Machine Learning Approaches for IC Fault Localization and Diagnosis

The conducted researchprovide a machine learning-driven method for thermal testing of integrated circuits. This method involves comparing the current temperature distribution with a reference distribution that is suitable for the energy state of the system. Furthermore, the research explores a methodology for positioning temperature sensors. This placement approach is advised for utilization in the identification and troubleshooting of issues with integrated circuits (ICs). Our failure localization method utilizes the average temperatures in several sub- areas of the integrated circuit, together with the interconnections between those temperatures. This research provides the simulation results for the sensor placement technique, fault detection, and localization. The results on the productivity of the testing approach are presented utilizing statistical methodologies.