ANNPR 2008 was sponsored by the IAPR Technical Committee TC3 on Neural Networks and Computational Intelligence and was co-organized by the TC3 Chair Simone Marinai (University of Florence, Italy) and Prof. Lionel Prevost of UPMC. The workshop received 57 submissions, from which 29 were accepted for presentation, including 18 oral presentations and 11 poster presentations. About 40 participants from five continents (Europe, Asia, North America, South America, and Africa) attended the workshop.

The third Workshop on ANNPR follows the first in Florence (Italy) in 2003 and the second in Ulm (Germany) in 2006. The background of this series of workshop is that artificial neural networks have been successfully applied to pattern recognition since the 1980s, and have received much attention from the research community. Many new machine learning methods, like support vector machines and ensemble methods, have joined neural networks to solve problems of classification, regression, clustering, feature extraction, and so on. The aim of the Workshop ANNPR is thus to provide a forum for exchanging advances in the joint areas of neural networks, machine learning, and pattern recognition.

The program of the ANNPR 2008 included seven oral sessions, one poster session, and an invited talk. The oral sessions were titled “Unsupervised Learning”, “Supervised Learning”, “Clustering”, “Neural Networks”, “Multiple Classifiers”, Speech Recognition”, and “Feature Selection”. Each oral presentation was allocated 30 minutes such that there was plenty of time for questions and discussion. There were indeed plenty of discussions on the presented topics, not only at the sessions, but also at the breaks and lunch/dinner times.

The invited talk, titled “Predicting Structured Outputs—A Reinforcement Learning Approach”, was given by Prof. Patrick Gallinari of UPMC. Structured prediction is involved in many pattern recognition applications, and poses a new challenge of machine learning. The talk by Prof. Gallinari summarized generative and discriminative models for structured prediction and presented the authors’ own work on successful use of reinforcement learning to solve this problem.

The active discussions at the workshop were partially due to the host’s nice organization: three lunches and a dinner were provided at the venue. The attendees also enjoyed a cruiser tour on the Seine River, a visit to the Museum d’Orsay, and the banquet at the Maison de l'Amérique Latine in the historic center of Paris.

During ANNPR 2008, the attendees discussed how to organize the next workshop of the series in 2010. Possibilities of holding ANNPR out of Europe and in combination with other large-scale conferences will be considered.

Workshop Report:  ANNPR 2008

TC3’s web site:

www.dsi.unifi.it/TC3

Report prepared by:

Cheng-Lin Liu

Co-chairs:

 

Simone Marinai (Italy)

Lionel Prevost (France)

Proceedings of the workshop have been published by Springer-Verlag in Lecture Notes in Computer Science Series (volume number 5064).

 

Click on the image to go to the publisher’s web site for this volume.