Lynceus, a developer of artificial intelligence (AI)-powered predictive process control, and NXP Semiconductors today presented a case study on “Qualifying AI for Automotive Production” at the Fab Owners Alliance (FOA) Q1 Collaborative Forum in Phoenix, Ariz. The presentation covered the importance of qualifying AI for production deployment, specifically for automotive-grade production.
The shift towards electric vehicles and autonomous driving means cars are increasingly packed with highly critical electronic components. Automotive chip manufacturers need to adapt by producing more reliably, while optimizing asset spending in a challenging market environment.
AI can in principle help improve yield and production capacity in existing fabs but to this date has not been adopted at full scale. The industry misses a robust qualification and introduction process, able to bridge the gap between a performant predictive model and a fully functioning AI production solution.
“To this day, AI solutions have not met the robustness and reliability standards of automotive-grade production, and the industry is missing a qualification framework to validate an AI solution that is ready for deployment in an automotive fab,” said David Meyer, CEO of Lynceus. “In collaboration with NXP, we are the first to fully qualify AI for automotive-grade production, and we have developed an actionable framework that can be repurposed for any AI application,” he added.
The presentation touched on a four-step playbook covering:
● Who needs to be involved and how
● What is a reliable model
● What supporting infrastructure is required
● How to assess associated manufacturing risks and opportunities
“Successfully deploying AI in a high-value manufacturing environment, such as a semiconductor fab, by focusing on predictive performance alone is not enough. Building a comprehensive qualification case evidencing the expected impact on manufacturing performance, the robustness of the system and the feasibility of scaling is a critical and often neglected step to fostering large scale adoption of AI,” Meyer stated.
FOA members were the first to learn of the advantages of Lynceus’ AI implementation and can now adopt the technology for each unique purpose. Members also reaped the value from this first use case and can apply learned lessons from this successful partnership.