Scholarlink

This paper introduces a self-supervised deep learning framework for real-time anomaly detection in quantum sensor networks, addressing the critical challenge of maintaining quantum system integrity without labeled training data. Our approach combines temporal convolutional networks with quantum state reconstruction layers to process high-dimensional sensor data streams while preserving quantum correlations. The architecture employs contrastive predictive coding to learn robust representations from unlabeled measurements, enabling detection of decoherence events, calibration errors, and external interference patterns. Experimental results on superconducting qubit networks demonstrate 92.4% detection accuracy with 18ms latency, outperforming classical autoencoder-based methods by 14.7% in F1-score while maintaining interpretability through quantum feature visualization. The system’s lightweight design enables deployment on edge quantum controllers, addressing the latency limitations of cloud-based monitoring solutions. This work bridges quantum information processing and edge AI, offering a practical tool for fault detection in next-generation quantum technologies [1].

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Self-supervised Deep Learning for Real-Time Anomaly Detection in Quantum Sensor Networks

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Lakshmojee Koduru,
  • Monu Sharma,
  • Krishna Charan Bajanthri Chikalaguriki,
  • Anushka Raj Yadav,
  • Shubneet,
  • Navjot Singh Talwandi

摘要

This paper introduces a self-supervised deep learning framework for real-time anomaly detection in quantum sensor networks, addressing the critical challenge of maintaining quantum system integrity without labeled training data. Our approach combines temporal convolutional networks with quantum state reconstruction layers to process high-dimensional sensor data streams while preserving quantum correlations. The architecture employs contrastive predictive coding to learn robust representations from unlabeled measurements, enabling detection of decoherence events, calibration errors, and external interference patterns. Experimental results on superconducting qubit networks demonstrate 92.4% detection accuracy with 18ms latency, outperforming classical autoencoder-based methods by 14.7% in F1-score while maintaining interpretability through quantum feature visualization. The system’s lightweight design enables deployment on edge quantum controllers, addressing the latency limitations of cloud-based monitoring solutions. This work bridges quantum information processing and edge AI, offering a practical tool for fault detection in next-generation quantum technologies [1].