Battle Against Fluctuating Quantum Noise: Compression-Aided Framework to Enable Robust Quantum Neural Network
DescriptionNoise has become a non-negligible issue in quantum computing. There are plenty of works mitigating the effect of noise. However, the fluctuating noise is still a challenge that will attack the reproducibility of performance. In this paper, we propose a 2-stage framework to resolve the issue. At the offline stage, We generate centroids by K-means-based clustering. We also use noise-aware compression to generate robust models as the initial model into a growable model repository. At the online stage, our framework automatically generated the approximate optimal model and reduced the overhead of training. Experiments show our effectiveness among baselines.
Event Type
Research Manuscript
TimeTuesday, July 11th10:40am - 10:55am PDT
Location3003, 3rd Floor
Quantum Computing