Our goal is to implement a memory-prediction framework as described in Jeff Hawkins' book On Intelligence
The On Intelligence book appeared in 2004. The framework proposed by Jeff Hawkins is a proposal to explain how the brain works.
The ideas put forth surprised me as being clear, simple and bore resemblance to my older work on modified markov models used for natural language processing.
I started implementing the MPF in december 2004. Until now (end of april 2005) I produced a simple predictor class which learns patterns and can be linked to other predictors.
Currently I'm wondering in several directions: