Zoom: https://oregonstate.zoom.us/s/98357211915
Lithography, the process of transferring chip designs onto silicon wafers, is a critical stage in semiconductor manufacturing. Computational lithography models the lithography system's behavior, focusing on resolution enhancement in response to the aggressive scaling dictated by Moore’s Law. This field encompasses two core challenges: system modeling and photomask optimization, both of which demand high-quality results and face immense computational overhead in terms of time and memory. In this talk, we will explore the key challenges in modern computational lithography and how AI is reshaping its landscape. Topics include: 1) AI-accelerated lithography simulations, 2) AI-assisted inverse design, 3) LLM-powered agents for lithography recipe development, and 4) future directions in AI semiconductor manufacturing and research.
Haoyu Yang is a Senior Research Scientist at NVIDIA Research, specializing in AI and large language models (LLM) for Electronic Design Automation (EDA), GPU-accelerated EDA, AI-driven MultiPhysics simulation, and AI applications in semiconductor manufacturing. Before joining NVIDIA, he was a Postdoctoral Fellow in the Department of Computer Science and Engineering at the Chinese University of Hong Kong, where he also earned his Ph.D. His accolades include the Best Paper Award from IEEE Transactions on Semiconductor Manufacturing (TSM) in 2022 and a Best Paper Nomination at IEEE ASPDAC in 2019. He was also the sole recipient of the prestigious 2019 Nick Cobb Memorial Scholarship, awarded by SPIE & Mentor Graphics for his groundbreaking work in AI for lithography.