Click here for Top of Page

Traditional pattern recognition techniques commonly rely on feature-based, vectorial representation of data. However, this approach suffers from several limitations, in particular when the objects being analysed are naturally described in terms of relations between parts, rather than by a fixed number of features. 

The SIMBAD workshop series aims at bringing together researchers and practitioners interested in the field of similarity-based pattern recognition and machine learning. The 1st SIMBAD workshop was held in Venice, in the beautiful context of the historical Dolfin Palace of Ca' Foscari University. It was organized by Marcello Pelillo (University of Venice, Italy) and Edwin Hancock (University of York, UK) and marked the end of the EU FP7 Projects SIMBAD (simbad-fp7.eu). The workshop was intended to be an informal forum where nearly 50 participants would meet to discuss fundamental questions that arise when abandoning the standard vectorial framework. In particular: 1) how can one obtain suitable similarity information from data representations that are more powerful than, or simply different from, the vectorial? 2) how can one use similarity information in order to perform learning and classification tasks? A wide range of problems and perspectives were covered, from supervised to unsupervised learning, from generative to discriminative models, and from theoretical issues to real-world practical applications.

During the meeting, three invited speakers contributed to the debate.

Marco Gori, University of Siena (Italy), gave a talk on the theory of support constraint machines, an extension of the classical kernel machines.

Ulrike Hahn, Cardiff University (UK), overviewed the psychological theories of similarity and key behavioural data.

John Shawe-Taylor, University College London (UK), gave a lecture on the imitations of kernel and multiple kernel learning.

All the oral presentations were filmed by Videolectures and are freely available online at videolectures.net/simbad2011_venice.

The next SIMBAD workshop will be held in York, UK, in July, 2013, and will be organized by Edwin Hancock and Marcello Pelillo.

Workshop Report:  SIMBAD 2011

Report prepared by Luca Rossi

Text Box: 1st IAPR International Workshop on
Similarity-Based Pattern Analysis and Recognition

September 2830, 2011
Venice, Italy

General Chairs:

 

Marcello Pelillo, IAPR Fellow (Italy)

Edwin Hancock, IAPR Fellow (UK)

Proceedings of the

workshop have been

published by Springer

in the series

Lecture Notes in

Computer Science

(Volume 7005)

 

Click here to go to the publisher’s

web site for the volume.

Right Arrow: Previous
Right Arrow: Next