KIWI is part of the department "Complex Systems, Artificial Intelligence and Robotics" of the LORIA laboratory
The KIWI team is interested in automatic analysis of digital traces, either to build predictive behavior models or to design personalized recommendation systems. Such work has many application domains.
Our goal is to provide students with adapted pedagogical resources, and teachers with feedback (learning analytics).
We design models to early diagnose pathologies (sports accidents, neurodegenerative diseases).
Digital media services
We help users to come up with personalized music, movies, and TV shows.
Recommender systems may be used to find relevant items, and to make users' decisions easier.
We improve the visitor experience in physical spaces by computing adapted pathes (museums, smart cities, ...).
Information and social networks
We improve the interactions between users and information systems / social networks.
Highlights of the KIWI team
An Outstanding Paper Award was given to Benjamin Gras, Armelle Brun, Boyer Anne for their paper “Identifying Grey Sheep Users in Collaborative Filtering: a Distribution-Based Technique”.ACM UMAP 2016
The KIWI team is part of the CrossCult H2020 european project, which has been ranked 1st out of 137 proposals in 2015.H2020 CrossCult
Lina Fahed successfully defended her PhD thesis entitled “Predicting and influencing the occurrence of events in a complex sequence” on October, the 27th.Lina Fahed
The KIWI team, in collaboration with Sailendra SAS, won the PriceMinister Rakuten competition on behavioral analysis and recommendation (1st in quality of recommendations, 2nd in calculation time). See the CNRS website for more information.PriceMinister