NeuroStrategy FAQ
This site is best viewed in a browser that conforms to web standards.
A strategy model is based on an artificial neural mesh (ANM). This intelligent software system is capable of receiving stimuli (data) via receptors and of processing them in a complex way in order to develop certain types of reactions. The manner of processing and the reactions are subject to a learning process that is partly controlled from outside and partly self-organizing.
Through training, such software components are able to develop characteristics
of artificial intelligence. Thus, for example, meshes
are capable of forming long-term and short-term memory. They can learn to
establish connections, to abstract and to generalize. While it is difficult
for human beings to think in terms of more than the usual three dimensions
of space, meshes normally act in highly
complex data spaces with a theoretically arbitrary number of dimensions.
Although they only have a minimal capacity in comparison to the human brain,
they are nevertheless able to solve problems, which our brain is unable or
barely able to manage.
Meshes can be encoded in a type of blueprint similar to the genome of a living organism. They can form populations and associations, unite as symbionts in order to solve a problem in cooperation as well as generate offspring, whose genome then consists of components of the parent genomes. Thus, they pass on part of their abilities and part of their knowledge to their descendants.
On the basis of these characteristics, it is possible to let a group or individual specimen of highly qualified meshes emerge in a process that is modelled on the evolution of life. These can then be used by traditional software to solve problems, which cannot be solved, or even represented with sufficient precision, by means of ordinary programming techniques.
Such tasks include, for example, the analysis of repositories of complex data and the learning of strategies for dealing with this data.
Accordingly, NeuroStrategy merely furnishes data feeds as well as the required environment of learning and evolution, within which the meshes are able to develop. The system then uses the artificial intelligences thus "cultivated" in order to learn how to trade effectively in financial instruments.
We assume that every security contains characteristic features in its development that differentiate it from other securities of the same type. Hence, NeuroStrategy examines every security individually and attempts to "cultivate" meshes, which, over the course of the process of learning and evolution, form a strategy for trading successfully with this particular security. The goal is to obtain an intelligent software component (the strategy model), which can subsequently be relied upon for issuing trading recommendations on a daily basis.
Whether such a component can actually be developed and how well it will work, of course, depends on how much information the mesh is able to extract from the data and whether the development of the security exhibits statistically verifiable characteristic features. In other words, there is no guarantee that NeuroStrategy will be able to develop a robust trading strategy for a particular security. If there is a possibility of developing intelligent trading behavior on the basis of the available data, however, NeuroStrategy is very likely to succeed in its task.
