[{"data":1,"prerenderedAt":423},["ShallowReactive",2],{"content-query-i4gVFwoJAr":3},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"date":10,"cover":11,"type":12,"category":13,"body":14,"_type":417,"_id":418,"_source":419,"_file":420,"_stem":421,"_extension":422},"/technology-blogs/zh/2544","zh",false,"","论文精讲 | 基于昇思MindSpore Quantum搭建量子神经网络分类器","该论文主要讨论了在量子神经网络的不同编码结构和编码策略在解决监督学习任务时的区别。文章对幅度编码和块编码在量子神经网络处理二分类问题时的表现进行对比。通过使用Fashion MNIST、MNIST、对称性保护拓扑态作为识别训练任务，以精度为标准，对不同的拟设线路和编码策略进行基准测试。","2023-06-05","https://obs-mindspore-file.obs.cn-north-4.myhuaweicloud.com/file/2023/06/08/dbf7650c412249cba8a66a5f89b2a76f.png","technology-blogs","大V博文",{"type":15,"children":16,"toc":414},"root",[17,25,42,50,54,62,67,72,82,90,104,112,125,140,152,160,165,172,180,188,198,206,214,222,232,239,247,254,259,264,272,280,285,290,295,300,305,310,315,320,325,330,335,340,345,350,355,359,364,369,374,379,384,389,394,401,406],{"type":18,"tag":19,"props":20,"children":22},"element","h1",{"id":21},"论文精讲-基于昇思mindspore-quantum搭建量子神经网络分类器",[23],{"type":24,"value":8},"text",{"type":18,"tag":26,"props":27,"children":28},"p",{},[29,35,37],{"type":18,"tag":30,"props":31,"children":32},"strong",{},[33],{"type":24,"value":34},"分享人：闫家舜",{"type":24,"value":36}," ｜",{"type":18,"tag":30,"props":38,"children":39},{},[40],{"type":24,"value":41},"学校：浙江大学",{"type":18,"tag":26,"props":43,"children":44},{},[45],{"type":18,"tag":30,"props":46,"children":47},{},[48],{"type":24,"value":49},"内容简介",{"type":18,"tag":26,"props":51,"children":52},{},[53],{"type":24,"value":9},{"type":18,"tag":26,"props":55,"children":56},{},[57],{"type":18,"tag":30,"props":58,"children":59},{},[60],{"type":24,"value":61},"相关论文",{"type":18,"tag":26,"props":63,"children":64},{},[65],{"type":24,"value":66},"**标题：**Quantum Neural Network Classifiers: A Tutorial",{"type":18,"tag":26,"props":68,"children":69},{},[70],{"type":24,"value":71},"**作者：**Weikang Li, Zhide Lu, Dong-Ling Deng",{"type":18,"tag":26,"props":73,"children":74},{},[75,80],{"type":18,"tag":30,"props":76,"children":77},{},[78],{"type":24,"value":79},"期刊:",{"type":24,"value":81}," SciPost Physics Lecture Notes（2022年8月17日）",{"type":18,"tag":26,"props":83,"children":84},{},[85],{"type":18,"tag":30,"props":86,"children":87},{},[88],{"type":24,"value":89},"论文复现代码",{"type":18,"tag":26,"props":91,"children":92},{},[93,95],{"type":24,"value":94},"**代码链接：**",{"type":18,"tag":96,"props":97,"children":101},"a",{"href":98,"rel":99},"https://gitee.com/mindspore/mindquantum/tree/research/paper%5C_recurrence/2023/21%5C_yjshun",[100],"nofollow",[102],{"type":24,"value":103},"https://gitee.com/mindspore/mindquantum/tree/research/paper\\_recurrence/2023/21\\_yjshun",{"type":18,"tag":26,"props":105,"children":106},{},[107],{"type":18,"tag":30,"props":108,"children":109},{},[110],{"type":24,"value":111},"点击下方链接观看视频：",{"type":18,"tag":26,"props":113,"children":114},{},[115],{"type":18,"tag":30,"props":116,"children":117},{},[118],{"type":18,"tag":96,"props":119,"children":122},{"href":120,"rel":121},"http://mp.weixin.qq.com/s?