超维计算概念、应用及研究进展

doi:10.12305/j.issn.1001-506X.2023.07.04

系统工程与电子技术 ›› 2023, Vol. 45 ›› Issue (7): 1938-1956.doi: 10.12305/j.issn.1001-506X.2023.07.04

超维计算概念、应用及研究进展

刘文波¹^,², 姚翼荣¹^,², 张弓³^,⁴^,*, 胡文³^,⁴

1. 南京航空航天大学自动化学院, 江苏南京 211106
2. 高速载运设施的无损检测监控技术工业和信息化部重点实验室, 江苏南京 211106
3. 南京航空航天大学电子信息工程学院, 江苏南京 211106
4. 雷达成像与微波光子技术教育部重点实验室, 江苏南京 211106

收稿日期:2022-03-14 出版日期:2023-06-30 发布日期:2023-07-11
通讯作者: 张弓
作者简介:刘文波(1968—), 女, 教授, 博士, 主要研究方向为控制与测量技术、信号处理、深度学习、非线性系统控制
姚翼荣(1998—), 男, 硕士研究生, 主要研究方向为信号处理、目标检测与识别、机器学习
张弓(1964—), 男, 教授, 博士, 主要研究方向为雷达信号处理、目标识别、SAR图像处理
胡文(1979—), 男, 副教授, 博士, 主要研究方向为雷达信号处理、新体制雷达
基金资助:
国家重点研发计划(2018YFB2003304);国家自然科学基金(62271255);国家自然科学基金(62271286);国家自然科学基金(61871218);上海航天科技创新基金(SAST2021-026);南京航空航天大学前瞻布局科研专项;雷达成像与微波光子技术教育部重点实验室开放课题基金(NJ20220008);雷达成像与微波光子技术教育部重点实验室开放课题基金(NJ20210008)

Concept, application, and research progress of hyperdimensional computing

Wenbo LIU¹^,², Yirong YAO¹^,², Gong ZHANG³^,⁴^,*, Wen HU³^,⁴

1. College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China
2. Key Laboratory of Non-Destructive Testing and Monitoring Technology for High-Speed Transport Facilities, Ministry of Industry and Information Technology, Nanjing 211106, China
3. College of Electronic and Information Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China
4. Key Laboratory of Radar Imaging and Microwave Photonics, Ministry of Education, Nanjing 211106, China

Received:2022-03-14 Online:2023-06-30 Published:2023-07-11
Contact: Gong ZHANG

摘要/Abstract

摘要：

超维计算是一种受大脑工作机制启发的新兴认知模型, 使用信息的高维、随机、全息分布式表示作为处理对象, 具有低运算成本、快速学习过程、高硬件友好性、强鲁棒性、不依赖大数据和优异的模型可解释性等优势, 在分类识别、信号处理、多任务学习、信息融合、智能决策等领域有着良好的应用前景。近年来, 超维计算受到的关注量持续增加, 展现出巨大的发展潜力, 为研究人员提供了一种新选择。本文详细介绍了超维计算的发展历史、基本原理和模型框架, 给出超维计算的典型应用实例, 并对超维计算现阶段存在的问题和未来可能的发展方向进行了探讨。

关键词: 超维计算, 分类识别, 分布式表示, 机器学习, 研究进展

Abstract:

Hyperdimensional computing (HDC) is an emerging brain-inspired cognitive model. It uses high-dimensional random holographic distributed representations of different entities as processing objects and has many advantages such as low computational cost, superfast learning process, high hardware-friendliness, strong robustness, independent of data volume, and excellent model interpretability. HDC is a promising method in many fields, such as target recognition, signal processing, multi-task learning, information fusion, and smart decision-making. In recent years, HDC has received increasing attention and shown great development potential, which provides researchers with a new choice. This paper introduces the development history, basic principles, and the specific model architecture of HDC, presents some typical application examples, and discusses the existing deficiencies and possible future directions of HDC.

Key words: hyperdimensional computing (HDC), category recognition, distributed representation, machine learning, research progress

中图分类号:

TP181

刘文波, 姚翼荣, 张弓, 胡文. 超维计算概念、应用及研究进展[J]. 系统工程与电子技术, 2023, 45(7): 1938-1956.

