Digital Implementation of Neural Network by Partial Reconfiguration

Digital Implementation of Neural Network by Partial Reconfiguration

C. Udhaya Kumar, P. Saravanan, N. Thiyagarajan, Veena Raj
Copyright: © 2023 |Pages: 35
DOI: 10.4018/978-1-6684-6596-7.ch009
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Artificial Neural Networks (ANNs) are becoming increasingly important in the present technological era due to their ability to solve complex problems, adapt to new inputs, and improve decision-skills for different domains. The human brain serves as a model for Artificial Neural Networks (ANNs), a type of machine learning, as a reference for both structure and function. The existing work on ANNs supports tasks, such as regression, classification and pattern recognition separately. The discussion aims at resolving the above highlighted issues related to various ANN architectural implementations, considering the dynamic function exchange feature of FPGAs. With the aid of Zynq SOC, CNN and DNN architectures are designed in its Processing System, and the structure is accelerated using Programmable Logic. It also solves the issues due to trojans on design files, by introducing cryptography within the accelerator.
Chapter Preview
Top

Dynamic Function Exchange

Asymmetric Key Encryption techniques can also be merged to the design to increase its data security. To implement dynamic function exchange with asymmetric key encryption based on an ANN algorithm, the following several steps are followed:

Develop the ANN algorithm: The first step would be to develop an ANN algorithm that can perform the desired task. For example, if there is a need to exchange functions between two components in a design, an ANN can be developed to learn and execute these functions.

Implement public key encryption: The next step would be to implement public key encryption into the ANN algorithm. This would involve encrypt and decrypt process, generating public and private keys, and the sender’s authenticity verification.

Exchange functions dynamically: Once the ANN algorithm is implemented with asymmetric key encryption, it can be used to exchange functions dynamically between different software components. For example, if one component needs a particular function from another component, it could request it, and the other component could dynamically exchange it using the ANN algorithm. Implementing dynamic function exchange with asymmetric key encryption based on an ANN algorithm can be a complex task, but it can provide a secure and flexible way to exchange code between different software components.

Complete Chapter List

Search this Book:
Reset