二进神经网络中的汉明球突及其线性可分性

对于二进神经网络,剖析其神经元的逻辑意义对网络的规则提取是十分重要的,而目前每个神经元所表达的线性结构的逻辑意义仍没有完全解决,一部分线性函数的结构及其逻辑意义尚不明确.本文在寻找线性可分结构的过程中,提出了汉明球突的概念,给出其是否线性可分的判定方法,并得到二进神经元与线性可分的汉明球突等价的充要条件,从而建立了判别线性可分的汉明球突的一般方法,并通过实例验证了该方法的有效性.

Applications of Jacobi Elliptic Function Expansion Method for Nonlinear Differential-Difference Equations

The Jacobi elliptic function expansion method is extended to derive the explicit periodic wave solutions for nonlinear differential-difference equations. Three well-known examples are chosen to illustrate the application of the Jacobi elliptic function expansion method. As a result, three types of periodic wave solutions including Jacobi elliptic sine function, Jacobi elliptic cosine function and the third elliptic function solutions are obtained. It is shown that the shock wave solutions and solitary wave solutions can be obtained at their limit condition.