Grant: Engage Grants Program
University of Victoria
$24,996.00
Content-aware recommender systems using semantic analysis
FY 2016-17
Summary
Every day millions of news stories are read online, and there is an increasing tendency for readers to stay on the content provider's page for only one story. This has significant impact on the revenue for the site, and the ability of online news organizations to stay relevant. Digital Cavalier Technology Services Inc. (the company) provides smart personalization enterprise solutions that generate content recommendations that encourage users to read additional stories. The company fuses several collaborative filtering based approaches in order to accomplish this goal and would like to incorporate semantic analysis to their recommender engine generating patentable IP and creating additional funding and revenue generation opportunities. In this project we propose a solution with three parts. Firstly, we will build a baseline that uses distributional semantics and average word distance to create content-based predictions, without knowledge of entities or events. Secondly, we propose a system that generates recommendations based on identifying named entity mentions so that stories that mention the same entities will be given a boosted relatedness score. Thirdly, we propose to not only recognize entities, but relationships and events, which is often difficult because we must correctly interpret the dependencies between an event and its arguments. However, by focusing in on the people and events in news articles, rather than treating all words equally, we can create better content-based predictions. The successful approaches will be integrated with the existing collaborative filtering engine using a ranking algorithm. The company's R&D team will build upon the deliverables of this research project using the expert knowledge gained through this engagement to develop spin-off technologies such as automatic keyword tagging, as well as demand-driven publishing.
Program
Engage Grants Program
Recipient
Location
British Columbia, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
502835-2016
Name nom
Fyshe, Alona
Department département
Computer Science
Organization id
7
Institution établissement
University of Victoria
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2016
Competition year année de concours
2016
Award amount
24996
Program id
EGP
Program name en
Engage Grants Program
Program name fr
Programme de subventions d'engagement partenariat
Group en
Research partnerships programs
Group fr
Programmes de partenariat
Committee code
1554
Committee name en
Pacific Internal Decision Committee
Committee name fr
Comité de décision interne pour le Pacifique
Area of application code
1207
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Information, computer and communication technologies
Area of application fr
Technologie de l'information, des ordinateurs et des réseaux de communications
Research subject code
2801
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Natural language and speech understanding
Research subject fr
Langage naturel et reconnaissance de la parole
Partie
1
Nb partie
1
Application title
Content-aware recommender systems using semantic analysis
Keyword
NSERC
Application summary
Every day millions of news stories are read online, and there is an increasing tendency for readers to stay on the content provider's page for only one story. This has significant impact on the revenue for the site, and the ability of online news organizations to stay relevant. Digital Cavalier Technology Services Inc. (the company) provides smart personalization enterprise solutions that generate content recommendations that encourage users to read additional stories. The company fuses several collaborative filtering based approaches in order to accomplish this goal and would like to incorporate semantic analysis to their recommender engine generating patentable IP and creating additional funding and revenue generation opportunities. In this project we propose a solution with three parts. Firstly, we will build a baseline that uses distributional semantics and average word distance to create content-based predictions, without knowledge of entities or events. Secondly, we propose a system that generates recommendations based on identifying named entity mentions so that stories that mention the same entities will be given a boosted relatedness score. Thirdly, we propose to not only recognize entities, but relationships and events, which is often difficult because we must correctly interpret the dependencies between an event and its arguments. However, by focusing in on the people and events in news articles, rather than treating all words equally, we can create better content-based predictions. The successful approaches will be integrated with the existing collaborative filtering engine using a ranking algorithm. The company's R&D team will build upon the deliverables of this research project using the expert knowledge gained through this engagement to develop spin-off technologies such as automatic keyword tagging, as well as demand-driven publishing.