__biz=MzI3NjAzMjA0NA==&mid=2247500760&idx=1&sn=bafa21041342678c9a1df456ba83d625&chksm=eb793c9bdc0eb58d76b33aec2923fe9cdf5022e8926bc8da9b3129c4627804b702aabcb0acba&scene=21#wechat_redirect",[100],[123],{"type":24,"value":124},"【直播】量子计算组会一起开 | 搭建量子神经网络分类器",{"type":18,"tag":26,"props":126,"children":127},{},[128,133,135],{"type":18,"tag":30,"props":129,"children":130},{},[131],{"type":24,"value":132},"01",{"type":24,"value":134}," ",{"type":18,"tag":30,"props":136,"children":137},{},[138],{"type":24,"value":139},"编码策略和拟设线路",{"type":18,"tag":26,"props":141,"children":142},{},[143],{"type":18,"tag":30,"props":144,"children":145},{},[146],{"type":18,"tag":147,"props":148,"children":151},"img",{"alt":149,"src":150},"image.png","https://fileserver.developer.huaweicloud.com/FileServer/getFile/cmtybbs/e64/154/b38/90a1d5d431e64154b387b3660e356ff5.20230608025343.07020802947202013641302416076424:50540607070120:2400:47CA5A45A3CF00BD02D5F14D18CFD06C7F4F880F7EBF9189C3D45B9745FBC6CC.png",[],{"type":18,"tag":26,"props":153,"children":154},{},[155],{"type":18,"tag":147,"props":156,"children":159},{"alt":157,"src":158},"cke_2764.png","https://fileserver.developer.huaweicloud.com/FileServer/getFile/cmtybbs/e64/154/b38/90a1d5d431e64154b387b3660e356ff5.20230608025420.20225064480794906674931739667793:50540607070120:2400:9BDF6BA4DCA9ED6230FC032C285D56442379C5D70E85BF2AC90D38C78BC923AC.png",[],{"type":18,"tag":26,"props":161,"children":162},{},[163],{"type":24,"value":164},"文章在相同的编码策略下，对四种不同的ansatz进行分类性能测试。ansatz中分为参数层和纠缠层两部分，四种ansatz参数层选取一致，纠缠层分别由CZ、CNOT、两层CNOT、多体系统下的时间演化构建。多体系统采用Aubry-André模型。",{"type":18,"tag":26,"props":166,"children":167},{},[168],{"type":18,"tag":147,"props":169,"children":171},{"alt":149,"src":170},"https://fileserver.developer.huaweicloud.com/FileServer/getFile/cmtybbs/e64/154/b38/90a1d5d431e64154b387b3660e356ff5.20230608025525.53569100559975533965287894290147:50540607070120:2400:8AAE0AFFEA6663906977267B8D43577DC8B08240266EB473ADAB9C421F2B04D2.png",[],{"type":18,"tag":26,"props":173,"children":174},{},[175],{"type":18,"tag":30,"props":176,"children":177},{},[178],{"type":24,"value":179},"02",{"type":18,"tag":26,"props":181,"children":182},{},[183],{"type":18,"tag":30,"props":184,"children":185},{},[186],{"type":24,"value":187},"分类任务",{"type":18,"tag":26,"props":189,"children":190},{},[191],{"type":18,"tag":30,"props":192,"children":193},{},[194],{"type":18,"tag":147,"props":195,"children":197},{"alt":149,"src":196},"https://fileserver.developer.huaweicloud.com/FileServer/getFile/cmtybbs/e64/154/b38/90a1d5d431e64154b387b3660e356ff5.20230608025636.17371028203909199249549128178769:50540607070120:2400:27D129B5535EF320FF656554EDB8544EB3A6756640B43FEA808BE103C6DFB89F.png",[],{"type":18,"tag":26,"props":199,"children":200},{},[201],{"type":18,"tag":147,"props":202,"children