# US 2026/0300681 A1 · Systems, methods, kits, and apparatuses for generative artificial intelligence, graphical neural networks, transformer models, and converging technology stacks in value chain networks

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

- Number: US 2026/0300681 A1
- Document ID: US20260300681A1
- Status: Published application
- Kind: A1
- Application number: 19/368,110
- Filed: 2025-10-24
- Published: 2026-10-01
- Earliest claimed: 2023-04-25
- Applicants: Strong Force VCN Portfolio 2019, LLC
- Inventors: Charles H. Cella; Andrew Bunin; Andrew Cardno; Joshua Dobrowitsky; Teymour S. El-Tahry; Leon Fortin, JR.; Andrew Locke; Jenna Parenti; Nicholas Rogosin; David Stein; Eric P. Vetter
- Web page: https://patentgrep.com/patent/US20260300681A1

## CPC

- G06N 3/042: Knowledge-based neural networks; Logical representations of neural networks
- G06N 3/0985: Hyperparameter optimisation; Meta-learning; Learning-to-learn
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A system may execute, by a generative artificial intelligence system, generative artificial intelligence algorithms trained on value chain network data. A system may receive input data including at least one of images, video, audio, text, programmatic code, and data, process the input data using the generative artificial intelligence algorithms to generate output content, wherein the output content includes at least one of structured prose, images, video, audio content, software source code, formatted data, algorithms, definitions, and context-specific structures, and generate an internal state of the generative artificial intelligence system, including a set of weights and/or biases as a result of prior processing. A system may provide the generated output content to a user interface for presentation to a user.

## Claims

### Claim 1

A generative artificial intelligence system for a value chain network, the system including one or more processors and one or more memories configured to perform operations comprising:
executing, by the generative artificial intelligence system, generative artificial intelligence algorithms trained on value chain network data;
receiving input data including at least one of images, video, audio, text, programmatic code, and data;
processing the input data using the generative artificial intelligence algorithms to generate output content, wherein the output content includes at least one of structured prose, images, video, audio content, software source code, formatted data, algorithms, definitions, and context-specific structures;
generating an internal state of the generative artificial intelligence system, including a set of weights and/or biases as a result of prior processing; and
providing the generated output content to a user interface for presentation to a user.

### Claim 2 (depends on 1)

The generative artificial intelligence system of claim 1, wherein the input data further includes natural-language expressions, single- or multidimensional shapes or models, real-world and/or virtual scene representations, LIDAR point-cloud representations, sensor inputs and/or outputs, vehicle and/or machine telemetry, geographic maps, authentication credentials, financial transactions, smart contracts, processing directives, and device configurations.

### Claim 3 (depends on 1)

The generative artificial intelligence system of claim 1, wherein the generative artificial intelligence system is configured to maintain contextual awareness across interactions to facilitate continuity in dialogue with users.

### Claim 4 (depends on 1)

The generative artificial intelligence system of claim 1, wherein the generative artificial intelligence system is configured to support interpretability and explainability of its outputs by providing a description of a basis of the output.

### Claim 5 (depends on 1)

The generative artificial intelligence system of claim 1, wherein the output includes a recommendation to improve one or more of the value chain network, an entity within the value chain network, an asset within the value chain network, and a value chain network environment.

### Claim 6 (depends on 1)

The generative artificial intelligence system of claim 1, further comprising a digital twin interface for a digital twin, wherein the digital twin interface is configured to enable access to the generative artificial intelligence system, and wherein the digital twin represents at least one of the value chain network, an entity within the value chain network, an asset within the value chain network, or an environment within the value chain network.

### Claim 7 (depends on 1)

The generative artificial intelligence system of claim 1, further comprising a digital twin module configured to generate a digital twin representing at least one of a value chain network, an entity within the value chain network, an asset within the value chain network, and an environment within the value chain network.

### Claim 8 (depends on 1)

The generative artificial intelligence system of claim 1, wherein the processor is further configured to communicate with robotic systems within the value chain network, and the memory stores instructions that enable the generative artificial intelligence system to translate human commands into robotic actions, thereby facilitating human-to-robot interaction.

### Claim 9 (depends on 1)

The generative artificial intelligence system of claim 1, wherein the output content is further tailored based on an executive role of the user.

