Web site : http://xmlmining.lip6.fr

Many domains and applications are concerned with complex data composed of elementary components linked according to some structural or logical organization. In biology for example, the 3D structure of proteins results from the interaction of its different components, phylogenetic trees are used to model species evolution and relations, RNA structures are often compared via tree edit distances. In the text domain, the diffusion of new data formats like XML and HTML has considerably changed the domains of Information Retrieval and Information Extraction (IE). For structured IR tasks, both the logical structure and the content information have to be considered simultaneously. For information extraction applications, structure plays a major role for identifying the relations between the different elements to be extracted. For building and querying heterogeneous XML databases, learning automatically from XML collections the relations between different formats and the transformations between different structured document representations is a key problem. Other application domains concerned with structured data do include image processing, multimedia (video), natural language processing, social networks, etc. Handling structured data has become a major challenge for these domains and different communities have been developing for some years their own methods for dealing with structured data. The Machine Learning community should be a major actor in this area. Among the many open problems for handling structured data, we will focus in this challenge on the two generic tasks of classification and clustering and one structure specific task which is Structure Mapping. The goal of the challenge is therefore to explore algorithmic, theoretical and practical issues regarding the classification, clustering and structure mapping of structured data.

Organisers