Wenbo LIU, Yirong YAO, Gong ZHANG, Wen HU. Concept, application, and research progress of hyperdimensional computing[J]. Systems Engineering and Electronics, 2023, 45(7): 1938-1956.

图/表 17

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

图11

图12

图13

图14

图15

图16

图17

参考文献 81

1	ROSENBLATT F . The perceptron: a probabilistic model for information storage and organization in the brain[J]. Psychological Review, 1958, 65 (6): 386. doi: 10.1037/h0042519
2	RUMELHART D E , MCCLELLAND J L . Parallel distributed processing: explorations in the microstructure of cognition[M]. Cambridge: MIT Press, 1986.
3	ANDERSON J A . A simple neural network generating an inte-ractive memory[J]. Mathematical Biosciences, 1972, 14 (3/4): 197- 220.
4	KOHONEN T . Correlation matrix memories[J]. IEEE Trans.on Computers, 1972, 100 (4): 353- 359.
5	FUKUSHIMA K . Neocognitron: a self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position[J]. Biological Cybernetics, 1980, 36, 193- 202. doi: 10.1007/BF00344251
6	OLSHAUSEN B A , FIELD D J . Emergence of simple-cell receptive field properties by learning a sparse code for natural images[J]. Nature, 1996, 381 (6583): 607- 609. doi: 10.1038/381607a0
7	BELL A J , SEJNOWSKI T J . An information-maximization approach to blind separation and blind deconvolution[J]. Neural Computation, 1995, 7 (6): 1129- 1159. doi: 10.1162/neco.1995.7.6.1129
8	KRIZHEVSKY A, SUTSKEVER I, HINTON G E. Imagenet classification with deep convolutional neural networks[C]//Proc. of the Conference on Advances in Neural Information Processing Systems, 2012, 25: 1097-1105.
9	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2016: 770-778.
10	SUTSKEVER I, VINYALS O, LE Q V. Sequence to sequence learning with neural networks[C]//Proc. of the Conference on Advances in Neural Information Processing Systems, 2014: 3104-3112.
11	HINTON G E , DENG L , YU D , et al. Deep neural networks for acoustic modeling in speech recognition: the shared views of four research groups[J]. IEEE Signal Processing Magazine, 2012, 29 (6): 82- 97. doi: 10.1109/MSP.2012.2205597
12	LI F F , FERGUS R , PERONA P . One-shot learning of object categories[J]. IEEE Trans.on Pattern Analysis and Machine Intelligence, 2006, 28 (4): 594- 611. doi: 10.1109/TPAMI.2006.79
13	LIU Z, LI J G, SHEN Z Q, et al. Learning efficient convolutional networks through network slimming[C]//Proc. of the IEEE International Conference on Computer Vision, 2017: 2736-2744.
14	ZEILER M D, FERGUS R. Visualizing and understanding convolutional networks[C]//Proc. of the European Conference on Computer Vision, 2014: 818-833.
15	ZHANG Y , TINˇO P , LEONARDIS A , et al. A survey on neural network interpretability[J]. IEEE Trans.on Emerging Topics in Computational Intelligence, 2021, 5 (5): 726- 742. doi: 10.1109/TETCI.2021.3100641
16	THOMAS A , DASGUPTA S , ROSING T . Theoretical foundations of hyperdimensional computing[J]. Journal of Artificial Intelligence Research, 2021, 72, 215- 249. doi: 10.1613/jair.1.12664
17	KANERVA P . Sparse distributed memory[M]. Cambridge: MIT Press, 1988.
18	KANERVA P . Hyperdimensional computing: an introduction to computing in distributed representation with high-dimensional random vectors[J]. Cognitive Computation, 2009, 1 (2): 139- 159. doi: 10.1007/s12559-009-9009-8
19	MASSE N Y , TURNER G C , JEFFERIS G S X E . Olfactory information processing in Drosophila[J]. Current Biology, 2009, 19 (16): R700- R713. doi: 10.1016/j.cub.2009.06.026
20	CARON S J C , RUTA V , ABBOTT L F , et al. Random convergence of olfactory inputs in the Drosophila mushroom body[J]. Nature, 2013, 497 (7447): 113- 117. doi: 10.1038/nature12063
21	SQUIRE L R . Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans[J]. Psychological Review, 1992, 99 (2): 195- 231. doi: 10.1037/0033-295X.99.2.195
22	RAHIMI A, KANERVA P, RABAEY J M. A robust and energy-efficient classifier using brain-inspired hyperdimensional computing[C]//Proc. of the International Symposium on Low Power Electronics and Design, 2016: 64-69.
