O USO DE INTELIGÊNCIA ARTIFICIAL NA OTIMIZAÇÃO DE ARQUITETURAS DE HARDWARE: melhorando o processamento, reduzindo custos e ampliando o acesso a tecnologias de processamento neural

Abstract

This study investigates the use of artificial intelligence (AI) techniques to optimize hardware architectures, focusing on processors with dedicated neural cores. The main objective is to understand how AI methods such as machine learning, genetic algorithms, and reinforcement learning can contribute to improving computational performance, energy efficiency, and reducing manufacturing costs of these processors. The research adopts a qualitative, exploratory, and descriptive approach, based on a systematic literature review and a comparative analysis of performance and efficiency benchmarks from different hardware platforms. Case studies demonstrating the practical application of these techniques in the design and manufacturing of neural processors are also considered. The results indicate that the use of AI in hardware development accelerates design cycles, optimizes resource utilization, and reduces failures, contributing to the development of more accessible and sustainable solutions. Furthermore, integrating AI into hardware design promotes the democratization of access to high-performance technologies, positively impacting sectors such as healthcare, industry, security, education, and the digital economy. It is concluded that the application of these techniques represents a promising path to overcoming the cost and complexity challenges of neural processors, strengthening their presence in mobile devices, embedded systems, and corporate environments.