Keynote Speaker 1: Alberto d'Onofrio, International Prevention Research Institute (iPRI), France
Human Behavior and the Spread of Infectious Diseases: a challenge for modeling
This talk concerns a fast growing research area: modeling the influence of information-driven human behavior on the spread and control of infectious diseases. In particular, we shall focus on two main and inter-related "core" topics: behavioral changes in response to global (or "perceived global"...) threats, and the pseudo-rational opposition to vaccines. Indeed, people are likely to change their behavior and their propensity to vaccinate themselves and their children based on information and, even more often, rumors about the spread of a disease. This, implicitly, induces a feedback that can deeply affect the dynamics of epidemics and endemics. In order to make realistic predictions, modelers must go beyond classical mathematical epidemiology, where, in anology with systems biology, the individuals are abstracted as particles in brownian motion.
Keynote Speaker 2: Elisa Fromont, Université de Lyon, Université de St-Etienne, France
Mine first to see better
I will explain how data mining techniques such as pattern mining or (semi-supervised) clustering can and should be used to improve fundamental computer vision tasks such as image classification, image or video retrieval or object tracking in videos. The main idea is to build on low level vision features such as segmentations or SIFT bag-of-visual-words to construct more discriminant and invariant "mid-level" descriptors. I will show examples of success stories that have used this pattern mining phase in the last years. On the algorithmic point of view, I will focus on a dynamic plane graph mining algorithm that integrates spatio-temporal constraints and can be used to help tracking objects in videos in an unsupervised way.