GBR workshop 2007

Alicante, June 14, 2007

Home

Call for participation

Program

Registration

Local Information

Program at a glance

• 9h00-9h30 Opening
• 9h30-11h30 Session #1
• 11h30-12h00 Coffee break
• 12h00-13h00 Session #2
• 13h00-14h30 Lunch
• 14h30-16h00 Session #3
• 16h00-16h30 Coffee break
• 16h30-18h00 Session #4: Discussion

Session 1: 9h30 -11h30

• Mining Frequent Closed Unordered Trees Through Natural Representations. José L. Balcazar, Albert Bifet, Antoni Lozano. Universitat Politècnica de Catalunya, Barcelona
• Learning with spectral representations and use of MDL principles. Edwin Hancock, University of York.
• Characterizing Implications of Injective Partial Orders. José L. Balcazar. Universitat Politècnica de Catalunya, Barcelona. Gemma C. Garriga. Helsinki University of Technology.
• Genetic Approximate Matching of Attributed Relational Graphs. Thomas Baerecke, Marcin Detynieck, Laboratoire d’Informatique de Paris 6. Stefano Berretti, Alberto Del Bimbo. University of Firenze.

Session 2: 12h-13h

• Molecular Graph Kernels for Drug Discovery. Anthony Demco, Craig Saunders, Alex Dolia. University of Southampton.
• Graphs Regularization for Data Sets and Images: Filtering and Semi-Supervised Classification. Vinh Thong Ta, Olivier Lezoray, Abderrahim Elmoataz. Caen.

Session 3: 14h30-16h

Graph Signature: A Simple Approach for Clustering Similar Graph. Rashid-Jalal Qureshi, Jean-Yves Ramel, Hubert Cardot Université François-Rabelais de Tours.

Szemerédi's Regularity Lemma and Pairwise Clustering. Anna Sperotto, University of Twente, Marcello Pelillo, Universitµa Ca' Foscari di Venezia.

Web People Search Disambiguation using Random Walks. José Iria, Fabio Ciravegna University of Sheffield.

Session 4: Discussion

The discussion is centered about the questions of graph distances, kernels and alignments.
What challenges?
Algorithmic issues ?
The participants to the workshop are encouraged to prepare about 5 slides addressing the following questions:

• What is my favourite distance, kernel for graphs (or at least one I want to talk about)?
• What are the modelising reasons for this? Is it useful? well adapted? powerful?
• What are the mathematical properties? Is it a metric? Other?
• What are the algorithmic consequences of using it?

A nice example would be great!