一种基于能量人工神经元模型的自生长、自组织神经网络

班晓娟; 刘浩; 徐卓然

doi:10.3724/SP.J.1004.2011.00615

一种基于能量人工神经元模型的自生长、自组织神经网络

doi: 10.3724/SP.J.1004.2011.00615

1.
北京科技大学信息工程学院北京 100083

详细信息

通讯作者:
刘浩

计量
- 文章访问数: 2511
- HTML全文浏览量: 62
- PDF下载量: 1824
- 被引次数: 0
出版历程
- 收稿日期: 2010-08-03
- 修回日期: 2010-12-27
- 刊出日期: 2011-05-20

An Energy Artificial Neuron Model Based Self-growing and Self-organizing Neural Network

1.
Information School of Engineering, University of Science and Technology Beijing, Beijing 100083

More Information

Corresponding author: LIU Hao

摘要

摘要: 本文结合近年生物学中神经科学的发展, 针对神经胶质细胞对生物神经元网络的生长提供能量支持的特性, 将神经胶质细胞的能量模型引入到人工神经元的概念模型中, 提出了能量人工神经元(Energy artificial neuron, EAN)的概念模型,并给出了数学表述. 同时,在能量人工神经元模型的基础上, 实现了一种新型自生长、自组织人工神经元网络 (EAN based self-growing and self-organizing neural network, ESGSONN), ESGSONN将神经元中的能量、网络的熵增量及样本与神经元权值的相似度的竞争作为生长的条件, 并对最优生长点中的获胜神经元进行单位步长调整. ESGSONN实现了快速生长、精确的样本数据分布密度保持、死神经元少的特性. 本文使用经典的16种动物实验(Ritter and Kohonen, 1989)验证了ESGSONN的正确性,并通过同SOFM、GCS等自组织网络的对比实验验证ESGSONN网络的特性. 最后,本文对ESGSONN在高维空间中的本质进行了讨论.
- 自组织 /
- 生长网络 /
- 能量人工神经元 /
- 无监督学习 /
- 高维空间 /
- 熵 /
- 聚类
Abstract: In this paper, we established a new artificial neuron model called EAN (energy artificial neuron) based on the energy concept from the glial cells according to the recent achievements in the neuroscience field. We suggested a way to demonstrate EAN model in mathematics. In addition, we realized a self-growing and self-organizing neural network based on the EAN model called ESGSONN (EAN based self-growing and self-organizing neural network). ESGSONN considers the energy in EAN, the entropy productions in the network and the similarity (between the sample and neurons' weights) as its conditions of growing and competitions. Its main features are described as below: rapid growing, probability density preserving and few superfluous neurons. A classical experiment of 16-kind animals (after Ritter and Kohonen, 1989) proved ESGSONN can work correctly. We showed the new features of ESGSONN by comparing it with the traditional self-organizing networks such as SOFM and GCS. Finally, we argued about the essence of this network in the high-dimensional space.
- Self-organizing /
- growing network /
- energy artificial neuron (EAN) /
- unsupervised leaning /
- high-dimensional space /
- entropy /
- clustering

HTML全文

参考文献(1)

[1]

Willshaw D J, Von D M C. How patterned neural connections can be set up by self-organization. Proceedings of the Royal Society of London, Series B, Biological Sciences, 1976, 194(1117): 431-445 [2] Kohonen T. Self-organized formation of topologically correct feature maps. Biological Cybernetics, 1982, 43(1): 59-69 [3] Kangas J A, Kohnen T K, Laaksonen J T. Variants of self-organizing maps. IEEE Transactions on Neural Networks, 1990, 1(1): 93-99 [4] Fritzke B. Growing cell structures --- a self-organizing network for unsupervised and supervised learning. Neural Networks, 1994, 7(9): 1441-1460 [5] Fritzke B. Growing grid --- a self-organizing network with constant neighborhood range and adaption strength. Neural Processing Letters, 1995, 2(5): 9-13 [6] Fritzke B. A growing neural gas network learns topologies. Advances in Neural Information Processing System. Cambridge: The MIT Press, 1995. 625-632[7] Choi D I, Park S H. Self-creating and organizing neural networks. IEEE Transactions on Neural Networks, 1994, 5(4): 561-575 [8] Martinetz T M, Schultan K J. A ``neural-gas'' network learns topologies. Artificial Neural Networks, 1991, 1: 397-402[9] Blackmore J, Miikkulainen R. Incremental grid growing: encoding high-dimensional structure into a two-dimensional feature map. In: Proceedings of the IEEE International Conference on Neural Networks. San Francisco, USA: IEEE, 1993. 450-455[10] Chow T W S, Wu S. Cell-splitting grid: a self-creating and self-organizing neural network. Neurocomputing, 2004, 57: 373-387 [11] Alahakoon D, Halgamuge S K, Srinivasan B. Dynamic self-organizing maps with controlled growth for knowledge discovery. IEEE Transactions on Neural Networks, 2000, 11(3): 601-614 [12] Shen F. An Algorithm for Incremental Unsupervised Learning and Topology Representation [Ph.D. dissertation], Tokyo Institute of Technology, Japan, 2006[13] Shen F, Hasegawa O. An incremental network for on-line unsupervised classification and topology learning. Neural Networks, 2006, 19(1): 90-106 [14] Shen F, Hasegawa O. An adaptive incremental LBG for vector quantization. Neural Networks, 2006, 19(5): 694-704 [15] Shen F, Oqura T, Hasegawa O. An enhanced self-organizing incremental neural network for online unsupervised learning. Neural Networks, 2007, 20(8): 893-903 [17] Pfrieger F W, Barres B A. Synaptic efficacy enhanced by glial cells in vitro. Science, 1997, 277(5332): 1684-1687 [18] Ullian E M, Sapperstein S K, Christopherson K S, Barres B A. Control of synapse number by glia. Science, 2001, 291(5504): 657-661 [19] Song H, Stevens C F, Gage F H. Astroglia induce neurogenesis from adult neural stem cells. Nature, 2002, 417(6884): 39-44 [20] Slezak M, Pfrieger F W, Soltys Z. Synaptic plasticity, astrocytes and morphological homeostasis. Journal of Physiology-Paris, 2006, 99(2-3): 84-91[16] Deitmer J W. Strategies for metabolic exchange between glial cells and neurons. Respiration Physiology, 2001, 129(1-2): 71-81 [21] Deitmer J W. Glial strategy for metabolic shuttling and neuronal function. BioEssays, 2000, 22(8): 747-752 3.0.CO;2-0" target=_blank>[22] Newman E A. Glial cell inhibition of neurons by release of ATP. The Journal of Neuroscience, 2003, 23(5): 1659-1666[23] Deitmer J W, Brockhaus J, Casel D. Modulation of synaptic activity in Purkinje neurons by ATP. The Cerebellum, 2006, 5(1): 49-54 [24] Bezzi P, Volterra A. A neuron-glia signaling network in the active brain. Current Opinion in Neurobiology, 2001, 11(3): 387-394 [25] McCulloch W S, Pitts W. A logical calculus of the ideas immanent in nervous activity. Bulletin of Mathematical Biology, 1943, 5(4): 115-133[26] Ritter H, Kohonen T. Self-organizing semantic maps. Biological Cybernetics, 1989, 61(4): 241-254

施引文献

资源附件(0)

访问统计