Devices and methods for increasing the speed and efficiency at which a computer is capable of modeling a plurality of random walkers using a particle method

Patent metadata

Field	Value
Publication number	US-11281964-B2
Application number	US-201816020619-A
Country	US
Kind code	B2
Filing date	Jun 27, 2018
Priority date	Jun 27, 2018
Publication date	Mar 22, 2022
Grant date	Mar 22, 2022

Abstract

A method for increasing a speed or energy efficiency at which a computer is capable of modeling a plurality of random walkers. The method includes defining a virtual space in which a plurality of virtual random walkers will move among different locations in the virtual space. The method also includes either assigning a corresponding set of ringed neurons in a spiking neural network to a corresponding virtual random walker, or assigning a corresponding set of ringed neurons to a point in the virtual space. Movement of a given virtual random walker is tracked by decoding differences between states of individual neurons in a corresponding given set of ringed neurons. A virtual random walk of the plurality of virtual random walkers is executed using the spiking neural network.

First claim

What is claimed is:

1. A method for modeling random walkers in a circuit, the method comprising:
defining a virtual space in which a plurality of virtual random walkers will move among different locations in the virtual space;
assigning a corresponding sets of ringed neurons in a spiking neural circuit to corresponding virtual random walkers such that there is a one-to-one correspondence between each set of ringed neurons and a virtual random walker, wherein movement of a given virtual random walker is tracked by decoding differences between states of individual neurons in a corresponding given set of ringed neurons, wherein the spiking neural circuit comprising a plurality of sets of ringed spiking neurons, and wherein the sets of ringed spiking neurons comprise a prime number of neurons; and
executing with the spiking neural circuit, a stochastic process comprising a Markov process that uses of the plurality of virtual random walkers, wherein executing includes tracking all movements of all of the plurality of virtual random walkers.

2. The method of claim 1 wherein the corresponding set of ringed neurons comprises a first set of ringed neurons and a second set of ringed neurons, and wherein differences between neuron positions in the first set and the second set determine a position of the corresponding virtual random walker.

3. The method of claim 1 wherein the corresponding set of ringed neurons comprises a first set of ringed neurons, a second set of ringed neurons, and a third set of ringed neurons, and wherein differences between neuron positions in the first set, the second set, and the third set determine a position of the corresponding virtual random walker.

4. The method of claim 1 further comprising:
using a set of neurons to introduce a spike delay to add or subtract a time increment when a given neuron in a given ring will trigger.

5. The method of claim 1 further comprising:
using a secondary circuit placed on all rings of a given walker to advance or stall triggering of the corresponding set of ringed neurons.

6. The method of claim 1 further comprising:
storing a result of the virtual random walk on a non-transitory computer readable storage medium.

7. The method of claim 6 further comprising:
using the result to model a physical process.

8. The method of claim 7 wherein the physical process is selected from the group consisting of: radiation transport, plasma dynamics, and molecular dynamics.

9. The method of claim 8 wherein the result is used to track a property selected from the group consisting of: path dependent behavior of particles, interactions of virtual random walkers with one another, interactions of virtual random walkers with an environment in which the virtual random walkers are walking, and combinations thereof.

10. The method of claim 6 further comprising:
using the result to perform a discrete simulation Monte Carlo (DSMC) simulation of a physical process.

11. The method of claim 6 further comprising:
using the result to model an application-specific boundary condition of a physical process.

12. The method of claim 6 further comprising:
using the result to model information propagation through a social network.

13. The method of claim 6 further comprising:
using the result to compute a property of a computerized graph database.

14. The method of claim 13 further comprising:
estimating a valuation of a financial asset.

15. The method of claim 13 further comprising:
finding one of a neighborhood or a clique within the computerized graph database.

16. The method of claim 1 , wherein the circuit comprises:
neuromorphic hardware; or
a field programmable logic array.

17. An application-specific integrated circuit (ASIC) configured to execute a stochastic process comprising a Markov process, the ASIC comprising:
a processor architecture that implements a spiking neural circuit comprising a plurality of sets of ringed spiking neurons, wherein each set of ringed spiking neurons is assigned to calculate a single corresponding property of a single corresponding object, and wherein the plurality of sets of ringed spiking neurons comprises a prime number of neurons, and
wherein the plurality of sets of ringed spiking neurons comprises a first set of ringed neurons and a second set of ringed neurons, and wherein differences between neuron positions in the first set and the second set determine a position of the corresponding virtual random walker.

18. The application-specific integrated circuit of claim 17 wherein the plurality of sets of ringed spiking neurons comprises a first set of ringed neurons, a second set of ringed neurons, and a third set of ringed neurons, and wherein differences between neuron positions in the first set, the second set, and the third set determine a position of the corresponding virtual random walker.

19. The application-specific integrated circuit of claim 17 wherein the processor architecture is further modified to use a spike delay to add or subtract a time increment when a given neuron in a given ring will trigger.

20. The application-specific integrated circuit of claim 17 wherein the processor architecture is further modified to use a secondary circuit placed on all rings of a given walker to advance or stall triggering of the corresponding set of ringed neurons.

21. The application-specific integrated circuit of claim 17 further comprising:
a non-transitory computer readable storage medium storing results of a virtual random walk.

22. The applications-specific integrated circuit of claim 17 , wherein the processor architecture is implemented in neuromorphic hardware.

Assignees

• Nat Tech & Eng Solutions Sandia Llc

Inventors

• Aimone James Bradley
• Severa William Mark
• Lehoucq Richard B
• Parekh Ojas D

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N3/0495

  Quantised networks; Sparse networks; Compressed networks · CPC title

• G06N3/049 Primary

  Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs · CPC title

• G06N3/006

  based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] · CPC title

• G06N3/08

  Learning methods · CPC title