The AI & Vision team of the ASM Center of Competency is pursuing a range of applied AI/ML projects with a focus on innovation and industrial applications. In addition to the internal R&D effort we also partner with various academic and industrial organizations, hosting master and PhD students, and working with our partners on projects at the forefront of AI research.
Advanced semiconductor manufacturers increasingly face challenges in controlling the quality of modern devices. Conventional automated inspection and process control techniques are becoming less suitable for highly variable and complex micro and nanoscale assemblies. ASM Pacific Technology develops both assembly equipment and automated visual inspection tools which are designed to fit seamlessly in a smart factory. The ultimate vision is a production environment with minimal or no human supervision.
On the inspection front the use of Deep Learning offers new possibilities for detecting subtle microscale defects. Yet even such a powerful tool, in its current form, has shown serious limitations when used in real world environments. Some of the major challenges that limit deployment of DL and other ML solutions for visual inspection in the semiconductor industry include: 1) the rarity of defect samples and 2) the high variability of manufactured devices. Techniques that require much less (or no) supervision are the only long-term viable solution for high-throughput / high-mix precision industries.
In this context, we aim at developing computational methods and workflows that process the readily available visual streams in a largely unattended way. To that end our R&D efforts are focused on: self-supervised learning, continual/curriculum learning, realistic visual simulations for data augmentation, and automated machine learning.
Dr Faysal Boughorbel is responsible for the AI & Vision team. He can be reached at faysal.boughorbel@asmpt.com