Add to Chrome
Add to Firefox
Add to EdgeFlexible Error Mitigation of Quantum Processes with Data Augmentation Empowered Neural Model
Nov 03, 2023
Neural networks have shown their effectiveness in various tasks in the realm of quantum computing. However, their application in quantum error mitigation, a crucial step towards realizing practical quantum advancements, has been restricted by reliance on noise-free statistics. To tackle this critical challenge, we propose a data augmentation empowered neural model for error mitigation (DAEM). Our model does not require any prior knowledge about the specific noise type and measurement settings and can estimate noise-free statistics solely from the noisy measurement results of the target quantum process, rendering it highly suitable for practical implementation. In numerical experiments, we show the model's superior performance in mitigating various types of noise, including Markovian noise and Non-Markovian noise, compared with previous error mitigation methods. We further demonstrate its versatility by employing the model to mitigate errors in diverse types of quantum processes, including those involving large-scale quantum systems and continuous-variable quantum states. This powerful data augmentation-empowered neural model for error mitigation establishes a solid foundation for realizing more reliable and robust quantum technologies in practical applications.
