Machine Learning Applications and Opportunities in IC Design Flow

摘要

Due to the increasing complexity of semiconductor industry, the traditional rule-based technology has faced its limits in solving Electronic Design Automation (EDA) problems, which have high dimensionality, discontinuous, and non-linearities. Instead, machine learning (ML) has been in rapidly growing demand, and has been gradually involved in EDA applications [1]. This talk shares several industrial applications of ML, including commercial EDA tools, and the great opportunity to expand the use of ML solution for design flow optimization as well.