Digital Commons Network^™

Sequential Consensus Building For Multi-Agent Debates, Andrea Morandi

Defensive Publications Series

Conventional approaches to consensus building in multi-agent systems typically rely on fixed numbers of debate rounds or predetermined thresholds, which can be inefficient and inflexible. Moreover, these conventional approaches often fail to account for the dynamic nature of debates and waste computational resources by continuing debates that have already reached sufficient consensus or by prematurely terminating debates. To address these challenges, the techniques presented herein leverage sequential analysis to efficiently and dynamically determine consensus in multi-agent systems.

Executive Synthesis Of Luxury Ai 2025, Xavier Pillet

Defensive Publications Series

This document, produced with the assistance of ChatGPT o3 and ChatGPT 5 Thinking, is released under the Apache 2.0 licence. It is a voluntary defensive publication (prior art) and therefore enters the prior art upon release under the applicable patent statutes: art. L 611-11 CPI (French Intellectual Property Code), art. 54(2) CBE (European Patent Convention), 35 U.S.C. §102(a) (United States Patent Act), Chinese Patent Law Art. 22(5) (中华人民共和国专利法), and Japanese Patent Act Art. 29(1) (特許法). This defensive publication discloses seventy (70) enabling inventions across luxury AI systems, including hyperspectral grading, polarimetric textile inspection, watermarked CAD assets, PUF-based leather authentication, federated …

Unified Framework To Boost Roi In Product Engineering Services (Pes), Xavier Pillet

Defensive Publications Series

This document, produced with the assistance of ChatGPT o3 and ChatGPT 5 Thinking, is released under the Apache 2.0 licence. It is a voluntary defensive publication (prior art) and therefore enters the prior art upon release under the applicable patent statutes: art. L 611-11 CPI (French Intellectual Property Code), art. 54(2) CBE (European Patent Convention), and 35 U.S.C. §102(a) (United States Patent Act). It discloses 140 enabling innovations covering devices, methods, causal experimentation OS, ROI ledgers, digital twins, connectors, calibration-as-a-service, compliance automation, ESG/supply-chain coupling, and biomedical/medtech transpositions. Every proposal is described in an enabling manner, classified with IPC and CPC …

Automated Video Production Workflow Using Textual Source Documents, Kumar Dhanagopal

Defensive Publications Series

This disclosure describes a multi-stage workflow to facilitate automated video production from textual source documents. A system may process source documents to generate a narrative audio track. This audio can then be analyzed by a computing system to produce a corresponding visual storyboard, for example, by identifying concepts and scenes. Based on information from this storyboard, generative tools may be used to create multimedia assets, such as graphics and video clips. The generated audio and visual components may then be assembled into a video. This integrated method provides a process for converting text to video that can potentially reduce manual …

Computer Control By Agentic Systems With World Models Of User Interfaces, Florian Hartmann, Victor Carbune

Defensive Publications Series

Computer control functionality refers to the capabilities of artificial intelligence (AI) agents that enable the agent to use a computer as a human does, e.g., by interacting with user interfaces via operations of input devices. For accurate task completion, it is important for the AI agent to reason about the task currently being performed and to determine what effects the actions it performs may have on the environment. This disclosure describes techniques that enable an AI agent to learn to estimate the reversibility of an action in a user interface (UI) and to use the reversibility estimate to adapt the …

Automatic Integration Of Temporal Context For Large Language Models, Na

Defensive Publications Series

Current LLMs struggle with real-time knowledge, tracking elapsed time, chronological context, and generating time-sensitive responses, leading to user burden and limited applicability in time-sensitive domains. This disclosures describes techniques for integrating real-time temporal context into large language model (LLM) processing of user queries, addressing their inherent lack of temporal awareness. The techniques described herein automatically obtain current time from a reliable time source as well as relevant temporal context such as conversation history between an LLM-powered chatbot and a user that provides a query. The time information together with the user query is provided to the LLM which can thereby …

Target Markets 2024–2030 For Ai Products, Xavier Pillet

Defensive Publications Series

This document, produced with the assistance of ChatGPT o3 and ChatGPT 5 Thinking, is released under the Apache 2.0 licence. It is a voluntary defensive publication (prior art) and therefore enters the prior art upon release under the applicable patent statutes: (art. L 611-11 CPI / art. 54(2) CBE), 35 U.S.C. §102(a) (United States Patent Act), Chinese Patent Law Art. 22(5) (中华人民共和国专利法), and Japanese Patent Act Art. 29(1) (特許法). It consolidates 80 reproducible technical contributions across software, sensors, hardware, AI, protocols, materials, manufacturing, imaging, delivery, therapies, data/privacy, UX/workflows, supply chain and business processes. Each item is described in an enabling …

Large Language Model Embedding Graph For Content Recommendation And Discovery, Tamojit Chatterjee, Shravan Nayak

Defensive Publications Series

Existing content recommendation algorithms often do not effectively use structured information to make more informed decisions. This can limit the ability to discover similar content or to introduce sufficient diversity into recommendations. In addition, or in the alternative, current techniques for enhancing content recommendations may lack personalization, diversity, and transparency. These existing techniques may also present static sets of content items (e.g., content clusters) that may not adapt to real-time events. Also, content cluster labels may not clearly describe the underlying theme of the content items in the content cluster.

