An Analytical Study on Machine Learning Approaches for Simulation-Based Verification
Rajesh Koti Mourya Vangara, Bhaskar Kakani, Sandeep Vuddanti · 2021
In the modern era, there is a fast-growing demand for new and complicated digital systems. Advancements in CMOS fabrication technologies has accommodated chip fabrication of complex digital system designs. To cope with this demand efficient, fast-paced and convenient design verification methods are required. Integrating Machine Learning techniques into simulation-based verification methods can solve several problems in digital system verification for the current industrial scenario. This paper discusses the evolution of optimization algorithms from elementary genetic algorithms and Bayesian Networks to the latest Reinforcement Learning based deep neural networks and artificial neural networks. These Algorithms have enhanced Functional Verification methods by reducing human involvement and increasing their efficiency drastically.