July 2022 – Aug. 2022, MIT
Hanrui Wang, Pengyu Liu, Jinglei Cheng, Zhiding Liang, Jiaqi Gu, Zirui Li, Yongshan Ding, Weiwen Jiang, Yiyu Shi, Xuehai Qian, David Z. Pan, Frederic T. Chong, Song Han
- ICCAD2022 Accepted.
- Proposed QuEst, which leverages graph transformers to predict noise impact on circuit fidelity on real machines inspired by the natural graph representation of quantum circuits.
- Compared with circuit simulators, we showed the graph transformers have over 200× speedup for estimating the circuit fidelity with an RMSE error lower than 0.04 which outperforms a simple neural network-based model by 0.02 on average.