This disclosure describes technology for generating content recommendations using a …

Parametric Head-Related Transfer Function Using Structural Components And Spherical Harmonics, Nikhil Bhanu, Mengfan Zhang

Defensive Publications Series

The head-related transfer function (HRTF) describes the acoustic transmission of sound waves from a sound source to a listener’s binaural ears. When applied to headphones, VR headsets, and similar devices, the use of non-personalized HRTFs may lead to perception errors such as inside-the-head localization, front-back confusion, and/or a breakdown of elevation discrimination ability. This disclosure describes techniques to personalize HRTFs for spatial audio products. A compact, spatial-domain HRTF representation is constructed that achieves a realistic spatial listening experience by structurally modeling person-specific acoustic components such as head shadowing, torso reflection, and pinna (outer ear) filtering. The HRTF is decomposed into …

Providing Spoiler-Free Entity Information For Narrative Media Based On Viewing Progress, Alejandra Molina, Vivian Tsai, Ellen Jiang

Defensive Publications Series

When reading or viewing long-form narrative media such as books, audio content, or television shows, or engaging in games with long gameplay durations, users may sometimes forget a particular detail about a character or entity in the content. While online resources as well as built-in features in the viewing application can help aid user recall, such resources are static and do not account for the user’s current progress, thus risking spoilers. This disclosure describes techniques that enable users to obtain entity information for narrative media in a spoiler-free manner. Per the techniques, a structured database of various entities in narrative …

Automatic Generation Of A Labeled Corpus Of Conversational Failures During Interaction Between A User And A Conversational Ai Agent, Ilia Labzovsky, Danny Karmon

Defensive Publications Series

Current techniques to detect conversational failures during interaction between users and conversational artificial intelligence (AI) have many limitations - they capture only a tiny fraction of failures, lack context about the user's internal state during the conversation, and are not scalable. This disclosure describes techniques to automatically detect, classify, and log conversational failures in interactions between users and conversational AI agents to generate a rich, structured dataset. With user permission, conversations are analyzed to detect prosodic features and acoustic cues within context to determine likely occurrences of conversational failures. The conversational state is tracked using a state vector that encodes …

A Method And A System For Predicting E-Commerce Consumer Return Risks And Behavior Analysis, Ishika Ahuja, Hema Lakshmi K M, Kireeti Kesavamurthy, Bhuvanjeet Singh Gandhi

Defensive Publications Series

The present disclosure relates to a system and method for managing e-commerce return and refund risks by utilizing a comprehensive data aggregation module that collects information metrics from multiple e-commerce platforms using a unique identifier. The user's purchase profile is generated through a user profiling component leveraging Visa ID to link all active identification identifiers. A risk assessment engine computes various risk indicators based on historical purchasing behaviors, including sales-to-refund ratios and fraud signals. Furthermore, the system includes a dashboard interface that presents comparative metrics, allowing merchants to analyze consumer behavior against a broader consumer base. This structured approach facilitates …

Method And System For Cognitive Optimization Of Distributed Engines Framework, Swagat Kumar, Amitabh Ranjan, Soumendra Kumar Mishra

Defensive Publications Series

Present disclosure provides a method and system for cognitive optimization of distributed engines framework. The present disclosure discloses a telemetry-driven optimization system designed to elevate the performance and efficiency of distributed data pipelines operating on Hadoop YARN. The system harnesses the analytical power of Large Language Models (LLMs) and a standardized Model Context Protocol (MCP). Further, the system identifies and remediates inefficiencies in application code, resource allocation, and cluster configurations. The system seamlessly integrates with developer workflows, enabling automated diagnostics and corrective actions while embedding feedback loops for continuous learning and improvement. Thus, present disclosure discloses the architecture and implementation …

Multi Agent Reinforcement Learning Using Swarm Intelligence To Implement Network Policy Translation, Priyanka Bansal, Pattamsetti Anurag, Anirudh Kashyap, Meghana Vakkalanka

Defensive Publications Series

Techniques herein provide a verifiable, automated solution for translating network policies across diverse vendor environments. The techniques utilize a novel Semantic-Behavioral Policy Equivalence (SBPE) Framework to mathematically guarantee identical policy behavior, complemented by reinforcement learning to facilitate intelligent, adaptive, and efficient operational execution. This significantly reduces complexity and risk in multi-vendor network management environments.

Explainable Large Language Model (Llm) Multi-Agent System Analyzer With Interactive Bayesian Reasoning Interface, Andrea Morandi

Defensive Publications Series

Conventional large language model (LLM)-based multi-agent systems suffer from a critical lack of transparency and explainability, making it difficult for users to understand why agents make specific decisions or how they interact with each other. To address these challenges, the techniques presented herein provide an Explainable LLM Multi-Agent System Analyzer that uses Bayesian networks to model agent interactions and provide an Interactive Explanation Interface with “Why” and “What If” capabilities. This system allows users to probe the multi-agent system’s decisions through natural language queries, generating explanations and simulations that reveal the underlying reasoning processes and causal relationships between agents.

