Fault tolerance estimation in digital circuits with visualised generative networks
Researchers have developed a novel numerical method for estimating the fault tolerance of digital circuits. This approach utilizes a generative adversarial network (GAN) to sample bitwise configurations and compare ideal outputs with realistic signals, thereby identifying deviations caused by error modes. The method allows for the evaluation of electronic design robustness by differentiating the impact of failures associated with various logical elements. AI