Description
Summary: | "Presented by Bojan Babic, Sr. Software Engineer at Groupon. Groupon is a dynamic Marketplace where we try to match millions of the deals organized in different verticals and taxonomies with the demand across 20 countries around the world. Modeling such complex relationships requires sophisticated machine learning models that utilize hundreds of user and deal features. Customers discover deals by directly entering the search query or browsing on the mobile or desktop devices. The purpose of this paper is to describe a series of techniques used to improve various parts of Search and Ranking algorithms by utilizing the embeddings representations of the user and deal features. The paper will describe improvements made in Query Understanding, Deal Classification, Similar Deals Recommendations and computation of an Image Propensity to Purchase that leverage respective embedding feature representations."--Resource description page
|
Item Description: | Title from resource description page (Safari, viewed October 6, 2020). - Place of publication from title screen |
Physical Description: | 1 streaming video file (27 min., 1 sec.) |