| Forecast 
Daily Market Commentary 
Free Password forlonger term projections
 
Longer term projections Investing
Philosophy
 Method
 
Artificial Intelligence Risk
 Bio
 Pension
Funds andEndowments
 
Our Investment Programs
 Contact
 
 
 | Artificial
            
            Intelligence: 
              
              
                
                  | The origins of
                    "artificial intelligence" programming sought to
                    replicate the way humans think and solve problems. 
                      
                        The
                        first program that received recognition was written by
                        Herbert Simon and Allen Newell in 1955 it was called The
                        Logic Theorist. 
                        Research
                        continued in this area and programming languages such as
                        LISP and PROLOGUE were born.
                        Expert
                        systems were developed that stored conditional rules and
                        information that could guide the user to an answer by
                        asking appropriate questions. Over time other methods
                  of  solving non-linear problems were also included in the
                  artificial intelligence fold. These additions tended to focus
                  on the "mechanics" of problem solving rather than on
                  the "human element" in logic that the very early
                  works tried to achieve. 
                   
                    
                      Fuzzy
                      logic introduced flexibility and non-linear responses
                      to rule based systems.
                    
                      Neural
                      networks are a type of  adaptive learning
                      programs. They combine a network of numerical processing
                      units each with its own weight, or importance, to act
                      together.  Through trial and error they search for
                      the answer to the problem.  Neural networks do not
                      need to have a set of rules to solve a problem, they
                      surmise the rules from a set of historical data. Genetic algorithms are
                      another type of problem solving program. They work in a
                      manner similar to the way nature allows for "survival
                      of the fittest". They allow hundreds or thousands of
                      possible solutions to compete with each other to find the
                      best solution. Unlike neural networks, you must have a
                      deeper understanding of the nature of the problem you are
                      trying to solve, in order to apply the genetic algorithms. There is a distinct difference between
                  the way humans think and the way most "artificial
                  intelligence" programs solve problems.  The stock
                  market problem is not just about economics or technical
                  analysis, it is about humans making decisions.  It is
                  the perception of what other investors are about to do.  Solving
                  the stock market problem is solving two problems. One is
                  trying to extrapolate the non-linear response to market
                  stimuli and the other is trying to replicate human response. 
                  Our model takes both options into consideration.  free
T-Index software
 |  To believe with certainty, we must begin by doubting.---
 
 
 
 
 
 |