Method and apparatus for producing programmable probability distribution function of pseudo-random numbers
First Claim
1. A method for reducing spikes in an artificial nervous system, the method comprising:
- generating, by an apparatus, a sequence of random numbers;
performing, by the apparatus, bit-level operations on the generated sequence of random numbers to generate a first sequence of pseudo-random numbers with a programmable probability distribution function (PDF), wherein performing the bit-level operations on the generated sequence of random numbers comprises;
performing AND operation on bits of the sequence of random numbers;
passing the ANDed bits through a cascade of OR gates to obtain bit-mask values; and
applying the bit-mask values to bits of the sequence of random numbers to generate the first sequence of pseudo-random numbers with the programmable PDF;
generating, by the apparatus, another sequence of random numbers;
performing, by the apparatus, bit-level operations between the first sequence and the other sequence to generate a second sequence of pseudo-random numbers with a programmable PDF; and
dropping, by the apparatus, spikes associated with one or more artificial neurons of the artificial nervous system in accordance with the PDF of the second sequence of pseudo-random numbers.
1 Assignment
0 Petitions
Accused Products
Abstract
Certain aspects of the present disclosure provide methods and apparatus for producing programmable probability distribution function of pseudo-random numbers that can be utilized for filtering (dropping and passing) neuron spikes. The present disclosure provides a simpler, smaller, and lower-power circuit than that typically used. It can be programmed to produce any of a variety of non-uniformly distributed sequences of numbers. These sequences can approximate true probabilistic distributions, but maintain sufficient pseudo-randomness to still be considered random in a probabilistic sense. This circuit can be an integral part of a filter block within an ASIC chip emulating an artificial nervous system.
-
Citations
32 Claims
-
1. A method for reducing spikes in an artificial nervous system, the method comprising:
-
generating, by an apparatus, a sequence of random numbers; performing, by the apparatus, bit-level operations on the generated sequence of random numbers to generate a first sequence of pseudo-random numbers with a programmable probability distribution function (PDF), wherein performing the bit-level operations on the generated sequence of random numbers comprises; performing AND operation on bits of the sequence of random numbers; passing the ANDed bits through a cascade of OR gates to obtain bit-mask values; and applying the bit-mask values to bits of the sequence of random numbers to generate the first sequence of pseudo-random numbers with the programmable PDF; generating, by the apparatus, another sequence of random numbers; performing, by the apparatus, bit-level operations between the first sequence and the other sequence to generate a second sequence of pseudo-random numbers with a programmable PDF; and dropping, by the apparatus, spikes associated with one or more artificial neurons of the artificial nervous system in accordance with the PDF of the second sequence of pseudo-random numbers. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8)
-
-
9. An apparatus, comprising:
-
a first circuit configured to generate a sequence of random numbers; a second circuit configured to perform bit-level operations on the generated sequence of random numbers to generate a first sequence of pseudo-random numbers with a programmable probability distribution function (PDF), wherein, to perform the bit-level operations on the generated sequence of random numbers, the second circuit is configured to; perform AND operation on bits of the sequence of random numbers; pass the ANDed bits through a cascade of OR gates to obtain bit-mask values; and apply the bit-mask values to bits of the sequence of random numbers to generate the first sequence of pseudo-random numbers with the programmable PDF; a third circuit configured to generate another sequence of random numbers; a fourth circuit configured to perform bit-level operations between the first sequence and the other sequence to generate a second sequence of pseudo-random numbers with a programmable PDF; and a fifth circuit configured to drop spikes associated with artificial neurons of an artificial nervous system comprising the apparatus in accordance with the PDF of the second sequence of Pseudo-random numbers. - View Dependent Claims (10, 11, 12, 13, 14, 15, 16)
-
-
17. An apparatus, comprising:
-
means for generating a sequence of random numbers; means for performing bit-level operations on the generated sequence of random numbers to generate a first sequence of pseudo-random numbers with a programmable probability distribution function (PDF), wherein means for performing bit-level operations on the generated sequence of random numbers comprises; means for performing AND operation on bits of the sequence of random numbers; means for passing the ANDed bits through a cascade of OR gates to obtain bit-mask values; and means for applying the bit-mask values to bits of the sequence of random numbers to generate the first sequence of pseudo-random numbers with the programmable PDF; means for generating another sequence of random numbers; means for performing bit-level operations between the first sequence and the other sequence to generate a second sequence of pseudo-random numbers with a programmable PDF; and means for dropping spikes associated with artificial neurons of an artificial nervous system comprising the apparatus in accordance with the PDF of the second sequence of pseudo-random numbers. - View Dependent Claims (18, 19, 20, 21, 22, 23, 24)
-
-
25. A non-transitory computer-readable medium having instructions executable to:
-
generate, by an apparatus, a sequence of random numbers; perform, by the apparatus, bit-level operations on the generated sequence of random numbers to generate a first sequence of pseudo-random numbers with a programmable probability distribution function (PDF), wherein performing the bit-level operations on the generated sequence of random numbers comprises; performing AND operation on bits of the sequence of random numbers; passing the ANDed bits through a cascade of OR gates to obtain bit-mask values; and applying the bit-mask values to bits of the sequence of random numbers to generate the first sequence of pseudo-random numbers with the programmable PDF; generate, by the apparatus, another sequence of random numbers; perform, by the apparatus, bit-level operations between the first sequence and the other sequence to generate a second sequence of pseudo-random numbers with a programmable PDF; and drop spikes associated with artificial neurons of an artificial nervous system comprising the apparatus in accordance with the PDF of the second sequence of pseudo-random numbers. - View Dependent Claims (26, 27, 28, 29, 30, 31, 32)
-
Specification