### Claim 10 (depends on 1)

The generative artificial intelligence system of claim 1, wherein the generated output includes 3D printing instructions for a set of 3D printers.

### Claim 11 (depends on 1)

The generative artificial intelligence system of claim 1, wherein the system is part of a Dual Process Artificial Neural Network (DPANN) architecture.

### Claim 12 (depends on 1)

The generative artificial intelligence system of claim 1, further comprising an augmented reality (AR) interface configured to overlay the generated output content onto a user's real-world view.

### Claim 13

A method for operating a generative artificial intelligence system within a value chain network, the method comprising:
executing, by the generative artificial intelligence system, generative artificial intelligence algorithms trained on value chain network data;
receiving input data including at least one of images, video, audio, text, programmatic code, and data;
processing the input data using the generative artificial intelligence system to generate output content, wherein the output content includes at least one of structured prose, images, video, audio content, software source code, formatted data, algorithms, definitions, and context-specific structures;
generating an internal state of the generative artificial intelligence system, including a set of weights and/or biases as a result of prior processing; and
providing the generated output content to a user interface for presentation to a user.

### Claim 14 (depends on 13)

The method of claim 13, further comprising receiving additional input data including natural-language expressions, single- or multidimensional shapes or models, real-world and/or virtual scene representations, LIDAR point-cloud representations, sensor inputs and/or outputs, vehicle and/or machine telemetry, geographic maps, authentication credentials, financial transactions, smart contracts, processing directives, and device configurations.

### Claim 15 (depends on 13)

The method of claim 13, further comprising maintaining contextual awareness across interactions to facilitate continuity in dialogue with users.

### Claim 16 (depends on 13)

The method of claim 13, further comprising supporting interpretability and explainability of its outputs by providing a description of a basis of the output.

### Claim 17 (depends on 13)

The method of claim 13, wherein the generated output includes a recommendation to improve one or more of a value chain network, an entity within a value chain network, an asset within a value chain network, and an environment within the value chain network.

### Claim 18 (depends on 13)

The method of claim 13, wherein the method further comprises:
interfacing with a digital twin via a digital twin interface, wherein the digital twin interface is configured to facilitate access to the generative artificial intelligence system; and
representing, by the digital twin, one or more of the value chain network, an entity within the value chain network, an asset within the value chain network, and an environment within the value chain network.

### Claim 19 (depends on 13)

The method of claim 13, wherein the method further comprises:
generating, through a digital twin module, a digital twin that represents at least one of a value chain network, an entity within the value chain network, an asset within the value chain network, and an environment within the value chain network.

### Claim 20 (depends on 13)

The method of claim 13, wherein the method further comprises:
communicating with one or more robotic systems within the value chain network;
translating human commands into robotic actions; and
facilitating human-to-robot interaction within the value chain network, thereby enabling the robotic systems to perform tasks in response to the human commands derived from the generated output content.

### Claim 21 (depends on 13)

The method of claim 13, wherein the method further comprises:
tailoring the output content based on an executive role of the user within an organization.

### Claim 22 (depends on 13)

The method of claim 13, wherein the generated output includes 3D printing instructions for a set of 3D printers.

### Claim 23 (depends on 13)

The method of claim 13, wherein the generative artificial intelligence system operates as part of a Dual Process Artificial Neural Network (DPANN) architecture.

### Claim 24 (depends on 13)

The method of claim 13, wherein the method further comprises:
interfacing with an augmented reality (AR) interface; and
overlaying, by the AR interface, the generated output content onto a user's real-world view.

### Claim 25

A system for improving a value chain networks using graph neural networks, the system including one or more processors and one or more memories configured to perform operations comprising:
receiving an input graph data set representing a value chain network, including nodes representing entities within the value chain and edges representing transactions or relationships between entities;
generating an embedding based on the input graph data set specific to a value chain context;
applying a plurality of graph neural network layers to the embedding to produce an output graph data set, wherein the output graph data set includes updates to nodes and edges reflecting changes in the value chain network;
performing a clustering analysis on the nodes or edges of the output graph data set to identify clusters representing highly interconnected segments of the value chain; and
outputting optimized value chain network graph data set including updated node properties, edge properties, or graph properties.

### Claim 26 (depends on 25)

The system of claim 25, wherein the system is further configured to perform dimensionality reduction on the graph data set based on a similarity matrix to facilitate the clustering analysis.