23	ZHANG S, WANG R, ZHANG J J, et al. Assessing robustness of hyperdimensional computing against errors in associative memory[C]//Proc. of the IEEE 32nd International Conference on Application-specific Systems, Architectures and Processors, 2021: 211-217.
24	HERSCHE M , LIPPUNER S , KORB M , et al. Near-channel classifier: symbiotic communication and classification in high-dimensional space[J]. Brain Informatics, 2021, 8 (1): 16. doi: 10.1186/s40708-021-00138-0
25	KARUNARATNE G , SCHMUCK M , LE GALLO M , et al. Robust high-dimensional memory-augmented neural networks[J]. Nature Communications, 2021, 12 (1): 2468. doi: 10.1038/s41467-021-22364-0
26	YAO Y R , LIU W B , ZHANG G , et al. Radar-based human activity recognition using hyperdimensional computing[J]. IEEE Trans.on Microwave Theory and Techniques, 2022, 70 (3): 1605- 1619. doi: 10.1109/TMTT.2021.3134992
27	GE L L , PARHI K K . Classification using hyperdimensional computing: a review[J]. IEEE Circuits and Systems Magazine, 2020, 20 (2): 30- 47. doi: 10.1109/MCAS.2020.2988388
28	HASSAN E , HALAWANI Y , MOHAMMAD B , et al. Hyper-dimensional computing challenges and opportunities for AI applications[J]. IEEE Access, 2021, 10, 97651- 97664.
29	IMANI M, KONG D, RAHIMI A, et al. VoiceHD: hyperdimensional computing for efficient speech recognition[C]//Proc. of the IEEE International Conference on Rebooting Computing, 2017.
30	JOSHI A, HALSETH J T, KANERVA P. Language geometry using random indexing[C]//Proc. of the International Symposium on Quantum Interaction, 2016: 265-274.
31	IMANI M , HWANG J , ROSING T , et al. Low-power sparse hyperdimensional encoder for language recognition[J]. IEEE Design and Test, 2017, 34 (6): 94- 101. doi: 10.1109/MDAT.2017.2740839
32	MA D N, GUO J M, JIANG Y, et al. HDTest: differential fuzz testing of brain-inspired hyperdimensional computing[C]//Proc. of the ACM/IEEE Design Automation Conference, 2021: 391-396.
33	YAO Y R , LIU W B , ZHANG G , et al. Fast SAR image recog-nition via hyperdimensional computing using monogenic mapping[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19, 4508205.
34	RAHIMI A , KANERVA P , BENINI L , et al. Efficient biosignal processing using hyperdimensional computing: network templates for combined learning and classification of ExG signals[J]. Proceedings of the IEEE, 2018, 107 (1): 123- 143.
35	RAHIMI A, BENATTI S, KANERVA P, et al. Hyperdimensional biosignal processing: a case study for EMG-based hand gesture recognition[C]//Proc. of the IEEE International Conference on Rebooting Computing, 2016.
36	BURRELLO A , SCHINDLER K , BENINI L , et al. Hyperdimensional computing with local binary patterns: one-shot learning of seizure onset and identification of ictogenic brain regions using short-time iEEG recordings[J]. IEEE Trans.on Biome-dical Engineering, 2019, 67 (2): 601- 613.
37	KIM Y, IMANI M, MOSHIRI N, et al. GenieHD: efficient DNA pattern matching accelerator using hyperdimensional computing[C]//Proc. of the Design, Automation and Test in Europe Conference and Exhibition, 2020: 115-120.
38	CHANG C Y , CHUANG Y , CHANG E J , et al. MulTa-HDC: a multi-task learning framework for hyperdimensional computing[J]. IEEE Trans.on Computers, 2021, 70 (8): 1269- 1284. doi: 10.1109/TC.2021.3073409
39	RÄSÄNEN O J , SAARINEN J P . Sequence prediction with sparse distributed hyperdimensional coding applied to the analysis of mobile phone use patterns[J]. IEEE Trans.on Neural Networks and Learning Systems, 2015, 27 (9): 1878- 1889.
40	RASANEN O , KAKOUROS S . Modeling dependencies in multiple parallel data streams with hyperdimensional computing[J]. IEEE Signal Processing Letters, 2014, 21 (7): 899- 903. doi: 10.1109/LSP.2014.2320573
41	SCHLEGEL K, MIRUS F, NEUBERT P, et al. Multivariate time series analysis for driving style classification using neural networks and hyperdimensional computing[C]//Proc. of the IEEE Intelligent Vehicles Symposium, 2021: 602-609.
