Organizer: Edmondo Trentin (trentin@fbk.eu)

Topics
Although Automatic Speech Recognition (ASR) systems based on hidden Markov models (HMMs) are popular and effective under many circumstances, they suffer from drawbacks that limit applicability of ASR technology in the real world. Between the end of the Eighties and the beginning of the Nineties, several researchers began applying Artificial Neural Networks (ANN) to ASR, with the aim to overcome such limitations. ANNs allowed for significant results on reduced-scale tasks, e.g. phoneme recognition, but they substantially failed in dealing with long time-sequences of speech signals. As a consequence, "hybrid" systems were proposed, either by combining HMMs and ANNs within a single architecture, in order to take advantage from the properties of both, or by exploiting symbolic information via hybrid neuro-fuzzy paradigms. The purpose of this Special Session is to bring together researchers in the ANN community whose work is rooted in the hybrid-based ASR area as well as to provide the wider audience of neural scientists with novel hybrid architectures and algorithms, presenting state-of-the-art reviews and challenging new directions in the field.

Tentative program

1. Yoshua Bengio
   Departement d'Informatique et Recherche Operationnelle
   Universite' de Montreal
   Montreal, Canada
   e-mail: bengioy@IRO.UMontreal.CA
   "Incorporating Neural Networks in Probabilistic Models of Sequential Data"

2. Piero Cosi
   Institute of Phonetics and Dialectology (C.N.R.)
   Padova, Italy
   e-mail: cosi@csrf.pd.cnr.it
   "Hybrid NN-HMM architectures for automatic digit recognition"

3. Nikola Kasabov
   Department of Information Science, University of Otago
   Dunedin, New Zealand
   e-mail: nkasabov@otago.ac.nz
   "A Hybrid System for Robust Recognition of Noisy Speech Based on Evolving
   Fuzzy Neural Networks and Adaptive Filtering"

4. Edmondo Trentin and Marco Matassoni
   ITC-irst (Centro per la Ricerca Scientifica e Tecnologica)
   Povo (TN), Italy
   e-mail: trentin@fbk.eu, matasso@fbk.eu
   "The regularized SNN-TA model for recognition of noisy speech"

5. Roberto Gemello, Dario Albesano and Franco Mana
   CSELT - Centro Studi E Laboratori Telecomunicazioni
   Torino, Italy
   e-mail: roberto.gemello@cselt.it, dario.albesano@cselt.it, franco.mana@cselt.it
   "Cselt hybrid HMM/Neural Networks technology
   for continuous speech recognition"

Schedule
This Special Session is part of the Technical Program of IJCNN 2000, to be held in Como (Italy) from 24th to 27th of July, 2000. The  program has not been defined in details, yet. Please refer to the IJCNN 2000 Official Site for up-to-date details and news.

Registration
To attend the Special Session you obviously need only a regular registration for the Conference! Please refer to the IJCNN 2000 Registration page for detailed information, deadlines, fees and student grants (on the same page you will also find tourist infos and Hotel reservation forms), as well as for your actual registration.