System and methods for facilitating pattern recognition

A method for facilitating pattern recognition through use of analysis by a plurality of layers organized into at least one module to form a hierarchy, the method comprising:

• collecting information through use of one or more sensors in the form of a time series for each of the one or more sensors;
• processing the each of the one or more time series using one or more overlapping filtering functions to encode the each of the one or more time series redundantly into a binary code to produce a population code, the population code including one or more clusters, wherein each of the one or more clusters comprises a set of many population activity patterns;
• communicating through a network of connections the information to a first layer in the at least one module at which the information is initially analyzed to identify one or more clusters in the input population code, wherein each cluster represents a sensory event, wherein said initially analyzing step
• applies a learning rule to the input population code such that each node in a first population of nodes in the first layer is activated by one or more clusters;
• upon activation of the node in the first population of nodes in the first layer, transferring the each of the one or more clusters, via one or more synapses, to a second layer in the at least one modules at which the each of the one or more clusters are evaluated to determine one or more temporal sequences, wherein the second layer of the at least one modules comprises a second population of nodes, wherein each node of the second population of nodes is connected to at least a second node in the second population of nodes through at least one recurrent synapse, wherein the at least one recurrent synapse operates using a plasticity rule, wherein one or more nodes in the second population of nodes is activated by one or more temporal sequences of sensory events represented in the first layer;
• transferring the one or more sequences from the second layer of the at least one module to a third layer in the at least one module to compute sequence predictions, wherein the sequence predictions are fed back into the first layer of the of the at least one modules to enhance an activity of correct predictions of the next sensory event in the sequence, wherein the sequence predictions are also fed back to a second layer of a second module one step down in the hierarchy;
• and
• distributing the one or more sequences from the second layer of the at least one modules to a first layer of a second module at a next stage in the hierarchy.