Scholarlink

<p>We present next-generation HQCNN architectures, including DMERA, HEA–QCNN, and (bQCNN) designs, significantly optimized for NISQ hardware. Benchmarked on medical imaging and MNIST datasets, the bQCNN achieves 93.6% accuracy, outperforming classical baselines (88.2%) and standard QCNNs (90.1%). Quantitative analysis reveals that quantum feature maps enhance inter-class separation by 87%, while integrated EM yields a 15% improvement in noise resilience. Furthermore, the bQCNN demonstrates superior generalization, narrowing the training-validation gap to less than 0.01. These findings establish a scalable, parameter-efficient framework for high-performance hybrid quantum-classical ML on near-term devices. The code and result analysis supporting this study are available at <a href="https://github.com/qprorepo/NewQCNNCode.git">https://github.com/qprorepo/NewQCNNCode.git</a>.</p>

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

Novel Quantum-Classical Hybrid Circuits for Adaptive Hardware-Based Machine Learning

  • 期刊论文

  • 发表日期：

    2026年7月14日

查看全文
  • Shyam Sihare,
  • Aswani Kumar Cherukuri

摘要

We present next-generation HQCNN architectures, including DMERA, HEA–QCNN, and (bQCNN) designs, significantly optimized for NISQ hardware. Benchmarked on medical imaging and MNIST datasets, the bQCNN achieves 93.6% accuracy, outperforming classical baselines (88.2%) and standard QCNNs (90.1%). Quantitative analysis reveals that quantum feature maps enhance inter-class separation by 87%, while integrated EM yields a 15% improvement in noise resilience. Furthermore, the bQCNN demonstrates superior generalization, narrowing the training-validation gap to less than 0.01. These findings establish a scalable, parameter-efficient framework for high-performance hybrid quantum-classical ML on near-term devices. The code and result analysis supporting this study are available at https://github.com/qprorepo/NewQCNNCode.git.