Vincent Aarts (ASML)

Quickly finding performance drivers in the high-dimensional data streams of advanced lithography systems

15:50-16:30

Extreme ultraviolet (EUV) lithography systems are helping to enable the most advanced chip manufacturing processes at major chipmakers. These systems are equipped with hundreds of sensors and actuators that generate thousands of signals to improve performance, run diagnostics and do predictive maintenance. Since such data streams are both extremely large and highly complex, conventional analytics will take too long to yield results, so we work with advanced data analytics methods.

In this talk, we’ll demonstrate an easy-to-use, scalable, Guided Diagnostics Workflow (GDW) built on advanced data analytics technology. The GDW assists domain experts in finding the root cause of performance issues at significantly reduced cycles of learning. The GDW is built to enable ‘citizen data scientists’ to perform machine learning and advanced statistics methods on these complex data sets, without requiring in-depth knowledge of data science or data engineering, all the while maintaining high flexibility of usage for different types of use cases.

Vincent Aarts studied electrical engineering and specialized in machine learning, signal processing and (nonlinear) dynamical systems engineering. During his studies, he was also an entrepreneur working on his own web company. He performed an internship in the Wearable Computing Lab at ETH Zürich in Switzerland and graduated cum laude for his master’s thesis performed at Philips Research Shanghai.

He has 8 years of experience in healthcare, performing research on the early detection of deterioration of patients using big data and leading a data science team developing contextual algorithms for a newly developed wearable biosensor. For this research, he lived for half a year in Boston, to collaborate with MIT to develop a large-scale machine learning and analytics framework using the Amazon cloud.

His goal is to transform data into knowledge, using data science. In doing so, he has over 10 patent applications. His interests are large-scale data analytics and visualization, context awareness and interpretable artificial intelligence.

He joined ASML in September 2018 as a senior data scientist and machine learning architect in the Data Science & Engineering group, where he defines and develops data science approaches to improve EUV source performance and diagnostics capabilities.