Hierarchy of Information

Abstract

The idea that some essential features of the process of thinking can be simulated by neural networks is quite popular in the cognitive community; see, for example, [11], [12], [93]. As with all theories, such an approach has strong as well as weak aspects. In particular, cognitive models based on neural networks can be criticized from the positions of artificial intelligence [185] (based on the computational, Turing’s machine [184], viewpoint of thought) or quantum cognitive theories, [148], [100], [150], [175] and chapters 5,6 (based on a field-like representation of thought), or cognitive architecture (based on the necessity of a linguistic medium, [77], [155], [199]), see [185] on the extended discussion. However, it seems that, despite all the critique, neural networks are very useful for describing (or at least simulating) the process of thinking.