Autonomous fault diagnosis system using learning with queries
暂无分享,去创建一个
We propose a powerful method of building a neural network fault diagnosis system that automatically collects training data (failure examples) to improve diagnosis. The learning-with-queries technique in a neural network is used to select the fault position and create training data that will improve the recognition rate of the diagnosis system. This technique is applicable to a fault diagnosis of a large-scale systems such as telecommunication switching systems.
[1] Eric B. Baum,et al. Constructing Hidden Units Using Examples and Queries , 1990, NIPS.
[2] Leslie G. Valiant,et al. A theory of the learnable , 1984, CACM.
[3] Eric B. Baum,et al. Neural net algorithms that learn in polynomial time from examples and queries , 1991, IEEE Trans. Neural Networks.