This is a project I collaborated on with my teammates @Mo Mo and @Jiaheng Tang for the 8th China College IC Competition - HuaDaJiuTian Cup in 2024. The project focuses on parametric structure capacitance extraction using machine learning techniques.
We designed a neural network that is both highly generalizable and relatively lightweight, based on the geometric and physical characteristics of capacitance. Despite there being several areas for improvement, such as the adoption of NAS (Neural Architecture Search), this project serves as a standard example of utilizing deep learning to solve practical problems.
Although we only received the second prize in the Shanghai regional competition, we were very close to advancing to the national finals. We have made this project public primarily to commemorate it as our first complete end-to-end project during our undergraduate studies.
Feel free to explore the code and documentation in this repository. If you have any questions or suggestions, please don't hesitate to reach out.
Special thanks to @Prof. Zhaori Bi and the organizers of the competition for providing this valuable learning opportunity.