A Learning Algorithm for Fault Tolerant Feedforward Neural Networks (Special Issue on Fault-Tolerant Computing)

スポンサーリンク

概要

• 論文の詳細を見る

• A new learning algorithm is proposed to enhance fault tolerance ability of the feedforward neural networks. The algorithm focuses on the links (weights) that may cause errors at the output when they are open faults. The relevances of the synaptic weights to the output error (i.e. the sensitivity of the output error to the weight fault) are estimated in each training cycle of the standard backpropagation using the Taylor expansion of the output around fault-free weights. Then the weight giving the maximum relevance is decreased. The approach taken by the algorithm described in this paper is to prevent the weights from having large relevances. The simulation results indicate that the network trained with the proposed algorithm do have significantly better fault tolerance than the network trained with the standard backpropagation algorithm. The simulation results show that the fault tolerance and the generalization abilities are improved.

• 社団法人電子情報通信学会の論文
• 1997-01-25

著者

• Hammadi Nait

  Graduate School Of Science And Technology Chiba University

• ITO Hideo

  Faculty of Engineering, Chiba University

• Ito Hideo

  Faculty Of Engineering Chiba University

関連論文

• X-Tolerant Test Data Compression for SOC with Enhanced Diagnosis Capability(Dependable Computing)
• A Learning Algorithm for Fault Tolerant Feedforward Neural Networks
• A Learning Algorithm for Fault Tolerant Feedforward Neural Networks (Special Issue on Fault-Tolerant Computing)
• Output Smoothing: A learning Algorithm for Fault Tolerant Neural Networks
• Hybrid Pattern BIST for Low-Cost Core Testing Using Embedded FPGA Core(Dependable Computing)
• Scan Design for Two-Pattern Test without Extra Latches(Dependable Computing)
• Fault-Tolerant Routing Algorithms for Hypercube Interconnection Networks
• FPGAs Complete Fault Diagnosis Based on Binary Tree BIST Method
• FPGAs Complete Fault Diagnosis Based on Binary Tree BIST Method
• Escape and Restoration Routing : Suspensive Deadlock Recovery in Interconnection Networks
• Low-Cost IP Core Test Using Tri-Template-Based Codes
• On the Activation Function and Fault Tolerance in Feedforward Neural Networks
• Deterministic Delay Fault BIST Using Adjacency Test Pattern Generation(Dependable Computing)
• Proposal of Testable Multi-Context FPGA Architecture(Dependable Computing)
• Redundant Design for Wallace Multiplier(Dependable Computing)

もっと見る 閉じる

スポンサーリンク