Hybrid Recommender System

The first task of the project is to increase the relevance of content recommended to Unono users and decrease the cost of manually generating these recommendations by Unono. Each App in Unono starts with a recommendation section, but this content is for the time being chosen manually by Unono team, and thus does not bring a personalized recommendation to each user. However, when a user can be identified, e.g. if logged in or through cookies, it is possible to generate personalized recommendations, based on the preferences inferred from the previous actions by the user and/or the content of the current page. The second task is the exploration of crowdsourcing used for soft-skill semantic tagging, in self-presentation video, in order to assist the process of matching job applicants and job offers. The objective of this part is to develop a functional platform where: (a) students can post their self-presentation video, (b) softs-skills can be evaluated by the community (crowdsourcing) (c) companies can have access to applicants’ profiles. The solution will integrate crowdsourced mechanisms for presentation type video, evaluating soft skills with semantic tags, in order to increase matching between applicants and job offers.
Unono
Idiap Research Institute
The Ark
Aug 01, 2015
Mar 31, 2016