In this paper, based on the statistical approach , the fault-tolerant behavior of feedback neural networks (FNN) with threshold neurons is discussed in detail. Firstly, the model of faults for FNN are defined. Secondly, the effect of link faults and error feedback on the fault-tolerant behavior are analyzed and corresponding formulas are given. Finally, the computer simulation shows the correctness of conclusions presented in this paper.
Index Terms:
Feedback Neural Networks, Fault-tolerant Analysis, Central Limit Theorem
Citation:
Tao Zhang, Dongcheng Hu, Shiyuan Yang, "Fault-Tolerant Analysis of Feedback Neural Networks with Threshold Neurons," ats, pp.209, Eighth Asian Test Symposium (ATS'99), 1999