AKRITOS is a software tool whose function and purpose is to model any nonrandom data with the simplest building algorithm. For this kind of software, the label Artificial Intelligence may be rightfully used, as it solves the tasks even humans sometimes can’t. It is powered by Critticall

Say you have a collection of a planet positions measured at different times! The task for AKRITOS is to deduce not only its orbit, but the governing laws of celestial mechanics at the same time.

Say you have a set of  exchange rates. A possible regularity will be spotted by AKRITOS. An evolution pressure will be applied by the control program toward the evolving the simplest algorithm which best explains the presented data and forecasts the next value.

The method for finding the simplest regression algorithms for a data set or sequence is the evolution of them, At first, a raw example is tried until a quite good approximation (or data explanation) evolves. It doesn’t stop there, even if everything perfectly fits together and the algorithm is displayed in black (not red anymore). It may always be a room for an improvement.

The power of a random change and a nonrandom selection is used, the results are very often excellent in a few hours of running. Sometimes even sooner, sometimes it runs over night to get the result or even more. This is the very nature of the (digital) evolution which is employed here. We think that the field of Artificial Intelligence has culminated now with this novel evolution approach.

The program will always look for an even simpler solution, no matter that a satisfactory one has already been found. The Occam Razor principle demands that (http://en.wikipedia.org/wiki/Occam%27s_razor).