| 引用本文: | 郑瑾,高庆,吕颜轩,董道毅,潘宇.基于参数化量子电路的量子卷积神经网络模型及应用[J].控制理论与应用,2021,38(11):1772~1784.[点击复制] |
| ZHENG Jin,GAO Qing,LV Yan-xuan,DONG Dao-yi,PAN Yu.Quantum convolutional neural network and applications for parameterized quantum circuits[J].Control Theory and Technology,2021,38(11):1772~1784.[点击复制] |
|
| 基于参数化量子电路的量子卷积神经网络模型及应用 |
| Quantum convolutional neural network and applications for parameterized quantum circuits |
| 摘要点击 3282 全文点击 871 投稿时间：2021-08-24 修订日期：2021-11-19 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2021.10786 |
| 2021,38(11):1772-1784 |
| 中文关键词 量子机器学习 量子神经网络 量子卷积神经网络 量子图卷积神经网络 |
| 英文关键词 quantum machine learning quantum neural network quantum convolutional neural network quantum graph convolutional neural network |
| 基金项目 国家自然科学基金项目(61903016, 61803132, 62173296), 德国亚历山大.冯.洪堡基金项目资助. |
|
| 中文摘要 |
| 量子神经网络结合了量子计算与经典神经网络模型的各自优势, 为人工智能领域的未来发展提供了一种
全新的思路. 本文提出一种基于参数化量子电路的量子卷积神经网络模型, 能够针对欧几里得结构数据与非欧几里
得结构数据, 利用量子系统的计算优势加速经典机器学习任务. 在MNIST数据集上的数值仿真结果表明, 该模型具
有较强的学习能力和良好的泛化性能. |
| 英文摘要 |
| Quantum neural networks have provided entirely new insight into the future of artificial intelligence by combining
the advantages of quantum computing technologies and classical neural network models. In this paper, a parameterized
quantum circuit based quantum convolutional neural network model is proposed, which can deal with both Euclidean
data and non-Euclidean data and accelerate classical machine learning tasks by taking the computational advantages of
quantum systems. Simulation results on the MNIST data set show that the model has strong learning ability and good
generalization performance. |
|
|
|
|
|