Neural Networks for Logic Circuits

ＴｈｅｎｅｕｒａｌｎｅｔｗｏｒｋｓｆｏｒＮＯＴ，ＡＮＤ，ＯＲ，ＮＡＮＤ，ＮＯＲ，ＸＯＲａｎｄＸＮＯＲｇａｔｅｓｗｅｒｅｐｒｅｓｅｎｔｅｄｉｎ［１］ａｎｄ［４］．ＴｈｅｌｏｇｉｃｆｕｎｃｔｉｏｎｏｆａｎｙｇａｔｅｃａｎｂｅｄｅｓｃｒｉｂｅｄｕｓｉｎｇＢ...

Neural networks for logic circuits

Bushnell and the author proposed the neural networks for NOT, AND, OR, NAND, NOR, XOR and XNOR gates. Using these neural networks, the neural networks of any logic circuits can be constructd. From this, the consistent signals in the logic circuits will be transformed into the global minimal points of a quadratic pseudo Boolean function. Thus the neural network application in the field of circuit modeling and automatic test pattern generation can be widened.