Our recent work, led by Ph.D. student, Han Zhou, to optimize the on-chip power grid networks considering the electromigration effects using machine learning based approach has been accepted by TCAD. This work represents our new approach to addressing  the challenging  EM failure and aging issues in the nanometer VLSI chip design. Our study shows that by leveragig the generative neural networks, we can significantly speedup the sensitivity based optimization processes by using the automatic differential feature of deep neural networks. The paper information is shown below:

H. Zhou, Y. Liu, W. Jin, and S. X.-D. Tan, “GridNetOpt: Fast full-chip EM-aware power grid optimization accelerated by deep neural networks}”, IEEE Transaction on Computer-Aided Design of Integrated Circuits and Systems (TCAD), accepted