Distributed computing for the analysis of complex systems
Team leader: Maurice Tchuenté
Affiliation: University of Yaounde I, Cameroon
Contact: maurice.tchuente@gmail.com
Co-team leader Inria: Jean-François Méhaut
Affiliation: Inria Grenoble
Contact: Inria’s associated team : MAGNET, ADAM (Lille), GCG (Grenoble Compiler Group), URBANET (Inria/INSA Lyon)
Objectives
The work of the team IDASCO was organized around three axes :
- Learning techniques for social network mining, especially with application to the detection and monitoring of communities, and recommendation systems;
- Search of traces of multimedia applications on embedded systems, generated by complex architectures centered around integrated components called Multi-Processor Systems-on-Chips. The aim is to detect anomalies useful for debugging these applications;
- Parallelization of complex system analysis algorithms. In fact, the treatment in reasonable time of wide networks requires computing power that goes beyond the capacity of a single-processor machine. Thus, the use of parallelism cannot be avoided. We are particularly interested in multi-core architectures that started to be commonly used in recent years. The objective is to propose a dedicated language providing ease of programming, with a compiler that produces optimized code for social network algorithms running on multi-core platforms.