### Claim 27 (depends on 25)

The system of claim 25, wherein the system is further configured to apply spectral clustering techniques during the clustering analysis.

### Claim 28 (depends on 25)

The system of claim 25, wherein the system is further configured to generate an output graph data set that includes graphlet degree vectors indicating recurring structures within the input graph data set.

### Claim 29 (depends on 25)

The system of claim 25, wherein the system is further configured to perform graph-to-graph translation to generate the output graph data set that represents a different graph data set from the input graph data set.

### Claim 30 (depends on 25)

The system of claim 25, wherein the system is further configured to utilize instance-level explanations based on gradients, features, perturbations, or decomposition to provide interpretability of the output graph data set.

### Claim 31 (depends on 25)

The system of claim 25, wherein the system is further configured to perform hyperparameter search processing to refine the graph neural network based on user feedback regarding the output graph data set.

### Claim 32 (depends on 25)

The system of claim 25, wherein the value chain network entities include at least two of products, suppliers, producers, manufacturers, retailers, businesses, owners, operators, operating facilities, customers, consumers, workers, mobile devices, wearable devices, distributors, resellers, supply chain infrastructure facilities, supply chain processes, logistics processes, reverse logistics processes, demand prediction processes, demand management processes, demand aggregation processes, machines, ships, barges, warehouses, maritime ports, airports, airways, waterways, roadways, railways, bridges, tunnels, online retailers, ecommerce sites, demand factors, supply factors, delivery systems, floating assets, points of origin, points of destination, points of storage, points of use, networks, information technology systems, software platforms, distribution centers, fulfillment centers, containers, container handling facilities, customs, export control, border control, drones, robots, robotic handling systems, 3D printers, vehicles, autonomous vehicles, hauling facilities, waterways, and port infrastructure facilities.

### Claim 33

A computer-implemented method for optimizing value chain networks using graph neural networks, comprising:
receiving, by a computing device, an input graph data set representing a value chain network, including nodes representing entities within the value chain and edges representing transactions or relationships between entities;
generating an embedding based on the input graph data set specific to a value chain context;
applying a plurality of graph neural network layers to the embedding to produce an output graph data set, wherein the output graph data set includes updates to nodes and edges reflecting changes in the value chain network;
performing a clustering analysis on the nodes or edges of the output graph data set to identify clusters representing highly interconnected segments of the value chain; and
outputting the optimized value chain network graph data set including updated node properties, edge properties, or graph properties.

### Claim 34 (depends on 33)

The computer-implemented method of claim 33, further comprising performing dimensionality reduction on the graph data set based on a similarity matrix to facilitate the clustering analysis.

### Claim 35 (depends on 33)

The computer-implemented method of claim 33, further comprising applying spectral clustering techniques during the clustering analysis.

### Claim 36 (depends on 33)

The computer-implemented method of claim 33, further comprising generating an output graph data set that includes graphlet degree vectors indicating recurring structures within the input graph data set.

### Claim 37 (depends on 33)

The computer-implemented method of claim 33, further comprising performing graph-to-graph translation to generate the output graph data set that represents a different graph data set from the input graph data set.

### Claim 38 (depends on 33)

The computer-implemented method of claim 33, further comprising utilizing instance-level explanations based on gradients, features, perturbations, or decomposition to provide interpretability of the output graph data set.

### Claim 39 (depends on 33)

The computer-implemented method of claim 33, further comprising performing hyperparameter search processing to refine the graph neural network based on user feedback regarding the output graph data set.

### Claim 40 (depends on 33)

The computer-implemented method of claim 33, wherein the value chain network entities include at least two of products, suppliers, producers, manufacturers, retailers, businesses, owners, operators, operating facilities, customers, consumers, workers, mobile devices, wearable devices, distributors, resellers, supply chain infrastructure facilities, supply chain processes, logistics processes, reverse logistics processes, demand prediction processes, demand management processes, demand aggregation processes, machines, ships, barges, warehouses, maritime ports, airports, airways, waterways, roadways, railways, bridges, tunnels, online retailers, ecommerce sites, demand factors, supply factors, delivery systems, floating assets, points of origin, points of destination, points of storage, points of use, networks, information technology systems, software platforms, distribution centers, fulfillment centers, containers, container handling facilities, customs, export control, border control, drones, robots, robotic handling systems, 3D printers, vehicles, autonomous vehicles, hauling facilities, waterways, and port infrastructure facilities.

