Problem of Efficient Initialization of Large Self-Organizing Maps Implemented in the CMOS Technology

被引:0
作者
Kolasa, Marta [1 ]
Dlugosz, Rafal [1 ]
Pedrycz, Witold [2 ]
机构
[1] UTP Univ Sci & Technol, Fac Telecommun Comp Sci & Elect Engn, Ul Kaliskiego 7, PL-85796 Bydgoszcz, Poland
[2] Univ Alberta, Dept Elect & Comp Engn, ECERF, Edmonton, AB T6G 2V4, Canada
来源
2015 IEEE 2ND INTERNATIONAL CONFERENCE ON CYBERNETICS (CYBCONF) | 2015年
关键词
Self-Organizing Maps; Initialization of Neuron Weights; CMOS implementation; full-custom ASIC; WEIGHT INITIALIZATION;
D O I
暂无
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Initialization of neuron weights is one of key problems in artificial neural networks (ANNs). This problem is particularly important in ANNs implemented as Application Specific Integrated Circuits (ASICs), where the number of the weights becomes large. When ANNs are implemented in software, the weights can be easily programmed. In contrast, in parallel systems of this type realized as ASICs it is necessary to provide programming and addressing lines to each weight that causes a large increase in the complexity of such designs. In this paper we present investigations that demonstrate that Self-Organizing Maps (SOMs) in many situations may be trained without the initialization (with zeroed weights). We present example results of several thousands simulations for different topologies of the SOM, for different neighborhood functions and two distance measures between the learning patterns and particular neurons in the input data space. Simulations were performed for zero initial values, for small values (up to 1 % of full scale range) and for neurons randomly distributed over the overall input data space. The results are comparable that allows to reduce the complexity of the SOM implemented in the CMOS technology.
引用
收藏
页码:36 / 41
页数:6
相关论文
共 50 条
  • [41] Integer Self-Organizing Maps for Digital Hardware
    Kleyko, Denis
    Osipov, Evgeny
    De Silva, Daswin
    Wiklund, Urban
    Alahakoon, Damminda
    2019 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), 2019,
  • [42] Fusion of Self-Organizing Maps with Different Sizes
    Pasa, Leandro Antonio
    Costa, Jose Alfredo F.
    de Medeiros, Marcial Guerra
    INTELLIGENT DATA ENGINEERING AND AUTOMATED LEARNING - IDEAL 2015, 2015, 9375 : 77 - 86
  • [43] Cartograms, Self-Organizing Maps, and Magnification Control
    Henriques, Roberto
    Bacao, Fernando
    Lobo, Victor
    ADVANCES IN SELF-ORGANIZING MAPS, PROCEEDINGS, 2009, 5629 : 89 - 97
  • [44] Self-Organizing Maps for Fraud Profiling in Leasing
    Bach, Mirjana Pejic
    Vlahovic, Nikola
    Pivar, Jasmina
    2018 41ST INTERNATIONAL CONVENTION ON INFORMATION AND COMMUNICATION TECHNOLOGY, ELECTRONICS AND MICROELECTRONICS (MIPRO), 2018, : 1203 - 1208
  • [45] Self-organizing maps in adaptive health monitoring
    Tamminen, S
    Pirttikangas, S
    Röning, J
    IJCNN 2000: PROCEEDINGS OF THE IEEE-INNS-ENNS INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOL IV, 2000, : 259 - 264
  • [46] Analyzing financial performance with self-organizing maps
    Back, B
    Sere, K
    Vanharanta, H
    IEEE WORLD CONGRESS ON COMPUTATIONAL INTELLIGENCE, 1998, : 266 - 270
  • [47] Additive Composition of Supervised Self-Organizing Maps
    Jean-Luc Buessler
    Jean-Philippe Urban
    Julien Gresser
    Neural Processing Letters, 2002, 15 : 9 - 20
  • [48] Self-organizing maps for the skeletonization of sparse shapes
    Singh, R
    Cherkassky, V
    Papanikolopoulos, N
    IEEE TRANSACTIONS ON NEURAL NETWORKS, 2000, 11 (01): : 241 - 248
  • [49] Characterization of Suicidal Behaviour with Self-Organizing Maps
    Leiva-Murillo, Jose M.
    Lopez-Castroman, Jorge
    Baca-Garcia, Enrique
    COMPUTATIONAL AND MATHEMATICAL METHODS IN MEDICINE, 2013, 2013
  • [50] Additive composition of supervised self-organizing maps
    Buessler, JL
    Urban, JP
    Gresser, J
    NEURAL PROCESSING LETTERS, 2002, 15 (01) : 9 - 20