42	RAHIMI A, KANERVA P, MILLAN J R, et al. Hyperdimensional computing for noninvasive brain-computer interfaces: blind and one-shot classification of EEG error-related potentials[C]//Proc. of the ACM/EAI International Conference on Bio-inspired Information and Communications Technologies, 2017: 19-26.
43	IMANI M, MORRIS J, BOSCH S, et al. AdaptHD: adaptive efficient training for brain-inspired hyperdimensional computing[C]//Proc. of the IEEE Biomedical Circuits and Systems Conference, 2019.
44	RAHIMI A , DATTA S , KLEYKO D , et al. High-dimensional computing as a nanoscalable paradigm[J]. IEEE Trans.on Circuits and Systems I: Regular Papers, 2017, 64 (9): 2508- 2521. doi: 10.1109/TCSI.2017.2705051
45	IMANI M, MESSERLY J, WU F, et al. A binary learning framework for hyperdimensional computing[C]//Proc. of the Design, Automation and Test in Europe Conference & Exhibition, 2019: 126-131.
46	SCHMUCK M , BENINI L , RAHIMI A . Hardware optimizations of dense binary hyperdimensional computing: remateria-lization of hypervectors, binarized bundling, and combinational associative memory[J]. ACM Journal on Emerging Technologies in Computing Systems, 2019, 15 (4): 32.
47	SALAMAT S , IMANI M , ROSING T . Accelerating hyperdimensional computing on FPGAs by exploiting computational reuse[J]. IEEE Trans.on Computers, 2020, 69 (8): 1159- 1171. doi: 10.1109/TC.2020.2992662
48	LI H, WU T F, RAHIMI A, et al. Hyperdimensional computing with 3D VRRAM in-memory kernels: device-architecture co-design for energy-efficient, error-resilient language recognition[C]//Proc. of the IEEE International Electron Devices Meeting, 2016.
49	WU T F, LI H, HUANG P C, et al. Brain-inspired computing exploiting carbon nanotube FETs and resistive RAM: hyperdimensional computing case study[C]//Proc. of the IEEE International Solid-State Circuits Conference, 2018: 492-494.
50	KARUNARATNE G , LE-GALLO M , CHERUBINI G , et al. In-memory hyperdimensional computing[J]. Nature Electro-nics, 2020, 3 (6): 327- 337. doi: 10.1038/s41928-020-0410-3
51	HOPFIELD J J . Neural networks and physical systems with emergent collective computational abilities[J]. Proceedings of the National Academy of Sciences, 1982, 79 (8): 2554- 2558. doi: 10.1073/pnas.79.8.2554
52	ACKLEY D H , HINTON G E , SEJNOWSKI T J . A learning algorithm for Boltzmann machines[J]. Cognitive Science, 1985, 9 (1): 147- 169. doi: 10.1207/s15516709cog0901_7
53	HINTON G E , OSINDERO S , TEH Y W . A fast learning algorithm for deep belief nets[J]. Neural Computation, 2006, 18 (7): 1527- 1554. doi: 10.1162/neco.2006.18.7.1527
54	HINTON G E , ANDERSON J A . Parallel models of associative memory: updated edition[M]. New York: Psychology Press, 2014.
55	MINSKY M , PAPERT S . Perceptrons[M]. Cambridge: MIT Press, 1969.
56	WERBOS P J . Backpropagation through time: what it does and how to do it[J]. Proceedings of the IEEE, 1990, 78 (10): 1550- 1560. doi: 10.1109/5.58337
57	RUMELHART D E , HINTON G E , WILLIAMS R J . Learning representations by back-propagating errors[J]. Nature, 1986, 323 (6088): 533- 536. doi: 10.1038/323533a0
58	LECUN Y , BOSER B , DENKER J S , et al. Backpropagation applied to handwritten zip code recognition[J]. Neural Computation, 1989, 1 (4): 541- 551. doi: 10.1162/neco.1989.1.4.541
59	HINTON G E . Mapping part-whole hierarchies into connectionist networks[J]. Artificial Intelligence, 1990, 46 (1/2): 47- 75.
60	SMOLENSKY P . Tensor product variable binding and the representation of symbolic structures in connectionist systems[J]. Artificial Intelligence, 1990, 46 (1/2): 159- 216.
61	PLATE T A . Holographic reduced representations[J]. IEEE Trans.on Neural Networks, 1995, 6 (3): 623- 641. doi: 10.1109/72.377968
62	KANERVA P. Binary spatter-coding of ordered k-tuples[C]//Proc. of the International Conference on Artificial Neural Networks, 1996: 869-873.
