Artificial Intelligence Applications and Reconfigurable Architectures

2Review of Artificial Intelligence Applications and Architectures

Rashmi Mahajan¹*, Dipti Sakhare¹ and Rohini Gadgil²

¹MIT Academy of Engineering, Alandi, Pune, Maharashtra, India

²Dr D Y Patil School of Engineering, Lohegaon, Pune, Maharashtra, India

Abstract

Advancements in artificial intelligence provide opportunities for smart system development in various fields. This increases the need for corresponding high-performance computing resources for the implementation. Simultaneous development in semiconductor technologies is providing sustainable implementing platforms. The aim of the present review is to give insight into the hardware architectures for AI applications. Herein, the main contributors are Application Specific Integrated Circuits (ASIC), Field Programmable Gate Array (FPGA), and General Processing Technologies (GPU). In comparison, the FPGA and GPU excel in flexibility for implementation. These technologies offer computation flexibility for the application development. Hence, the chapter presents comparison about the two flexible architectures.

The application area separates artificial intelligence with the subdomains machine learning (ML) and deep learning (DL). In the chapter, hardware platforms are analyzed with deep learning applications as well.

Keywords: AI, deep learning, FPGA, GPU

2.1 Introduction

Artificial intelligence (AI) is a versatile domain with various applications in automation, image recognition, and data processing. Furthermore, AI plays ...