CompanyDAPHNE : DATA PARALLELISMNEURAL NETWORK SIMULATOR

نویسندگان

  • Paolo Frasconi
  • Marco Gori
  • Giovanni Soda
چکیده

In this paper we describe the guideline of Daphne, a parallel simulator for supervised recurrent neural networks trained by Backpropagation through time. The simulator has a modular structure, based on a parallel training kernel running on the CM-2 Connection Machine. The training kernel is written in CM Fortran in order to exploit some advantages of the slicewise execution model. The other modules are written in serial C code. They are used for designing and testing the network, and for interfacing with the training data. A dedicated language is available for deening the network architecture, which allows the use of linked modules. The implementation of the learning procedures is based on training example paral-lelism. This dimension of parallelism has been found to be eeective for learning static patterns using feedforward networks. We extend training example parallelism for learning sequences with full recurrent networks. Daphne is mainly conceived for applications in the eld of Automatic Speech Recognition, though it can also serve for simulating feedforward networks.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

The efficiency of Artificial Neural Network, Neuro-Fuzzy and Multivariate Regression models for runoff and erosion simulation using rainfall simulator

1- INTRODUCTION According to the complexity of environmental factors related to erosion and runoff, correct simulation of these variables find importance under rain intensity domain of watershed areas.  Although modeling of erosion and runoff by Artificial Neural Network and Neuro-Fuzzy based on rainfall-runoff and discharge-sediment models were widely applied by researchers, scrutinizing Arti...

متن کامل

Comparative study of computer simulation softwares

One of the methods for analyzing systems is simulation. Network simulation is a technique that models the behavior of the network by performing transaction calculations between different network entities and using mathematical formulas and taking observations from network products. A network simulator is a piece of software or hardware that predicts the behavior of a computer network without a ...

متن کامل

Internet network design for quality of service guarantee using Data Envelopment Analysis (DEA)

By developing the new services such as VoIP and Videoconference, using a mechanism is needed to support the quality of service of the application programs. Different models have been presented to guarantee the quality of service. Among these, the differentiated services can be mentioned which was presented by IETF. In the architecture of the differentiated services, no admission control mechani...

متن کامل

An Extensible Remote Graphical Interface for an ATM Network Simulator

The University of California, Santa Barbara, is developing a 40 gigabit per second ATM switch as part of its Thunder and Lightning network project. To use the high bandwidth e ciently, new network protocols are being developed and simulated on the Thunder and Lightning network protocol simulator. Due to the extreme memory and computational requirements of the simulator, the display, unlike most...

متن کامل

PANNS - A Parallelized Arti cial Neural Network Simulator

Thomas Fuerle and Erich Schikuta Institute of Applied Computer Science and Information Systems, University of Vienna, Rathausstr. 19/4, A-1010, Vienna, Austria Email: [email protected], [email protected] Abstract This paper describes the speci cation and implementation of PANNS (Parallel Arti cial Neural Network Simulator), a general purpose tool for building, running and ana...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1992