63	GAYLER R W. Multiplicative binding representation operators and analogy[C]//Proc. of the Workshop on Advances in Analogy Research: Integration of Theory and Data from the Cognitive Computational and Neural Sciences, 1998.
64	PLATE T A . Holographic reduced representations: distributed representation for cognitive structures[M]. Stanford: CSLI, 2003.
65	LEVY S D, GAYLER R. Vector symbolic architectures: a new building material for artificial general intelligence[C]//Proc. of the Conference on Artificial General Intelligence, 2008: 414-418.
66	IMANI M, HUANG C, KONG D, et al. Hierarchical hyperdimensional computing for energy efficient classification[C]//Proc. of the ACM/ESDA/IEEE Design Automation Confe-rence, 2018.
67	QUASTHOFF U, RICHTER M, BIEMANN C. Corpus portal for search in monolingual corpora[C]//Proc. of the International Conference on Language Resources and Evaluation, 2006: 1799-1802.
68	KOEHN P. Europarl: a parallel corpus for statistical machine translation[C]//Proc. of the Machine Translation Summit X, 2005: 79-86.
69	KLEYKO D , OSIPOV E , SENIOR A , et al. Holographic graph neuron: a bioinspired architecture for pattern processing[J]. IEEE Trans.on Neural Networks and Learning Systems, 2016, 28 (6): 1250- 1262.
70	MANABAT A X, MARCELO C R, QUINQUITO A L, et al. Performance analysis of hyperdimensional computing for cha-racter recognition[C]//Proc. of the International Symposium on Multimedia and Communication Technology, 2019.
71	IMANI M, KIM Y, WORLEY T, et al. HDCluster: an accurate clustering using brain-inspired high-dimensional computing[C]//Proc. of the Design, Automation and Test in Europe Conference & Exhibition, 2019: 1591-1594.
72	KLEYKO D , RAHIMI A , RACHKOVSKIJ D A , et al. Classification and recall with binary hyperdimensional computing: tradeoffs in choice of density and mapping characteristics[J]. IEEE Trans.on Neural Networks and Learning Systems, 2018, 29 (12): 5880- 5898. doi: 10.1109/TNNLS.2018.2814400
73	IMANI M, SALAMAT S, KHALEGHI B, et al. SparseHD: algorithm-hardware co-optimization for efficient high-dimensional computing[C]//Proc. of the IEEE 27th Annual International Symposium on Field-Programmable Custom Computing Machines, 2019: 190-198.
74	MITROKHIN A , SUTOR P , SUMMERS-STAY D , et al. Symbolic representation and learning with hyperdimensional computing[J]. Frontiers in Robotics and AI, 2020, 7, 63. doi: 10.3389/frobt.2020.00063
75	ALONSO P, SHRIDHAR K, KLEYKO D, et al. HyperEmbed: tradeoffs between resources and performance in NLP tasks with hyperdimensional computing enabled embedding of n-gram statistics[C]//Proc. of the International Joint Conference on Neural Networks, 2021.
76	KLEYKO D , KHEFFACHE M , FRADY E P , et al. Density encoding enables resource-efficient randomly connected neural networks[J]. IEEE Trans.on Neural Networks and Learning Systems, 2020, 32 (8): 3777- 3783.
77	MORRIS J, LUI H W, STEWART K, et al. HyperSpike: hyperdimensional computing for more efficient and robust spiking neural networks[C]//Proc. of the 2022 Design, Automation & Test in Europe Conference & Exhibition, 2022: 664-669.
78	KLEYKO D , FRADY E P , KHEFFACHE M , et al. Integer echo state networks: efficient reservoir computing for digital hardware[J]. IEEE Trans.on Neural Networks and Learning Systems, 2022, 33 (4): 1688- 1701. doi: 10.1109/TNNLS.2020.3043309
79	HSU P K, YU S. In-memory 3D NAND flash hyperdimensional computing engine for energy-efficient SARS-CoV-2 genome sequencing[C]//Proc. of the IEEE International Memory Workshop, 2022.
80	DUBREUIL T, AMARI P, BARRAUD S, et al. A novel 3D 1T1R RRAM architecture for memory-centric hyperdimensional computing[C]//Proc. of the IEEE International Memory Workshop, 2022.
81	DATTA S , ANTONIO R A G , ISON A R S , et al. A programmable hyper-dimensional processor architecture for human-centric IoT[J]. IEEE Journal on Emerging and Selected Topics in Circuits and Systems, 2019, 9 (3): 439- 452. doi: 10.1109/JETCAS.2019.2935464

