Grant: Collaborative Research and Development Grants
University of Alberta
$29,958.00
Large-scale learning for online advertising and web recommendations
EF 2016-17
Résumé
Google is one of the most profitable companies in the world, thanks largely to its online advertising business. They and other online advertisers agree that predicting which ads user are likely to click on - and charging based on that - is crucial to their success. The click-through rate predictors these companies use are the largest and most-valuable machine-learning systems on the planet. Their exact operation is also a closely-guarded secret, meaning new entrants must develop their predictors essentially from scratch. Canadian companies wishing to enter this lucrative market thus face a very difficult challenge. One such company, HybridForge, Inc., is our partner on this grant. This research proposal seeks to lower the entry barrier to this market, by exploring the design space of solutions to its particular challenges: the high dimensionality and sparsity of the inputs to the predictor, the cold-start problem, and the extreme skewness in advertising traffic data. In collaboration with HybridForge, we will create a prototype predictor that responds to all of these challenges, and then we will evaluate each component on public large-scale learning benchmarks. It is not currently possible to evaluate the entire system on internet advertising data, as this is closely held by the advertising companies. Instead, we will evaluate the predictor on the closely-related problem of recommender systems. The knowledge generated from this research will directly assist Canadian firms seeking to enter the online advertising market, as well as those developing recommender systems. HybridForge, in particular, intends to pursue commercialization of the prototype recommender as a new element of their Web design business, while an online advertising play is the firm's core long-term strategy.
Programme
Collaborative Research and Development Grants
Bénéficiaire
Emplacement
Alberta, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
488241-2015
Name nom
Dick, Scott
Department département
Electrical and Computer Engineering
Organization id
9
Institution établissement
University of Alberta
Province en
Alberta
Province fr
Alberta
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2016
Competition year année de concours
2015
Award amount
29958
Program id
CRDPJ
Program name en
Collaborative Research and Development Grants
Program name fr
Subventions de recherche et développement coopérative
Group en
Research partnerships programs
Group fr
Programmes de partenariat
Committee code
96
Committee name en
RPP Internal Decision Cttee
Committee name fr
Cté de décision interne - PPR
Area of application code
802
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Information systems and technology
Area of application fr
Systèmes d'information et technologie de l'information
Research subject code
2800
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Artificial intelligence (computer vision, use2603)
Research subject fr
Intelligence artificielle (vision artificielle, utiliser 2603)
Partie
1
Nb partie
5
Application title
Large-scale learning for online advertising and web recommendations
Keyword
NSERC
Application summary
Google is one of the most profitable companies in the world, thanks largely to its online advertising business. They and other online advertisers agree that predicting which ads user are likely to click on - and charging based on that - is crucial to their success. The click-through rate predictors these companies use are the largest and most-valuable machine-learning systems on the planet. Their exact operation is also a closely-guarded secret, meaning new entrants must develop their predictors essentially from scratch. Canadian companies wishing to enter this lucrative market thus face a very difficult challenge. One such company, HybridForge, Inc., is our partner on this grant. This research proposal seeks to lower the entry barrier to this market, by exploring the design space of solutions to its particular challenges: the high dimensionality and sparsity of the inputs to the predictor, the cold-start problem, and the extreme skewness in advertising traffic data. In collaboration with HybridForge, we will create a prototype predictor that responds to all of these challenges, and then we will evaluate each component on public large-scale learning benchmarks. It is not currently possible to evaluate the entire system on internet advertising data, as this is closely held by the advertising companies. Instead, we will evaluate the predictor on the closely-related problem of recommender systems. The knowledge generated from this research will directly assist Canadian firms seeking to enter the online advertising market, as well as those developing recommender systems. HybridForge, in particular, intends to pursue commercialization of the prototype recommender as a new element of their Web design business, while an online advertising play is the firm's core long-term strategy.