Probing technique for neural net fault detection

The weight shifting technique for recovering faulty neural network has been proposed based on the assumption that the weights of the faulty links are known. In this paper, a technique for detecting the faulty links and determining the faulty weights in single-output two-layered feedforward neural networks is presented. This type of network architecture has been used in various applications, especially in business-related applications. The determination of faulty weights is achieved through slight neuron modification and the use of probing vectors. The technique can be implemented on a neural network chip.

[1]  Peter Alan Lee,et al.  Fault Tolerance , 1990, Dependable Computing and Fault-Tolerant Systems.

[2]  C. H. Sequin,et al.  Fault tolerance in artificial neural networks , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[3]  Amnon Yariv,et al.  A CCD based neural network integrated circuit with 64K analog programmable synapses , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[4]  Ling Rothrock.  Modeling human perceptual decision-making using an artificial neural network , 1992, [Proceedings 1992] IJCNN International Joint Conference on Neural Networks.

[5]  Chalapathy Neti,et al.  Maximally fault tolerant neural networks , 1992, IEEE Trans. Neural Networks.

[6]  Chidchanok Lursinsap,et al.  Recovering faulty self-organizing neural networks: by weight shifting technique , 1994, Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN'94).

[7]  Chidchanok Lursinsap,et al.  Weight shifting techniques for self-recovery neural networks , 1994, IEEE Trans. Neural Networks.

[8]  T. Tanigawa,et al.  Stock price pattern matching system-dynamic programming neural networks approach , 1992, [Proceedings 1992] IJCNN International Joint Conference on Neural Networks.

[9]  Jacek M. Zurada,et al.  Introduction to artificial neural systems , 1992 .