[1]	王湖升, 陈伯孝, 叶倾知. 基于箔条干扰实测数据的对抗方法研究[J]. 系统工程与电子技术, 2023, 45(7): 2010-2021.
[2]	王春政, 胡明华, 杨磊, 赵征. 空中交通延误预测研究综述[J]. 系统工程与电子技术, 2022, 44(3): 863-874.
[3]	张心宇, 刘源, 宋佳凝. 基于LSTM神经网络的短期轨道预报[J]. 系统工程与电子技术, 2022, 44(3): 939-947.
[4]	赵凡, 金虎. 基于GAN的通信干扰波形生成技术[J]. 系统工程与电子技术, 2021, 43(4): 1080-1088.
[5]	宋波, 叶伟, 孟祥辉. 基于多智能体强化学习的动态频谱分配方法综述[J]. 系统工程与电子技术, 2021, 43(11): 3338-3351.
[6]	彭芳, 吴军, 王帅, 向建军. 基于SVRGD的机载预警雷达自适应波束形成算法[J]. 系统工程与电子技术, 2021, 43(1): 83-90.
[7]	张菁, 何友, 彭应宁, 李刚. 新一代战斗机技术特征论证[J]. 系统工程与电子技术, 2018, 40(8): 1754-1759.
[8]	颛孙少帅, 杨俊安, 刘辉, 黄科举. 基于正强化学习和正交分解的干扰策略选择算法[J]. 系统工程与电子技术, 2018, 40(3): 518-525.
[9]	王桂胜, 任清华, 姜志刚, 刘洋, 徐兵政. 基于信号特征空间的TDCS干扰分类识别[J]. 系统工程与电子技术, 2017, 39(9): 1950-1958.
[10]	王邑, 孙金标, 肖明清, 罗继勋. 基于类型2区间模糊K近邻分类器的动态武器目标分配方法研究[J]. 系统工程与电子技术, 2016, 38(6): 1314-1319.
[11]	顾斌杰, 潘丰. 改进的v-支持向量回归机的v解路径算法[J]. 系统工程与电子技术, 2016, 38(1): 205-214.
[12]	唐宁，蔡晋，李原，张开富. 基于扩散先验分布的成组技术分类识别方法[J]. Journal of Systems Engineering and Electronics, 2012, 34(4): 827-832.
[13]	王海华, 沈晓峰. 一种新的雷达辐射源信号脉内特征提取方法[J]. Journal of Systems Engineering and Electronics, 2009, 31(4): 809-811.

超维计算概念、应用及研究进展

Concept, application, and research progress of hyperdimensional computing

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 17

参考文献 81

相关文章 13

编辑推荐

Metrics

本文评价