A Dialogue-based Approach Enhanced with Situation Awareness and Reinforcement Learning for Ubiquitous Access to Linked Data

The main barrier to a mainstream adoption of Semantic Web and Linked Data is the difficulty for users to search and retrieve the required information in this huge network of data. This work proposes a novel approach for Ubiquitous Browsing and Searching Linked Data. The proposed approach lays on a conceptual communication model, namely Interactive Alignment, for disambiguating both users' intentions and requests in the context of an information-seeking dialogue among humans and machines. More in details, the alignment between humans' intentions and machine comprehension is improved by identifying situations the users are involved in and considering users' situated preferences. Situation Awareness techniques are employed to identify and handle perceptions about occurring situations and Reinforcement Learning algorithms are exploited in order to elicit and acquire part of the user's mental model regarding her situated preferences. An ISU-based Dialogue System Architecture has been chosen to handle human-computer interaction and allow interactive alignment. Furthermore, the paper proposes a case study in which users are customers in U-commerce scenarios and they are looking for products or services to purchase.