":205},{"alt":203,"src":204},"cke_7514.png","https://fileserver.developer.huaweicloud.com/FileServer/getFile/cmtybbs/e64/154/b38/90a1d5d431e64154b387b3660e356ff5.20230608025704.30565693268163170628589428043923:50540607070120:2400:DA91E035B41D10302EE9C7B255249BE757648AC8204E26EB1EAC7B4B7F279959.png",[],{"type":18,"tag":26,"props":207,"children":208},{},[209],{"type":18,"tag":30,"props":210,"children":211},{},[212],{"type":24,"value":213},"03",{"type":18,"tag":26,"props":215,"children":216},{},[217],{"type":18,"tag":30,"props":218,"children":219},{},[220],{"type":24,"value":221},"分类结果",{"type":18,"tag":26,"props":223,"children":224},{},[225,230],{"type":18,"tag":30,"props":226,"children":227},{},[228],{"type":24,"value":229},"幅度编码",{"type":24,"value":231},"下，不同的线路深度和ansatz对应的四种数据集的分类表现为下表，Ent1、Ent2、Ent3分别为前面提到三种纠缠线路(a)、(b)、(c)。",{"type":18,"tag":26,"props":233,"children":234},{},[235],{"type":18,"tag":147,"props":236,"children":238},{"alt":149,"src":237},"https://fileserver.developer.huaweicloud.com/FileServer/getFile/cmtybbs/e64/154/b38/90a1d5d431e64154b387b3660e356ff5.20230608025729.63113459703213698255415814625481:50540607070120:2400:48D9D3EA8ADE77AD72638E02CE23F62D1AC163D23BB53EE56CBDC1288A860910.png",[],{"type":18,"tag":26,"props":240,"children":241},{},[242],{"type":18,"tag":147,"props":243,"children":246},{"alt":244,"src":245},"cke_11341.png","https://fileserver.developer.huaweicloud.com/FileServer/getFile/cmtybbs/e64/154/b38/90a1d5d431e64154b387b3660e356ff5.20230608025805.59658809102607832156126825768027:50540607070120:2400:FE2915954F4E5D553CD505AC58FBB4E3046EBE3DCFFC4CD813CCA60A9DA69F6E.png",[],{"type":18,"tag":26,"props":248,"children":249},{},[250],{"type":18,"tag":147,"props":251,"children":253},{"alt":149,"src":252},"https://fileserver.developer.huaweicloud.com/FileServer/getFile/cmtybbs/e64/154/b38/90a1d5d431e64154b387b3660e356ff5.20230608025817.47895935273132596505547237373057:50540607070120:2400:F9FBD6C9AEC39D4C4EE0B9269015BEF59A4C5AB5F43D106D2FD1F7A64A1DFB9D.png",[],{"type":18,"tag":26,"props":255,"children":256},{},[257],{"type":24,"value":258},"（来源：论文原文）",{"type":18,"tag":26,"props":260,"children":261},{},[262],{"type":24,"value":263},"可见，不同ansatz对不同数据集的解析能力有所差异，但分类精度基本上和线路深度正相关。块编码存在最优权重因子，分类表现上略逊于幅度编码。",{"type":18,"tag":26,"props":265,"children":266},{},[267],{"type":18,"tag":30,"props":268,"children":269},{},[270],{"type":24,"value":271},"04",{"type":18,"tag":26,"props":273,"children":274},{},[275],{"type":18,"tag":30,"props":276,"children":277},{},[278],{"type":24,"value":279},"复现结果",{"type":18,"tag":26,"props":281,"children":282},{},[283],{"type":24,"value":284},"我们可以在昇思MindSpore Quantum上将上述论文结果复现并做出以下改进： 1. 测量操作选择测量末位两量子比特计算与标签的交叉熵，避免增加对训练集与测试集额外预处理操作，减小误差引入，简化代码，增加可读性； 2. 通过继承训练网络封装类TrainOneStepCell实现模型搭建和训练，增加了代码扩展性并易于测试； 3. 对整个过程建立主类进行封装，易于移植； 4. 