You are here

Research Topics

Approximate Graph Matching

Graph matching arises in many domains from the comparison of biological networks to scene recognition problems. Efficient methods are available for exact matching and people are now interested in approximate graph matching. Given an input query graph, the objective is to find this query in a target graph, allowing some approximations : node gaps, structural differences,…

Constraint Programming for Data Mining

Constraint programming has seduced the data mining community with its highly declarative approach and flexibility. However, these advantages do not offer a guarantee of efficiency. Evidence of this is that the CP-based methods have not yet been as efficient as the specialized methods. In our lab we investigate thoses problem and provide new Constraints and methods to tackles pattern mining problem. Our recent result has shown how profitable the hybridization of CP and Data mining can be for both communities.

Global Constraints

Global Constraints and Filtering algorithms

Local Search Framework LS(Graph) for CSPs on graphs

The objective of this research is to extend the Comet language by constructing a framework called LS(Graph) which aims at simplifying the modeling of CSPs on graphs. The goal of LS(Graph) is to strengthen the modeling features of Comet for an important class of CSP (CSPs on graphs) and to enhance compositionality.


Publications related to OscaR Solver (


Application, heuristics, constraints