ijcai-99 Home Page

D1 - Neural Networks for Data Structures Principles and Applications

Sunday, AM

Paolo Frasconi & Alessandro Sperduti

The purpose of the tutorial is to examine the state of the art in the use of connectionist networks for processing data structures and to present a unified view of formalisms and tools for dealing with rich data representations, covering connectionist architectures for data structures, learning algorithms, and applications. In particular, we will show that it is possible to represent and classify structured information very naturally. Moreover, it is possible to formalize several supervised models for classification of structures which stem very naturally from well known models, such as back propagation through time networks, real-time recurrent networks, simple recurrent networks, recurrent cascade correlation networks, and neural trees.

Because many concepts and formal tools are inherited from the theoretical framework of recurrent networks for sequence processing, the tutorial will begin with a review of basic concepts underpinning recurrent neural networks, for those attendees which are not familiar with such a class of models. Algorithms for training recursive neural networks are presented as generalizations of gradient computation algorithms for recurrent nets, and complexity as well computational issues are discussed. Finally, examples of applications in chemistry, structural pattern recognition, and theorem proving are presented.

Prerequisite knowledge:
Although we will briefly review the essential concepts for data structure and neural networks, we will assume that the attendants will be familiar with data structures; we also assume basic knowledge of linear algebra and calculus for the treatment of some neural network paradigms.

Paolo Frasconi received the M.Sc. degree in Electronic Engineering in 1990 and the Ph.D. degree in Computer Science in 1994, both from the University of Florence, Italy. He is currently Associate Professor with Dipartimento di Ingegneria Elettrica ed Elettronica at the University of Cagliari, Italy. He was Assistant Professor with the Dipartimento di Sistemi e Informatica at the University of Florence, Italy. In 1992 he was a Visiting Scholar in the Department of Brain and Cognitive Science at the Massachusetts Institute of Technology, Cambridge. In 1994 he was a Visiting Scientist at Centro Studi e Laboratori Telecomunicazioni (CSELT), Turin. His current research interests include neural networks, Markovian models, and graphical models, with particular emphasis on problems involving learning about sequential and structured information. Paolo Frasconi is the author of around 50 refereed papers mainly in the areas of graphical models for learning, neural networks, pattern recognition, artificial intelligence.

Alessandro Sperduti received his university education from the University of Pisa, Italy ("Laurea" and Doctoral degrees in 1988 and 1993, respectively, all in Computer Science). In 1993 he spent a period at the International Computer Science Institute, Berkeley, supported by a postdoctoral fellowship. In 1994 he moved back to the Computer Science Department, University of Pisa, where he was Assistant Professor, and where he presently is Associate Professor. His research interests include pattern recognition, image processing, neural networks, hybrid systems. In the field of hybrid systems his work has focused on the integration of symbolic and connectionist systems. He contributed to the organization of several workshops on this subject and he served also in the program committee of conferences on neural networks. Alessandro Sperduti is the author of around 50 refereed papers mainly in the areas of neural networks, fuzzy systems, pattern recognition, and image processing.


Webmaster: Sven Olofsson, sveno@dsv.su.se
Last modified: Mar 16, 1999