The CSA approach to knowledge representation in neural networks

New generation computers are to a large extent dependent on the progress of neural network research. The biggest problem of neural networks is the lack of representational power and incompatibility with the conventional AI. We propose to analyze neural networks using the Calculus of Self-Modifiable Algorithms which is more general than neural networks. We demonstrate why neural networks can be interpreted as a subclass of self-modifiable algorithms and how they work as self-modifiable algorithms. For illustration, basic neural nets models are described in a uniform way using this new approach