Generalization and selection of examples in feedforward neural networks

被引:26
|
作者
Franco, L [1 ]
Cannas, SA [1 ]
机构
[1] Univ Nacl Cordoba, Fac Matemat Astron & Fis, RA-5000 Cordoba, Argentina
关键词
D O I
10.1162/089976600300014999
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In this work, we study how the selection of examples affects the learning procedure in a boolean neural network and its relationship with the complexity of the function under study and its architecture. We analyze the generalization capacity for different target functions with particular architectures through an analytical calculation of the minimum number of examples needed to obtain full generalization (i.e., zero generalization error). The analysis of the training sets associated with such parameter leads us to propose a general architecture-independent criterion for selection of training examples. The criterion was checked through numerical simulations for various particular target functions with particular architectures, as well as for random target functions in a nonoverlapping receptive field perceptron. In all cases, the selection sampling criterion lead to an improvement in the generalization capacity compared with a pure random sampling. We also show that for the parity problem, one of the most used problems for testing learning algorithms, only the use of the whole set of examples ensures global learning in a depth two architecture. We show that this difficulty can be overcome by considering a tree-structured network of depth 2 log(2) (N) - 1.
引用
收藏
页码:2405 / 2426
页数:22
相关论文
共 50 条
  • [41] A New Formulation for Feedforward Neural Networks
    Razavi, Saman
    Tolson, Bryan A.
    IEEE TRANSACTIONS ON NEURAL NETWORKS, 2011, 22 (10): : 1588 - 1598
  • [42] Global Optimization of Feedforward Neural Networks
    LIANG Xun XIA Shaowei Department of Automation
    JournalofSystemsScienceandSystemsEngineering, 1993, (03) : 273 - 280
  • [43] Injecting Chaos in Feedforward Neural Networks
    Ahmed, Sultan Uddin
    Shahjahan, Md.
    Murase, Kazuyuki
    NEURAL PROCESSING LETTERS, 2011, 34 (01) : 87 - 100
  • [44] Interpolation representation of feedforward neural networks
    Li, HX
    Li, LX
    Wang, JY
    MATHEMATICAL AND COMPUTER MODELLING, 2003, 37 (7-8) : 829 - 847
  • [45] Injecting Chaos in Feedforward Neural Networks
    Sultan Uddin Ahmed
    Md. Shahjahan
    Kazuyuki Murase
    Neural Processing Letters, 2011, 34 : 87 - 100
  • [46] A NEW MODEL OF FEEDFORWARD NEURAL NETWORKS
    WANG, DX
    TAI, JW
    PHYSICS LETTERS A, 1992, 162 (01) : 41 - 44
  • [47] Topology of Learning in Feedforward Neural Networks
    Gabella, Maxime
    IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2021, 32 (08) : 3588 - 3592
  • [48] Quantum generalisation of feedforward neural networks
    Wan, Kwok Ho
    Dahlsten, Oscar
    Kristjansson, Hler
    Gardner, Robert
    Kim, M. S.
    NPJ QUANTUM INFORMATION, 2017, 3
  • [49] On the fault tolerance of feedforward neural networks
    Huazhong Ligong Daxue Xuebao, SUPPL. 2 (22-24):
  • [50] Function evaluation with feedforward neural networks
    Logan, D
    Argyrakis, P
    INTERNATIONAL JOURNAL OF COMPUTER MATHEMATICS, 1998, 67 (1-2) : 201 - 222