔʹ͋Γɺ͍͍΍ͭΛબͼ͍ͨ • ύϥϝλͷ૊Έ߹ΘͤͰ͔ͳΓީิ͕ଟ͍(~100)ͨΊɺී ௨ʹόϯσΟοτΞϧΰϦζϜΛ࢖͏ͷ͸ඇޮ཰ • ʮͦͦ͜͜Αͦ͞͏ʯͳީิΛߴ଎ʹൃݟ͢ΔͨΊͷख๏ ΛఏҊ

Slide 63

Slide 63 text

In-app Purchase Prediction Using Bayesian Personalized DwellDay Ranking • ژେ+CyberAgent • ΞϓϦ಺՝ۚͦ͠͏ͳϢʔβΛਪఆ͢ΔλεΫ • Bayesian personalized rankingΛݩʹͨ͠ख๏ΛఏҊ

Slide 64

Slide 64 text

Learning from Multi-User Activity Trails for B2B Ad Targeting • Yahoo • B2Bʹ͓͚Δ঎඼ͷߪೖϓϩηε͸ɺෳ਺ͷਓ͕ؒؔΘͬ ͨෳࡶͳ෺ʹͳΔ • ͜ͷΑ͏ͳϓϩηεʹରԠͨ͠޿ࠂͷ഑৴ઓུ͕ඞཁ • ಉ͡૊৫ʹ͍ΔΦʔσΟΤϯεΛਪఆɺͦΕΒͷߦಈΛ૯ ߹ͯ͠CV༧ଌ͢ΔϞσϧΛఏҊ

Slide 65

Slide 65 text

Modeling Advertiser Bidding Behaviors in Google Sponsored Search with a Mirror Attention Mechanism • SSPଆ͔ΒೖࡳऀͷߦಈΛϞσϦϯά͢Δ • ৽ػೳΛϦϦʔεͨ͠ࡍɺͦΕ͕௕ظతʹ޿ࠂओʹͲͷΑ ͏ͳӨڹΛٴ΅͔͢஌Γ͍ͨ…… ͱ͍͏ͷ͕Ϟνϕʔ γϣϯΒ͍͠

Slide 66

Slide 66 text

Optimal bidding: a dual approach • ৽͍͠ೖࡳઓུͷఏҊ

Slide 67

Slide 67 text

·ͱΊ • ޿ࠂܥ࿦จ͕ͨ͘͞Μ͋Δ͜ͱ͕Θ͔Γ·ͨ͠ • ͍͔͕Ͱ͔ͨ͠? • དྷ೥಄͘Β͍ʹ޿ࠂܥ࿦จಡॻձΛ΍Γ͍ͨͷͰɺڵຯ͋ Δਓ͸͓੠͕͚͍ͩ͘͞