Machine Learning in Logic Circuit Diagnosis

Our book’s finally going to be published in Feb 2023! Pre-order it now from Amazon or Barnes & Noble. It showcases the use of different machine learning (ML) techniques that have been successful in effective diagnosis of electronic circuit failures. I, along with my colleagues at Carnegie Mellon, have written a chapter on the use of ML to diagnose logic circuit failures.

Authors: Shawn Blanton, Qicheng Huang, Chenlei Fang, Soumya Mittal

Summary: ML has been used in logic-circuit diagnosis for over a decade. Many different types of ML have been deployed including, support vector machine, decision trees and decision forests, deep neural networks, k-nearest neighbors, and k-means clustering. In this chapter, we review the use of ML in diagnosis by my PhD lab, the Carnegie Mellon University Advanced Test Chip Laboratory (ACTL). The chapter is partitioned into three categories, namely, pre-diagnosis, during-diagnosis (my pioneering PhD work!), and post-diagnosis to characterize when and how a given methodology enhances the classic outcomes of diagnosis that include localization, failure behavior identification, and root cause of failure. The collective work demonstrates that ML is a valuable technology for optimizing diagnostic results and the use of diagnostics for follow-on activities. Even after a decade of using ML in diagnosis, there remains a great deal of ongoing and future application of ML in diagnosis.