Method and apparatus for neural learning of natural multi-spike trains in spiking neural networks
First Claim
Patent Images
1. A method of neural learning, comprising:
- modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; and
adapting the weight of the synapse based on a resource associated with the synapse, wherein a value of the resource is modified as a function of time relative to adapting the weight of the synapse, such that the value of the resource is recovered over time towards a previous value before the modification.
2 Assignments
0 Petitions
Accused Products
Abstract
Certain aspects of the present disclosure support a technique for neural learning of natural multi-spike trains in spiking neural networks. A synaptic weight can be adapted depending on a resource associated with the synapse, which can be depleted by weight change and can recover over time. In one aspect of the present disclosure, the weight adaptation may depend on a time since the last significant weight change.
48 Citations
80 Claims
-
1. A method of neural learning, comprising:
-
modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; and adapting the weight of the synapse based on a resource associated with the synapse, wherein a value of the resource is modified as a function of time relative to adapting the weight of the synapse, such that the value of the resource is recovered over time towards a previous value before the modification. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14)
-
-
15. An apparatus for neural learning, comprising:
-
a first circuit configured to modulate an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; and a second circuit configured to adapt the weight of the synapse based on a resource associated with the synapse, wherein a value of the resource is modified as a function of time relative to adapting the weight of the synapse, such that the value of the resource is recovered over time towards a previous value before the modification. - View Dependent Claims (16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28)
-
-
29. An apparatus for neural learning, comprising:
-
means for modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; and means for adapting the weight of the synapse based on a resource associated with the synapse, wherein a value of the resource is modified as a function of time relative to adapting the weight of the synapse, such that the value of the resource is recovered over time towards a previous value before the modification. - View Dependent Claims (30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42)
-
-
43. A computer program product for neural learning, comprising a non-transitory computer-readable medium comprising code for:
-
modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; and adapting the weight of the synapse based on a resource associated with the synapse, wherein a value of the resource is modified as a function of time relative to adapting the weight of the synapse, such that the value of the resource is recovered over time towards a previous value before the modification. - View Dependent Claims (44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56)
-
-
57. A method of neural learning, comprising:
-
modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; adapting the weight of synapse based on a resource consideration associated with the synapse and spike timing of the pre-synaptic neuron and the post-synaptic neuron; and determining the resource consideration based on one or more timing relations between prior spikes of the pre-synaptic and post-synaptic neurons. - View Dependent Claims (58, 59, 60)
-
-
61. An apparatus for neural learning, comprising:
-
a first circuit configured to modulate an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; a second circuit configured to adapt the weight of synapse based on a resource consideration associated with the synapse and spike timing of the pre-synaptic neuron and the post-synaptic neuron; and a third circuit configured to determine the resource consideration based on one or more timing relations between prior spikes of the pre-synaptic and post-synaptic neurons. - View Dependent Claims (62, 63, 64)
-
-
65. An apparatus for neural learning, comprising:
-
means for modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; means for adapting the weight of synapse based on a resource consideration associated with the synapse and spike timing of the pre-synaptic neuron and the post-synaptic neuron; and means for determining the resource consideration based on one or more timing relations between prior spikes of the pre-synaptic and post-synaptic neurons. - View Dependent Claims (66, 67, 68)
-
-
69. A computer program product for neural learning, comprising a non-transitory computer-readable medium comprising code for:
-
modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; adapting the weight of synapse based on a resource consideration associated with the synapse and spike timing of the pre-synaptic neuron and the post-synaptic neuron; and determining the resource consideration based on one or more timing relations between prior spikes of the pre-synaptic and post-synaptic neurons. - View Dependent Claims (70, 71, 72)
-
-
73. A method of neural learning, comprising:
-
modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; adapting the weight of synapse based on a resource consideration associated with the synapse and spike timing of the pre-synaptic neuron and the post-synaptic neuron; and determining the resource consideration based on an amount of time since a prior commitment to a change of the weight. - View Dependent Claims (74)
-
-
75. An apparatus for neural learning, comprising:
-
a first circuit configured to modulate an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; a second circuit configured to adapt the weight of synapse based on a resource consideration associated with the synapse and spike timing of the pre-synaptic neuron and the post-synaptic neuron; and a third circuit configured to determine the resource consideration based on an amount of time since a prior commitment to a change of the weight. - View Dependent Claims (76)
-
-
77. An apparatus for neural learning, comprising:
-
means for modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; means for adapting the weight of synapse based on a resource consideration associated with the synapse and spike timing of the pre-synaptic neuron and the post-synaptic neuron; and means for determining the resource consideration based on an amount of time since a prior commitment to a change of the weight. - View Dependent Claims (78)
-
-
79. A computer program product for neural learning, comprising a non-transitory computer-readable medium comprising code for:
-
modulating an input from a pre-synaptic neuron to a post-synaptic neuron by a weight of a synapse connecting the pre-synaptic neuron and the post-synaptic neuron; adapting the weight of synapse based on a resource consideration associated with the synapse and spike timing of the pre-synaptic neuron and the post-synaptic neuron; and determining the resource consideration based on an amount of time since a prior commitment to a change of the weight. - View Dependent Claims (80)
-
Specification