Recommender systems are becoming an integral part of our life by continuously monitoring and updating user models to recommend information that may be useful without requiring explicit user actions. However, is unclear whether Intelligent personal assistants can infer multiple complex tasks and how useful they are considering they provide little or no explanations and control to users. These are still timely and open challenges considering unexplored comprehensive monitoring of users beyond specific applications and realistic evaluation in field appropriation. The overall question of this project is whether an entity based approach can provide effective representations for an open and editable user model towards providing resourceful control of recommendations with benefits for trust , satisfaction , task performance while empowering choice of users and minimising risks such as exploitative persuasion.

2023

2027