Apollo-Engine

The Apollo technology emulates in general the way a simple biological  brain functions. All the relevant neuronal parameters like time coding, single spike data transmission, inhibitory connections, feedback loops  etc. are incorporated in the model.

Depending on the actual application, Apollo uses a network of significantly more than one million interconnected neurons!

In this respect, Apollo is unique on the market – it’s neither a pure  Artificial Neural Network (ANN) nor a mathematically motivated approach to Artificial Intelligence such as expert systems or  Fuzzy Logic, nor does it work like a “semantic net” also based on mathematics (e.g. Bayes- classifier)  as  they’re offered as “state of the art” on the market.

The architecture described above enables Apollo to solve such tasks eminently well that the human brain is also able to solve, in contrast to mathematical computer solutions. Examples are content recognition of arbitrary texts based on “text comprehension”, speaker-independent voice recognition, scene analysis and object recognition in complex images, and the processing of other physical signals (e.g. acoustic  and vibration analysis in condition monitoring).

Apollo is protected by various patents ; internally it is extremely complex, but externally very easy to operate, and it can be quickly and easily integrated into an existing IT environment.

Despite the high complexity of the scope of functionality, Apollo is so compact that it runs on notebooks and even on high-end PDAs.