Tutorials 

The tutorials will be held on SUNDAY, March 14, 2010 at ISIAM (Boulevard Hassan 1er, Quartier Dakhla, Agadir).

The schedule is as follows:

    -       14:00 – 16:00 – Tutorial 1

    -       16:00 – 16:30 – Coffee Break

    -       16:30 – 18:30 – Tutorial 2

Tutorial 1: Concept-based Learning Models Professor Sergei O. Kuznetsov, State University Higher School of Economics, Department of Applied Mathematics and Information Science, Moscow, Russia Presenter's Website Presenter's Biography Abstract: The importance of lattices in inductive learning was noticed in early 1970s in the works on antiunification, which give rise to Inductive Logic Programming. Techniques related to extraction of knowledge from data were among the mainstream of Formal Concept Analysis (FCA) from the very beginning in early 1980s, when researchers started to study the notion of attribute implication and implication bases. Generating bases of implications from contexts, e.g. in the process of  Attribute exploration, is a learning procedure, as well as generating bases of association rules, the latter being very closely related to concept lattices and their diagrams. In our tutorial we describe several learning models that are based on concept lattices, such as learning implications and association rules, JSM-method of hypothesis generation, learning with pattern structures, etc., as well as some well-known machine learning models that are naturally described in terms of concept lattices, among them version spaces and induction of decision trees. We employ some open-source software that realize learning models related to concept lattices and consider practical examples from several application domains.

Tutorial 2: Social Network Analysis. Professor Sergei Obiedkov, State University Higher School of Economics and Information Science, Russia Presenter's Website Presenter's DBLP page Presenter's Biography Dr. Camille ROTH, CNRS (Centre National de la Recherche Scientifique), Paris, France Presenter's Website Presenter's Biography Abstract: Social network analysis (SNA) is an active interdisciplinary research area, in which mathematical sociology and computer science play a notable role. The recent interest in social networks has been supported by significant advances in computing capabilities and electronic data availability for several social systems: scientists, webloggers, online customers, computer-based collaboration-enhancing devices, inter alia. In particular, knowledge networks, i.e., interaction networks where agents produce or exchange knowledge, are the focus of many current studies, both qualitative and quantitative. Among these, community-detection issues such as finding agents sharing sets of identical patterns are a key topic. Social network analysis is proficient in methods aimed at discovering, describing, and plausibly organizing various kinds of social communities. At the same time, conceptual structures can yield a fruitful insight in this regard, be it in relation to affiliation networks (actors belonging to the same organizations, participating in identical events), interaction networks (groups of agents related to similar groups of agents), or epistemic communities (i.e., actors dealing with identical topics, such as scientific communities or weblogs). Indeed, some applications of concept lattices in sociology have been proposed since the early 1990s. The tutorial covers main topics of social network analysis, such as community detection, role identification, data mining in networks, link prediction, etc., with respect to various kinds of networks, such as affiliation, collaboration, citation, one-mode and two-mode networks, folksonomies, etc. We describe traditional SNA approaches to each of the topics, as well as present FCA-based approaches if such have been developed. We indicate the opportunities for formal concept analysis in social network research by proposing possible bridges between these frameworks and by presenting issues of mathematical sociology which could benefit from conceptual structures, so as to facilitate collaboration between the two fields.