Section: New Results

Users modeling and designing interaction

Contrary to lookup search engines that help users to retrieve specific items (e.g., names, numbers, short statements, or specific documents), Exploratory Search Systems (ESSs) are search engines that help users to explore a topic of interest. Exploratory search (ES) tasks are open-ended, multi-faceted, and iterative like learning or topic investigation [59] . Currently, the evaluation methods of ESSs are not entirely adapted to the special features of ES tasks, and do not effectively assess that ESSs support users in performing those tasks. Our research goal is to elaborate methods that effectively lead to this assessment. Two research actions were undertaken this year to contribute to achieve this goal.

This work was undertaken in the context of the VISIIR ANR project (http://www.agence-nationale-recherche.fr/?Projet=ANR-13-CORD-0009 ), leaded by the MinD team (I3S, UNS), with Stéphanie Lopez and Frédéric Precioso. One of the objectives of VISIIR is to design an interactive image search system based on user's eye movements. Detected by an eye-tracker, these movements allow the system to infer the image that the user is going to select; VISIIR's aim is to replace user's mouse clicks as a selection mode by implicit eye-clicks. Since ES behaviors can be observed in image search tasks, we designed an eye-tracking user test protocol on Discovery Hub (http://discoveryhub.co/ ) in order to: 1) increase our understanding of the ES process (at the cognitive and perceptual-motor levels); 2) verify if identified characteristics of gaze trajectories allow to infer the images that the user is going to select in ES tasks (as opposed to lookup search tasks).

This work was undertaken in the context of the PhD of Emilie Palagi. In [41] we introduced our approach for designing a user-centered evaluation method for ESSs. Our method takes into account users's ES behavior and is based on a cognitive model of an ES task. We will specially work on Discovery Hub (Wimmics project – Inria) and 3cixty (https://www.3cixty.com/ ) (EURECOM project) ESSs.

Together with Célia da Costa Pereira (I3S, UNS) and Mauro Dragoni of FBK, Trento, who visited our team for three months from April to June 2014, we have further refined our approach to concept-level sentiment analysis based on fuzzy logic [12] .

In the framework of the Semantic Educloud project, we developed a Web ontology for the description and representation of serious games. Such ontology describes the functional and design elements of the game, the profile and virtual context of the players and furthermore the datasets from the Web of data that the game can query. The ontology has been evaluated through a prototype, which is basically a serious game quiz based on DBpedia. As future work, it is planned to implement state-of-the-art recommendation algorithms of Linked Data resources that take into account the context and players' profile. Furthermore an integration with the EDUCLOUD platform is considered: EDUCLOUD is an emerging initiative in Sophia Antipolis, for the implementation of a platform of digital educational content accessible through the cloud from a 3D portal of resources, and any interface devices (tablets, smartphones, PCs). [44]