Program

The program of this school, spread over 5 days (from Monday, October 30 to Friday, November 3), is mainly composed of fundamental courses (mornings) supplemented by advanced presentations (afternoons) on specific research topics. The program also contains presentations on applications/case studies/projects as well as a special session, in the form of posters and short talks for PhD students. The teaching language is English.


Abstract:

This lecture is composed of two distinct parts, and concentrates on the problem of modelling uncertainty and reasoning with it. 


In the first part, I will introduce the main uncertainty theories as well as some motivations to go beyond the basic models that are sets and probabilities. I will then introduce some reasoning and information treatment problems, as well as corresponding proposed solutions, providing pointers to go further into their studies. Treated problems include the modelling of available information, merging or fusing information, conditioning, modelling independence, decision. 


In the second part of the course, I will concentrate on the role of imprecision and uncertainty models in learning problems. I will mainly treat two aspects: on the one hand how to obtain cautious predictions, either through the learning of natively cautious models, or through the use of post-hoc approaches providing cautious predictions; on the other hand what becomes the learning problem when data themselves become uncertain, detailing what are the associated challenges but also why considering uncertain data may actually represent an interesting opportunity. 













Coffee breaks are scheduled every day, in the morning from 10:30 a.m. to 10:45 a.m. and in the afternoon from 4 p.m. to 4:15 p.m.

Lunch breaks are scheduled from 12 p.m. to 2 p.m. Dinners are scheduled for 7:30 p.m.