Accelerate Functional Coverage Closure Using Co-simulation and Machine Learning

Explore innovative techniques for accelerating functional coverage closure in complex design verification through a conference talk that delves into co-simulation and machine learning approaches. Learn about the limitations of constrained-random verification and discover how machine learning-based solutions can enhance test generation processes. Gain insights into using MATLAB for co-simulation and machine learning in random verification environments. Understand the basic concepts and flow of co-simulation-based verification, and explore the challenges of test selection in constrained randomized testing. Examine an autoencoder-based novel test selection system and its application to a channel estimation block in a 5G radio receiver, demonstrating up to 2x reduction in simulated tests compared to traditional methods. Benefit from the expertise of two FPGA engineers working on cutting-edge projects in machine learning and 5G telecommunications systems at Nokia Kraków Research & Development.

Accelerate Functional Coverage Closure Using... - Robert Synoczek, Szymon Madej - code::dive 2023