Abstract

AI Is Here and It’s Your Friend
Deep learning neural networks that can master the game of Go, artificial intelligence chatbots that engage in natural language conversations, machine learning tools that can accurately diagnose cancer - tools that once lived only in the realm of science fiction are now being applied to real-world applications. And the semiconductor industry, with its advanced process nodes, increasingly complex geometries and shorter time to market windows, is primed to take advantage of this emerging technology. Analytics have always played a key role in semiconductor design and manufacturing. Detecting and mitigating faults at every stage of the manufacturing process drives the entire economic model. However, historically, these faults required human-intensive failure analysis techniques, which could take weeks or even months for root cause discovery and subsequent corrective actions. With the emergence of inference servers at the edge, and deep learning models that can analyze data and implement fixes in real-time, semiconductor design, manufacturing and test is benefitting with faster time to market and higher yields. This presentation will discuss applying machine learning, artificial intelligence and deep learning models throughout the semiconductor ecosystem to realize efficiencies not previously possible, ultimately ensuring that the pace of innovation continues to accelerate.