Memristor-based neuromorphic circuits and unconventional computing

Even if traditional solid state electronics has provided good computational systems, that are still in progress, and performances of new generations of the computers are much higher than it was expected, for example, 20 years ago, a lot of activities can be currently observed in the field of unconventional computing. There are several reasons of this activity. The main one is connected to the fact that architecture and properties of the computers are very different to the best existing intelligent system - human brain. Among differences, we can underline that in computers memory and processor are distinct systems in computers. Therefore, information plays a passive role: it can be recorded, accessed, erased, but it cannot vary the processor properties. Instead, in the brain the same elements are used both for memorizing and processing of the information. Such organization allows parallel processing and, even more important, learning - modification of connections within the processor, making it ready for solving similar problems in the future.