量子谱回归算法

潘世杰; 高飞; 万林春; 秦素娟; 温巧燕

doi:10.7544/issn1000-1239.2021.20210366

量子谱回归算法

¹(网络与交换技术国家重点实验室(北京邮电大学) 北京 100876)
²(密码科学技术国家重点实验室北京 100878) (panshijie@bupt.edu.cn)

基金项目: 中央高校基本科研业务费专项资金(2019XD-A01);国家自然科学基金项目(61976024,61972048)

详细信息

中图分类号: O413
计量
- 文章访问数: 616
- HTML全文浏览量: 4
- PDF下载量: 247
出版历程
- 发布日期: 2021-08-31

Quantum Algorithm for Spectral Regression

¹(State Key Laboratory of Networking and Switching Technology (Beijing University of Posts and Telecommunications), Beijing 100876)
²(State Key Laboratory of Cryptology, Beijing 100878)

Funds: This work was supported by the Fundamental Research Funds for the Central Universities (2019XD-A01) and the National Natural Science Foundation of China (61976024, 61972048).

摘要

摘要: 子空间学习是机器学习领域的重要研究方向.为了降低子空间学习的复杂度，Cai等人提出了谱回归降维框架，并针对结合标签构造对应图的子空间学习提出了高效谱回归.近年来，量子计算的发展使进一步降低子空间学习算法的复杂度成为了可能.Meng等人率先提出了量子谱回归算法(MYXZ算法).MYXZ算法用了稀疏哈密顿量模拟技术来处理由权重矩阵生成的矩阵，但这个矩阵在较多的情况下是稠密矩阵.针对这种情况，指出了MYXZ算法的局限性，提出了一个改进的量子谱回归算法.改进算法采用了量子奇异值估计技术，在处理稠密矩阵时相对MYXZ算法有多项式加速.另外，提出了一个新的量子算法，对经典的高效谱回归进行加速.新算法能处理的这类问题是MYXZ算法无法处理的.新算法利用了量子岭回归和量子矩阵向量乘技术，在相同的参数条件下相对经典算法具有多项式加速效果.
- 量子算法 /
- 量子机器学习 /
- 谱回归 /
- 子空间学习 /
- 稠密矩阵
Abstract: Subspace learning is an important research direction in the field of machine learning. In order to reduce the complexity of subspace learning, Cai et al. proposed a spectral regression dimensionality reduction framework, and proposed efficient spectral regression for subspace learning that constructed their graph by incorporating the label information. In recent years, the development of quantum computing has made it possible to further reduce the complexity of subspace learning algorithms. Meng et al. pioneered the quantum algorithm for spectral regression (MYXZ algorithm). In this article, the limitations of the MYXZ algorithm are pointed out. We point out that MYXZ algorithm uses sparse Hamiltonian simulation technology to process the matrix generated by the weight matrix, but this matrix is a dense matrix in some cases. In response to this situation, we propose an improved quantum spectral regression algorithm. Our improved algorithm uses quantum singular value estimation technology, which has a polynomial acceleration compared with MYXZ algorithm when dealing with dense matrices. In addition, we propose a new quantum algorithm to accelerate the classical efficient spectral regression. The problems that our new algorithm can handle cannot be handled by MYXZ algorithm. Our algorithm utilizes quantum ridge regression and quantum matrix vector multiplication technology. Under the same parameter conditions, our algorithm has polynomial acceleration effect compared with the classical algorithm.
- quantum algorithm /
- quantum machine learning /
- spectral regression /
- subspace learning /
- dense matrix

HTML全文

参考文献(0)

施引文献(15)

期刊类型引用(6)

1.	徐雪峰，郭广伟，黄余. 改进全卷积神经网络的遥感图像小目标检测. 机械设计与制造. 2024(10): 38-42 . 百度学术
2.	刘雯雯，汪皖燕，程树林. 融合项目热门惩罚因子改进协同过滤推荐方法. 计算机技术与发展. 2023(03): 15-19 . 百度学术
3.	冯勇，刘洋，王嵘冰，徐红艳，张永刚. 面向用户需求的生成对抗网络多样性推荐方法. 小型微型计算机系统. 2023(06): 1192-1197 . 百度学术
4.	冯晨娇，宋鹏，张凯涵，梁吉业. 融合社交网络信息的长尾推荐方法. 模式识别与人工智能. 2022(01): 26-36 . 百度学术
5.	韩迪，陈怡君，廖凯，林坤玲. 推荐系统中的准确性、新颖性和多样性的有效耦合与应用. 南京大学学报(自然科学). 2022(04): 604-614 . 百度学术
6.	甘亚男，耿生玲，郝立. 超贝叶斯图模型及其联结树的构建. 青海师范大学学报(自然科学版). 2021(02): 42-48 . 百度学术