ThermoPhoton: Fast 3D Thermal Simulation of Photonic Integrated Circuits via Operator Learning
Published in 2025 IEEE/ACM International Conference On Computer Aided Design (ICCAD), 2025
Abstract: Thermal crosstalk is a critical issue in Photonic Integrated Circuits (PICs). Traditional Finite Element Method (FEM) simulations are accurate but computationally expensive. In this work, we propose ThermoPhoton, a learning-based framework that utilizes operator learning to predict the 3D thermal distribution of PICs. Our method achieves millisecond-level simulation speed with high accuracy, enabling efficient thermal-aware design automation.
Recommended Citation:
```bibtex @INPROCEEDINGS{11240705, author={Guan, Weilong and Huang, Li and Lin, Yuxuan and Wu, Yuchao and Tong, Yeyu and Ma, Yuzhe}, booktitle={2025 IEEE/ACM International Conference On Computer Aided Design (ICCAD)}, title={ThermoPhoton: Fast 3D Thermal Simulation of Photonic Integrated Circuits via Operator Learning}, year={2025}, pages={1-8}, doi={10.1109/ICCAD66269.2025.11240705} }
Recommended citation: W. Guan, L. Huang, Y. Lin, Y. Wu, Y. Tong, Y. Ma. (2025). "ThermoPhoton: Fast 3D Thermal Simulation of Photonic Integrated Circuits via Operator Learning." 2025 IEEE/ACM International Conference On Computer Aided Design (ICCAD). pp. 1-8.
Download Paper