# US 2023/0419074 A1 · Methods and systems for neural and cognitive processing

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2023/0419074 A1
- Document ID: US20230419074A1
- Status: Published application
- Kind: A1
- Application number: 18/132,937
- Filed: 2023-04-10
- Published: 2023-12-28
- Earliest claimed: 2014-05-22
- Inventors: Lee J. Scheffler
- Web page: https://patentgrep.com/patent/US20230419074A1

## CPC

- G06N 3/04: Architecture, e.g. interconnection topology
- G06N 3/082: modifying the architecture, e.g. adding, deleting or silencing nodes or connections
- G06N 3/09: Supervised learning
- G06N 3/098: Distributed learning, e.g. federated learning
- G06N 3/10: Interfaces, programming languages or software development kits, e.g. for simulating neural networks
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Provided herein is a system for creating, modifying, deploying and running intelligent systems by combining and customizing the function and operation of reusable component modules arranged into neural processing graphs which direct the flow of signals among the modules, analogous in part to biological brain structure and operation as compositions of variations on functional components and subassemblies.

## Claims

### Claim 1

(canceled)

### Claim 2

A neural process development, deployment, and execution system comprising:
a development system that facilitates a user creating, modifying, saving and restoring a neural process specification and includes:
a neural module library storing a plurality of module types including input type modules, output type modules, processing type modules, and memory type modules; and
a design creation and editing canvas configured as a graphical design editor configured to define the neural process specification as a set of neural modules connected by a set of neural links;
a neural computing system that:
provides deployment, running, managing, controlling, monitoring, or visualizing the neural process specification;
processes a neural processing specification based on at least one link signal strength parameter and one or more of time-parameters or a set of dependency parameters; and
includes and at least one of a link signal pattern memory storage sub-system or a memory sequence and matching sub-system;
a user interface through which a user controls deploying a neural process specification to one or more neural computing system instances, the user interface supporting user control for a plurality of controls selected from a list consisting of starting, stopping, pausing, stepping and resuming execution, modifying execution parameters, modifying real or virtual execution rate of stamp increments, saving a neural process specification including current execution state, and restoring a saved neural process specification including a saved execution state; and
an interface that facilitates concurrent access to the neural process specification by the development system, the neural computing system, and the user interface.

### Claim 3 (depends on 2)

The system of claim 2, wherein the link signal pattern memory storage sub-system stores patterns detected that are outside of a pattern match score threshold.

### Claim 4 (depends on 2)

The system of claim 2, wherein the memory sequence and matching sub-system produces pattern matching confidence for a plurality of patterns.

### Claim 5 (depends on 2)

The system of claim 2, wherein the interface that facilitates concurrent access enables changes made to the neural process specification by a user of the development system to impact an instance of the neural process specification being executed by the neural computing system.

### Claim 6 (depends on 5)

The system of claim 5, wherein the neural computing system adapts execution of the neural process specification responsive to the changes made to the neural process specification.

### Claim 7 (depends on 5)

The system of claim 5, wherein changes made through one of the user interface or the development system cause a corresponding change in processing the neural process specification.

### Claim 8 (depends on 2)

The system of claim 2, wherein the neural computing system pre-computes signal and module interdependencies based on the set of dependency parameters and invokes module functions that preserve those dependencies.

### Claim 9 (depends on 2)

The system of claim 2, wherein the system is adapted to incorporate external system functions as modules.

### Claim 10 (depends on 2)

The system of claim 2, wherein the one or more neural computing system instances includes a neural process specification interpreter.

### Claim 11 (depends on 2)

The system of claim 2, wherein the one or more neural computing system instances includes a neural process specification compiler.

### Claim 12 (depends on 2)

The system of claim 2, wherein the one or more neural computing system instances includes a neural process specification configuration generator.

### Claim 13 (depends on 2)

The system of claim 2, wherein the development system includes one or more of a text editor, a program editor, an outline editor, a graphical editor, an immersive environment, an experiential environment, and a virtual reality environment.

### Claim 14

A system comprising:
a development system that facilitates user manipulation of a neural process specification and including:
a library of neural modules storing a plurality of module types including input type modules, output type modules, processing type modules, and memory type modules; and
a specification design module configured for a user to define a neural process specification as a set of neural modules connected by a set of neural links;
a neural computing environment that provides two or more of deployment, running, managing, controlling, monitoring, visualizing, saving and restoring neural process specifications, wherein the neural computing environment processes a neural processing specification based on a plurality of parameters including time-parameters, a set of dependency parameters, and at least one link signal strength parameter;
a user interface through which a user controls deploying a neural process specification to one or more neural computing environment instances, the user interface supporting user control for at least two of starting, stopping, pausing, stepping and resuming execution, modifying execution parameters, modifying real or virtual execution rate of stamp increments, saving a neural process specification including current execution state, and restoring a saved neural process specification including a saved execution state; and
an interface that facilitates access to the neural process specification during development and execution thereof.

### Claim 15 (depends on 14)

The system of claim 14, wherein the neural computing environment includes and at least one of a memory pattern storage sub-system or a memory sequence and matching sub-system.

### Claim 16 (depends on 15)

The system of claim 15, wherein the memory pattern storage sub-system stores patterns detected that are outside of a pattern match score threshold.

### Claim 17 (depends on 15)

The system of claim 15, wherein the memory sequence and matching sub-system produces pattern matching confidence for a plurality of patterns.

### Claim 18 (depends on 14)

The system of claim 14, wherein the development system defines a first neural process specification for center-surround lateral inhibition visual process used by biological visual cortexes to highlight edges and the neural computing environment executes the first neural process specification by performing a center-surround matrix computation function.

### Claim 19 (depends on 14)

The system of claim 14, wherein the development system defines a second neural process specification for working memory and the neural computing environment executes the second neural process specification by executing one of a working memory module from the library of neural modules or a filter module with a feedback loop.

### Claim 20

A method of development, deployment, and execution of a neural process specification, the method comprising:
storing, in a library of neural modules, a plurality of module types including input type modules, output type modules, processing type modules, and memory type modules;
defining with a neural process specification development system the neural process specification as a set of neural modules connected by a set of neural links;
accessing the neural process specification through an interface that facilitates concurrently developing, deploying, and executing the neural process specification;
deploying the neural process specification as an executable neural graph to an instance of a neural computing system that facilitates deployment, running, managing, controlling, monitoring, visualizing, saving and restoring neural process specifications;
processing the neural process specification with the instance of a neural computing system based on a link signal strength parameter and one or more of a plurality of parameters including time-parameters or a set of dependency parameters, the processing including storing data patterns in a link signal pattern memory storage sub-system; and
providing a user interface through which a user is enabled to control one or more aspects of the processing including starting, stopping, pausing, stepping and resuming processing, modifying one or more of the plurality of parameters, saving and restoring a saved neural process specification including a saved execution state, and deploying the neural process specification.

### Claim 21 (depends on 20)

The method of claim 20, further comprising pre-computing signal and module interdependencies based on the set of dependency parameters and invoking module functions that preserve those dependencies.
