Robust Quantum Neural Networks Against Dynamic Noise Landscape in the NISQ Era

Shijin Duan, Gaowen Liu, Charles Fleming, Ramana Rao Kompella, Xiaolin Xu, Shaolei Ren

Keywords: Quantum Neural Network, Noise-Aware Training, Dynamic Noise

Abstract: Quantum machine learning, an emerging field in the noisy intermediate-scale quantum (NISQ) era, faces significant challenges in error mitigation during training and inference stages. Current noise-aware training (NAT) methods typically assume static error rates in quantum neural networks (QNNs), often neglecting the inherently dynamic nature of such noise. By addressing this oversight, our work recognizes the dynamics of noise in the NISQ era, evidenced by fluctuating error rates across different times and qubits. Moreover, QNN performance can vary markedly depending on the specific locations of errors, even under similar error rates. This variability underscores the limitations of static NAT strategies in addressing the dynamic nature of noisy environments. We propose a novel NAT strategy that adapts to both standard and fatal error conditions, cooperating with a low-complexity search strategy to efficiently locate fatal errors during optimization. Our approach marks a significant advancement over current NAT methods by maintaining robust performance in fatal error scenarios. Evaluations validate the efficacy of our strategy against fatal errors, while maintaining performance comparable to state-of-the-art NAT approaches under various error rates.

Primary Area: other topics in machine learning (i.e., none of the above)

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 7194