News center

[AI-Powered Logic Circuit Redesign Competes with Leading Global Universities and Companies, Demonstrating Global Semiconductor Design Capabilities]

 The research team from Professor Seokhyeong Kang’s laboratory (CAD & SoC Design Laboratory, CSDL) in the Department of Electrical Engineering and Semiconductor Engineering at POSTECH (Pohang University of Science and Technology) achieved 2nd place overall at the "MLCAD 2025 (7th ACM/IEEE International Symposium on Machine Learning for Computer-Aided Design)" contest co-hosted by Arizona State University (ASU) and NVIDIA.

 The MLCAD symposium held this year in Santa Cruz, California, is an international academic event that shares research achievements in AI-powered semiconductor and electronic design automation. This contest was held as a satellite event of "MLCAD 2025."

 The contest was conducted under the theme "ReSynthAI: Physical-Aware Logic Resynthesis for Timing Optimization Using AI," focusing on redesigning logic circuits and improving performance using AI. Global companies including Samsung, NXP, and IMEC, along with 27 universities and companies worldwide including UT Austin, Texas A&M, UC Davis, CUHK, and Fudan University participated.

 The POSTECH CSDL research team consisted of nine members: Professor Seokhyeong Kang, MS-Ph.D students Kyungjun Min, Mungyu Choi, and Jeonghyun Choi, and MS. students Howoo Jang, Jeyeong Park, and Junhwan Jang. In this competition, the team achieved 1st place in 5 out of 12 benchmark circuit categories, scoring 713 points and securing 2nd place overall. They received an NVIDIA RTX 5090 GPU as a prize. The gap with 1st place Fudan University (764 points) was only 51 points, while Dutch semiconductor company NXP ranked 3rd with 482 points.

 This achievement demonstrates POSTECH's world-class competitiveness in semiconductor design automation and is expected to further elevate Korea's technological standing in AI-based system semiconductor design. Professor Seokhyeong Kang said, "This award is the result of our laboratory's continuous efforts and students' creative challenges. We will continue to do our best to maintain world-class competitiveness in AI-based semiconductor design."