"Profiling users by activity tracker data collected by consumer wearable devices"

Activity trackers use proprietary algorithms that can usually estimate steps, distance, calorie burn, and hours of sleep. They are often considered as cost-effective solutions to support weight loss strategies. However, in terms of behavior adaptation to these goals, these devices do not provide specific features beyond monitoring the achievement of daily goals, such as steps walked and caloric outtake. By analyzing a large dataset of signals collected by these devices, we evaluate recent technologies such as Recurrent Neural Networks, Time2Vec, Deep Echo State networks and Transformers; with the goal of modeling user behaviors and predicting relevant patterns useful in the context of recommender systems.

Bio: Fabio Gasparetti is currently Associate Professor at the AI Lab of ROMA TRE University. Previously visiting scientist and researcher at Nokia Bell Labs at Cambridge UK and Xerox PARC. Since 2017 he is member of the Italian IA-Gov laboratory. The A.I. initiative, formerly named Task Force, is promoted by AgID (Agency for Digital Italy - Presidency of the Council of Ministers). He is member of SIGCHI Italy, the Italian ACM SIGCHI Chapter. He got his PhD in 2005 doing research on Adaptive Web Search. He has collaborated and is still collaborating in National and International research projects regarding Internet technologies. Principal research interests include User Modeling and User Adapted Interaction, Human-computer personalized interaction, Recommender Systems and Context-awareness.