### Claim 41

A system for improving a value chain network (VCN) using transformer models, the system including one or more memories and one or more processors configured to perform operations comprising:
receiving an input data set representing a VCN including nodes and edges;
encoding the input data set using an encoder of a transformer model to generate a contextual representation;
decoding the contextual representation using a decoder of the transformer model to generate an output data set indicative of VCN improvements; and
outputting the improved VCN data set for facilitating decision-making in the VCN.

### Claim 42 (depends on 41)

The system of claim 41, wherein the system is further configured to apply multi-head attention mechanisms to determine relationships between nodes and edges within the VCN.

### Claim 43 (depends on 41)

The system of claim 41, wherein the system is further configured to perform autoregressive prediction of VCN configurations using the output data set.

### Claim 44 (depends on 41)

The system of claim 41, wherein the system is further configured to utilize positional encoding to maintain a sequence order of the VCN data within the transformer model.

### Claim 45 (depends on 41)

The system of claim 41, wherein the system is further configured to implement layer normalization within the transformer model.

### Claim 46 (depends on 41)

The system of claim 41, wherein the system is further configured to adjust a set of transformer model parameters based on a set of training data representing historical VCN performance.

### Claim 47 (depends on 41)

The system of claim 41, wherein the system is further configured to generate predictions for VCN resource allocation based on the output data set.

### Claim 48

A computer-implemented method for improving a value chain network (VCN) using transformer models, comprising:
receiving, by a computing device, an input data set representing a VCN including nodes and edges;
encoding the input data set using an encoder of a transformer model to generate a contextual representation;
decoding the contextual representation using a decoder of the transformer model to generate an output data set indicative of VCN optimizations; and
outputting the improved VCN data set for facilitating decision-making in the VCN.

### Claim 49 (depends on 48)

The computer-implemented method of claim 48, further comprising applying multi-head attention mechanisms to determine relationships between nodes and edges within the VCN.

### Claim 50 (depends on 48)

The computer-implemented method of claim 48, further comprising performing autoregressive prediction of VCN configurations using the output data set.

### Claim 51 (depends on 48)

The computer-implemented method of claim 48, further comprising utilizing positional encoding to maintain a sequence order of the VCN data within the transformer model.

### Claim 52 (depends on 48)

The computer-implemented method of claim 48, further comprising implementing layer normalization within the transformer model.

### Claim 53 (depends on 48)

The computer-implemented method of claim 48, further comprising adjusting a set of transformer model parameters based on a set of training data representing historical VCN performance.

### Claim 54 (depends on 48)

The computer-implemented method of claim 48, further comprising generating predictions for VCN resource allocation based on the output data set.

### Claim 55

A dynamic vision system for a mobile device, the dynamic vision system comprising:
a variable focus liquid lens optical assembly;
a control system configured to adjust one or more optical parameters and data collected from the variable focus liquid lens optical assembly in real time; and
a processing system that dynamically learns on a training set of outcomes, parameters, and data collected from the variable focus liquid lens optical assembly to train one or more machine learning models to recognize an object relating to the mobile device.

### Claim 56 (depends on 55)

The dynamic vision system of claim 55 wherein
the variable focus liquid lens assembly is continuously adjusted by the control system based on environment factors and on feedback from the processing system to generate an object concept.

### Claim 57 (depends on 56)

The dynamic vision system of claim 56 wherein
the object concept includes contextual intelligence about the object and its environment and provides superior object recognition by the dynamic vision system.

### Claim 58 (depends on 55)

The dynamic vision system of claim 55 wherein
a first machine learning model is used to optimize collection of signals by the variable focus liquid lens optical assembly and a second machine learning model is used to operate on the signals to achieve a desired vision outcome.

### Claim 59 (depends on 55)

The dynamic vision system of claim 55 wherein
the processing system receives real-time adjustable data streams from the variable focus liquid lens optical assembly to generate situational awareness or create out-of-focus images of the object so as to capture rich metadata and contextual intelligence about the object and its environment.

### Claim 60 (depends on 55)

The dynamic vision system of claim 55 wherein
the control system and the processing system are integrated with the variable focus liquid lens optical assembly.

