布尔神经网络的一种高效率学习算法

A Efficient Algorithm for Boolean Neural Network

把二进神经网络学习算法推广至一般情形，利用汉明球及立方体的空间覆盖生成隐层神经元并对空间集合的相交、汉明球与低维空间的笛卡尔积在神经网络中的表现形式进行了深入探讨，得出几个旨在提高学习效率和减少布尔函数实现复杂性的有用结论，并融合形成完整的学习算法。

The learning algorithm of the binary neural network is extended to general case on which hidden neurons are generated by introducing space cover of Hamming spheres and hyper cubes.Several problems such as intersection of space set,Cartesian product of Hamming sphere and low dimension space are deeply discussed,thus getting some useful results which are aiming to decrease complexity of implementing Boolean function and improve the learning efficiency.Finally these strategies are combined into the learning algorithm.