Special Session – Machine Learning in Test: A Survey of Analog, Digital, Memory, and RF Integrated Circuits

This is the authors' abstract. We don't add key points for this paper.

Integrated circuit (IC) testing presents complex problems that, when ICs become large, are exceptionally difficult to solve by traditional computing techniques. To deal with unmanageable time complexity, engineers often rely on human “hunches” and “heuristics” learned through experience. Training machines to adopt these human skills is called machine learning (ML). This survey examines applications of ML to testing analog, digital, memory, radio frequency (RF), and other application based ICs. This survey then highlights significant challenges and potential research directions.