Multistability of competitive neural networks with different time scales

Multistability is an important property of recurrent neural networks. It plays a crucial role in some applications, such as decision making, association memory, etc. This paper studies multistability of a class of neural networks with different time scales under the assumption that the activation functions are unsaturated piecewise linear functions. Using local inhibition to the synaptic weights of the networks, it is shown that the trajectories of the network are bounded. A global attractive set which may contain multi-equilibrium points is obtained. Complete convergence is proved by constructing an energy-like function. Simulations are employed to illustrate the theory.

[1]  Anke Meyer-Bäse,et al.  Singular Perturbation Analysis of Competitive Neural Networks with Different Time Scales , 1996, Neural Computation.

[2]  Mao Ye,et al.  Complete Convergence of Competitive Neural Networks with Different Time Scales , 2005, Neural Processing Letters.

[3]  H. Sebastian Seung,et al.  Permitted and Forbidden Sets in Symmetric Threshold-Linear Networks , 2003, Neural Computation.

[4]  Zhang Yi,et al.  Multistability Analysis for Recurrent Neural Networks with Unsaturating Piecewise Linear Transfer Functions , 2003, Neural Computation.

[5]  Shun-ichi Amari,et al.  Field theory of self-organizing neural nets , 1983, IEEE Transactions on Systems, Man, and Cybernetics.

[6]  Heiko Wersing,et al.  Dynamical Stability Conditions for Recurrent Neural Networks with Unsaturating Piecewise Linear Transfer Functions , 2001, Neural Computation.

[7]  Zhang Yi,et al.  Convergence Analysis of Recurrent Neural Networks , 2003, Network Theory and Applications.

[8]  S. Grossberg.  Competition, Decision, and Consensus , 1978 .

[9]  C. Koch,et al.  Recurrent excitation in neocortical circuits , 1995, Science.

[10]  B. V. K. Vijaya Kumar,et al.  Emulating the dynamics for a class of laterally inhibited neural networks , 1989, Neural Networks.

[11]  Richard H. R. Hahnloser,et al.  On the piecewise analysis of networks of linear threshold neurons , 1998, Neural Networks.

[12]  Xiaohui Xie,et al.  Selectively Grouping Neurons in Recurrent Networks of Lateral Inhibition , 2002, Neural Computation.

[13]  Anke Meyer-Bäse,et al.  Global exponential stability of competitive neural networks with different time scales , 2003, IEEE Trans. Neural Networks.

[14]  Richard Hans Robert Hahnloser,et al.  Digital selection and analogue amplification coexist in a cortex-inspired silicon circuit , 2000, Nature.