Spiking network apparatus and method with bimodal spike-timing dependent plasticity

Abstract

Apparatus and methods for learning in response to temporally-proximate features. In one implementation, an image processing apparatus utilizes bi-modal spike timing dependent plasticity in a spiking neuron network. Based on a response by the neuron to a frame of input, the bi-modal plasticity mechanism is used to depress synaptic connections delivering the present input frame and to potentiate synaptic connections delivering previous and/or subsequent frames of input. The depression of near-contemporaneous input prevents the creation of a positive feedback loop and provides a mechanism for network response normalization.

1 . A method of detecting one or more representations of an object in a sequence of frames with a spiking neuron network, the method comprising:
 communicating a spiking signal to a neuron via a plurality of connections; and   based on a response generated by the neuron, the response based on a first portion of the spiking signal corresponding to a first frame of a sequence of frames:
 depressing a first connection of the plurality of connections, the first connection providing the first portion of the spiking signal; and 
 potentiating a second connection of the plurality of connections, the second connection providing a second portion of the spiking signal corresponding to a second frame of the sequence of frames, the second frame adjacent the first frame; 
   wherein the spiking signal is configured based on one or more frames of the sequence of frames.   
     
     
         2 . The method of  claim 1 , wherein:
 individual ones of the sequence of frames are characterized by luminance and chromaticity;   the object is characterized by object chromaticity; and   individual ones of the representations of the object are characterized by the object chromaticity.   
     
     
         3 . The method of  claim 1 , wherein:
 the object is characterized by object luminance; and   individual ones of the representations of the object are characterized by one or more different luminance values.   
     
     
         4 . The method of  claim 1 , wherein:
 individual frames of the sequence of frames are separated by a 40 ms time interval;   the second frame precedes or succeeds the first frame; and   the first frame and the second frame comprise two representations of the object that are separated at least by the time interval from one another.   
     
     
         5 . A computerized spiking neuron apparatus configured to encode sensory input comprising a plurality of views of an object, the apparatus comprising a plurality of computer-readable instructions configured to, when executed:
 encode individual ones of the plurality of views into a spike output by a plurality of first layer neurons;   provide the spike output to a second layer neuron via a plurality of connections associated with individual ones of the plurality of first layer neurons;   based on a response generated by the second layer neuron:
 depress a first connection of the plurality of connections, the first connection configured to provide a portion of the spike output to the second layer neuron within a first time interval from the response; and 
 potentiate a second connection of the plurality of connections, the second connection configured to provide a portion of the spike output to the second layer neuron within a second time interval from the response. 
   
     
     
         6 . The apparatus of  claim 5 , wherein the first and second connections collectively act as a single connection. 
     
     
         7 . The apparatus of  claim 5 , wherein:
 the second layer neuron is operable in accordance with a stochastic neuron response process characterized by a response generation probability;   the depression of the first connection is configured to reduce the response generation probability; and   the potentiation of the second connection is configured to increase the response generation probability.   
     
     
         8 . The apparatus of  claim 5 , wherein individual ones of the plurality of views correspond to views of the object undergoing a transformation, the transformation being selected from the group consisting of (i) a translational operation, (ii) a rotational operation, and (iii) a scaling operation. 
     
     
         9 . The apparatus of  claim 5 , wherein the first time interval overlaps, at least in part, the second time interval. 
     
     
         10 . The apparatus of  claim 5 , wherein the first time interval is adjacent the second time interval and the first time interval comprises the response. 
     
     
         11 . The apparatus of  claim 5 , wherein the second interval is configured to cover a time period outside the first interval. 
     
     
         12 .- 17 . (canceled) 
     
     
         18 . A non-transitory computer-readable apparatus configured to store one or more processes thereon, the one or more processes comprising a plurality of instructions configured to, when executed:
 send an encoded digital signal to a neuron via one or more connections;   receive a response from the neuron;   based at least in part on the response, determine a first and second interval;   demote a first connection of the one or more connections during the first interval; and   promote a second connection of the one or more connection during the second interval;   wherein the encoded digital signal is configured based on one or more transformations of an object in a sequence of frames.   
     
     
         19 . The computer-readable apparatus of  claim 18 , wherein the first and second intervals comprise a period of overlap. 
     
     
         20 . The computer-readable apparatus of  claim 18 , wherein:
 the first interval is based on a portion of the response corresponding to a first frame of the sequence of frames;   the second interval is based on another portion of the response corresponding to at least a second frame of the sequence of frames; and   the first and second intervals are determined so as to associate the one or more transformations with the object.   
     
     
         21 . The computer-readable apparatus of  claim 18 , wherein:
 the connection promotion is characterized by a first gain;   the connection demotion is characterized by a second gain, the first gain being greater than the second gain; and   the connection promotion is characterized by a second interval overlapping the first interval.   
     
     
         22 . The computer-readable apparatus of  claim 18 , wherein:
 the connection is characterized by a weight parameter configured to affect generation of the another response by another neuron;   a larger weight corresponds to a greater probability of the response being generated;   the connection promotion comprises increasing the weight parameters; and   the connection demotion comprises reducing the weight parameter.   
     
     
         23 . The computer-readable apparatus of  claim 18 , wherein:
 the encoded digital signal comprises one or more spikes; and   the first time interval is configured based on one or more first difference values comprising times extending between the response and individual ones of the one or more spikes.   
     
     
         24 . The computer-readable apparatus of  claim 23 , wherein:
 the difference values comprise a value of zero;   the second time interval being configured based on one or more second difference values comprising other times extending between the response and the time of other ones of the one or more spikes; and   at least a portion of the second difference values are greater than a respective portion of the first difference values.   
     
     
         25 . The computer-readable apparatus of  claim 18 , wherein:
 the second time interval being configured based on one or more second difference values comprising other times extending between the response and the time of other ones of one or more spikes; and   at least another portion of the second difference values are smaller than another respective portion of one or more first difference values.   
     
     
         26 . The computer-readable apparatus of  claim 25 , wherein:
 the encoded digital signal comprises a plurality of spike trains;   the sequence of frames is characterized by a frame refresh rate; and   for a value of the frame refresh rate of 25 Hz:
 the first time interval is selected from the range between 1 and 100 ms inclusive; and 
 the second time interval is selected from the range between 5 and 1000 ms inclusive.