### Claim 61 (depends on 55)

The dynamic vision system of claim 55 wherein
the optical parameters adjusted by the control system include at least one of focal length, specularity, color, environment, and lens shape.

### Claim 62 (depends on 61)

The dynamic vision system of claim 61 wherein
adjustment of the optical parameters result in a change to at least one of spherical aberration, field curvature, coma, chromatic aberration, distortion, vignetting, ghosting, flaring, and diffraction.

### Claim 63 (depends on 55)

The dynamic vision system of claim 55 wherein:
the processing system is configured to learn on a set of outcomes, parameters, and data from the variable focus liquid lens optical assembly to derive a configuration of the liquid lens optical assembly, and
the configuration includes at least one of liquid lens materials, geometry, shape, optical properties, performance, and design.

### Claim 64 (depends on 55)

The dynamic vision system of claim 55 wherein
the machine learning models are embodied on a semiconductor chip that is integrated into the mobile device that houses the variable focus liquid lens optical assembly.

### Claim 65 (depends on 55)

The dynamic vision system of claim 55 wherein:
the machine learning models are pre-trained on a separate system, such as in a cloud computing environment, such as using a large training data set of visual information and/or outcomes, to perform a set of machine vision tasks, and
the pre-trained machine learning models are deployed on a device or system that includes the variable focus liquid lens optical assembly.

### Claim 66 (depends on 55)

The dynamic vision system of claim 55 wherein the mobile device includes at least one of a smartphone, a tablet, a smart watch, an augmented reality (AR) device, and a handheld gaming console.

### Claim 67

A method for dynamically adjusting a vision system in a mobile device, the method comprising:
adjusting a variable focus liquid lens optical assembly;
controlling, by a control system, one or more optical parameters and data collected from the variable focus liquid lens optical assembly in real time; and
processing, by a processing system, a set of outcomes, parameters, and data collected from the variable focus liquid lens optical assembly to train one or more machine learning models to recognize an object relating to the mobile device.

### Claim 68 (depends on 67)

The method of claim 67, further comprising continuously adjusting the variable focus liquid lens assembly based on environment factors and feedback from the processing system to generate an object concept.

### Claim 69 (depends on 68)

The method of claim 68, wherein generating the object concept includes incorporating contextual intelligence about the object and its environment to provide superior object recognition by the dynamic vision system.

### Claim 70 (depends on 67)

The method of claim 67, further comprising using a first machine learning model to optimize collection of signals by the variable focus liquid lens optical assembly and a second machine learning model to operate on the signals to achieve a desired vision outcome.

### Claim 71 (depends on 67)

The method of claim 67, further comprising receiving real-time adjustable data streams from the variable focus liquid lens optical assembly by the processing system to generate situational awareness or create out-of-focus images of the object to capture rich metadata and contextual intelligence about the object and its environment.

### Claim 72 (depends on 67)

The method of claim 67, wherein the control system and the processing system are integrated with the variable focus liquid lens optical assembly.

### Claim 73 (depends on 67)

The method of claim 67, further comprising adjusting optical parameters by the control system, including at least one of focal length, specularity, color, environment, and lens shape.

### Claim 74 (depends on 73)

The method of claim 73, wherein the adjustment of the optical parameters results in a change to at least one of spherical aberration, field curvature, coma, chromatic aberration, distortion, vignetting, ghosting, flaring, and diffraction.

### Claim 75 (depends on 67)

The method of claim 67, further comprising configuring the processing system to learn on a set of outcomes, parameters, and data from the variable focus liquid lens optical assembly to derive a configuration of the liquid lens optical assembly, including at least one of liquid lens materials, geometry, shape, optical properties, performance, and design.

### Claim 76 (depends on 67)

The method of claim 67, wherein the machine learning models are embodied on a semiconductor chip that is integrated into the mobile device housing the variable focus liquid lens optical assembly.

### Claim 77 (depends on 67)

The method of claim 67, further comprising pre-training the machine learning models on a separate system, such as in a cloud computing environment, using a large training data set of visual information and/or outcomes to perform a set of machine vision tasks, and deploying the pre-trained machine learning models on a device or system that includes the variable focus liquid lens optical assembly.

### Claim 78 (depends on 67)

The method of claim 67, wherein the mobile device includes at least one of a smartphone, a tablet, a smartwatch, an augmented reality (AR) device, and a handheld gaming console.
