Neural networks (NNs) are applied in more and more application fields, due both to their increasing popularity in the scientific community and to the availability of suitable computing units [1-3]. In practical applications, where the training data set is large, execution times become often unacceptably long on serial machines. For this reason several projects have been developed concerning architectures devoted to simulating NNs. The project we are illustrating in this paper belongs to this category, since it aims to implement on a Field Programmable Gate Array (FPGA) a neural processor, based on the Totem NC3003 processor architecture.
Implementing a neural processor on programmable logic device
BERA, MARCO;DANESE, GIOVANNI;LEPORATI, FRANCESCO;SPELGATTI, ALVARO;
2003-01-01
Abstract
Neural networks (NNs) are applied in more and more application fields, due both to their increasing popularity in the scientific community and to the availability of suitable computing units [1-3]. In practical applications, where the training data set is large, execution times become often unacceptably long on serial machines. For this reason several projects have been developed concerning architectures devoted to simulating NNs. The project we are illustrating in this paper belongs to this category, since it aims to implement on a Field Programmable Gate Array (FPGA) a neural processor, based on the Totem NC3003 processor architecture.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