针对基于MNIST数据集的分类问题，对参数层和纠缠层进行改进，获得更高精度。",{"type":18,"tag":26,"props":286,"children":287},{},[288],{"type":24,"value":289},"以幅度编码Block Depth=2识别Fashion MNIST及MNIST数据集为例，复现结果为：",{"type":18,"tag":26,"props":291,"children":292},{},[293],{"type":24,"value":294},"FashionMNIST Ent1",{"type":18,"tag":26,"props":296,"children":297},{},[298],{"type":24,"value":299},"FashionMNIST Ent2",{"type":18,"tag":26,"props":301,"children":302},{},[303],{"type":24,"value":304},"FashionMNIST Ent3",{"type":18,"tag":26,"props":306,"children":307},{},[308],{"type":24,"value":309},"MNIST Ent1",{"type":18,"tag":26,"props":311,"children":312},{},[313],{"type":24,"value":314},"MNIST Ent2",{"type":18,"tag":26,"props":316,"children":317},{},[318],{"type":24,"value":319},"MNIST Ent3",{"type":18,"tag":26,"props":321,"children":322},{},[323],{"type":24,"value":324},"原文结果",{"type":18,"tag":26,"props":326,"children":327},{},[328],{"type":24,"value":329},"0.994",{"type":18,"tag":26,"props":331,"children":332},{},[333],{"type":24,"value":334},"0.989",{"type":18,"tag":26,"props":336,"children":337},{},[338],{"type":24,"value":339},"0.980",{"type":18,"tag":26,"props":341,"children":342},{},[343],{"type":24,"value":344},"0.546",{"type":18,"tag":26,"props":346,"children":347},{},[348],{"type":24,"value":349},"0.831",{"type":18,"tag":26,"props":351,"children":352},{},[353],{"type":24,"value":354},"0.848",{"type":18,"tag":26,"props":356,"children":357},{},[358],{"type":24,"value":279},{"type":18,"tag":26,"props":360,"children":361},{},[362],{"type":24,"value":363},"0.996",{"type":18,"tag":26,"props":365,"children":366},{},[367],{"type":24,"value":368},"0.990",{"type":18,"tag":26,"props":370,"children":371},{},[372],{"type":24,"value":373},"0.982",{"type":18,"tag":26,"props":375,"children":376},{},[377],{"type":24,"value":378},"0.677",{"type":18,"tag":26,"props":380,"children":381},{},[382],{"type":24,"value":383},"0.844",{"type":18,"tag":26,"props":385,"children":386},{},[387],{"type":24,"value":388},"0.868",{"type":18,"tag":26,"props":390,"children":391},{},[392],{"type":24,"value":393},"分类精度规律基本与原文结论一致，改进后较原文结果有不同程度的提高。",{"type":18,"tag":26,"props":395,"children":396},{},[397],{"type":18,"tag":147,"props":398,"children":400},{"alt":149,"src":399},"https://fileserver.developer.huaweicloud.com/FileServer/getFile/cmtybbs/e64/154/b38/90a1d5d431e64154b387b3660e356ff5.20230608025913.58826524563500337350434936045357:50540607070120:2400:658F60FFCF8311B2BE2603841BC85BBE1E8F0FFB5D1309EEA883C95D4829F2C2.png",[],{"type":18,"tag":26,"props":402,"children":403},{},[404],{"type":24,"value":405},"扫码查看论文复现代码",{"type":18,"tag":26,"props":407,"children":408},{},[409],{"type":18,"tag":30,"props":410,"children":411},{},[412],{"type":24,"value":413},"欢迎各位开发者积极投稿，与大家一起分享项目成果、任务开发过程中的经验和感悟，参与者可获得专属礼品，还有机会申请昇思MindSpore优秀开发者或布道师，有兴趣的小伙伴请添加联系小助手：LLT66TT（备注“专题投稿”加速通过）。",{"title":7,"searchDepth":415,"depth":415,"links":416},4,[],"markdown","content:technology-blogs:zh:2544.md","content","technology-blogs/zh/2544.md","technology-blogs/zh/2544","md",1776506121994]