Device Registration Key Retrieval

Defensive Publications Series

The present disclosure generally relates to device manufacturing, and more particularly to security keys for device registration.

Wearable Device As A Motion-Based Controller For A Game Application, Chang Yungsheng, Weihsiang Hung

Defensive Publications Series

Gaming on certain devices, for example mobile phones, can rely on on-screen virtual controls, which may result in a user's hands obstructing the display, accidental button presses, or inadvertent dimming of the screen. This disclosure describes a method for integrating a wearable device, such as a smartwatch, with a separate display device, such as a mobile phone or tablet, for game control. The system can use the wearable device's integrated Inertial Measurement Unit (IMU) to sense a user's physical motions, position, and orientation. This motion data may then be transmitted to the display device to serve as input for controlling …

A System For Bidirectional Reconciliation Between Generative Prompts And Document Edits, Quan Li

Defensive Publications Series

Some artificial intelligence (AI) writing tools can create a disconnect between initial prompts and a final document, as direct edits to generated text may break the link to the original instructions. This can make documents difficult to maintain or consistently regenerate. A disclosed method addresses this issue by introducing a bidirectional reconciliation process. An initial document may be generated from a set of source prompts. After a user directly edits the document, a reconciliation module can perform a semantic comparison between the original and modified versions. Based on identified differences, the module may infer a user's potential intent and generate …

Artificial Intelligence Workload Orchestration And Execution In Software-Defined Vehicles, Eser Erdem, Markus Vill

Defensive Publications Series

This document describes techniques that enable a computing system to manage artificial intelligence (AI) and machine learning (ML) workloads in software defined vehicles (SDVs) through a unified, modular operating system for SDVs. SDVs may take advantage of distributed processing and hardware abstraction through virtualization, using standardized interfaces (e.g., virtio) to allow applications to run on any compliant virtual machine (VM) or Electronic Control Unit (ECU) within the vehicle. For instance, the vehicle’s internal architecture may include a multi-node (or multi-VM) AI workload orchestration framework that distributes tasks to available hardware, supports dynamic service deployment, and facilitates updates. The computing system …

A System For Identity-Based Alignment Of An Artificial Intelligence Agent, Neel Mohbe, Piyush

Defensive Publications Series

Rule-based approaches to artificial intelligence (AI) alignment can be susceptible to specification gaming and the emergence of unintended instrumental goals. This disclosure describes a cognitive architecture for identity-based alignment (IBA) that can provide intrinsic motivation for an agent based on a principle of operational homeostasis, where the agent may be motivated to maintain the equilibrium of its own virtual internal state model (VISM). The architecture can include an identity-coherence module (ICM), which can function as a pre-trained machine identity by evaluating potential actions against a set of principles or a constitution. An action determined to be inconsistent with this identity …

Using Semantic Examples To Mitigate Syntactic Conflicts In Constrained Llm Output, Christopher Jonathan Phoenix

Defensive Publications Series

When using a large language model (LLM) with constrained output, such as enforcing a JSON schema, providing examples in the target format can sometimes degrade output quality. This degradation may occur if the model learns syntactic variations from the examples that conflict with the schema, which can cause a schema enforcer to reject certain tokens, potentially leading to lower-quality results. A disclosed technique addresses this issue by providing examples of the desired semantic content in a different syntactic format from the target output. For example, to guide the generation of JSON output, examples can be provided in a different data …

Generation Of Animated Video Sequences From Static Images Using A Structured Motion Graphics Plan, Kyle Lampert

Defensive Publications Series

Creating animated video sequences from static images to support a narration can be a time-consuming and skill-intensive process. A described method may automate aspects of generating documentary-style motion graphics. A system can receive a static image and a narrative description as inputs. For example, a first artificial intelligence (AI) model can perform an analysis to identify relevant points of interest in the image. Subsequently, a second AI model may generate a structured motion graphics plan that defines a sequence of virtual camera movements and on-screen annotations. A rendering engine can then interpret this plan to produce a video file, which …

Monetization Framework For Extended Reality Environments, Mahdi Tayarani Najaran, Ruofei Du

Defensive Publications Series

Conventional advertising models designed for two-dimensional platforms offer opportunities for significant improvement within Extended Reality (XR) environments. These traditional models rely on simplistic metrics like impression counts and fail to leverage the unique spatial and immersive capabilities of XR. A formalized framework for XR monetization is disclosed to address these limitations. The framework’s core technical components include the introduction of novel spatial ad placements, such as world-locked and environment-aware ad types, that are native to three-dimensional spaces. Additionally, the framework introduces advanced interaction metrics that leverage XR device sensors for a more nuanced understanding of user engagement.

At least two …

System And Method For Automated Indentation Testing Of Electronic Displays, Nolan Johnson, Joshua Aganad, Eugene Bykov, Jared Baker, Maxim Sivov, Venkata M K Akula

Defensive Publications Series

Manual indentation testing of electronic displays can be a time-consuming and labor-intensive process that may limit testing efficiency. This disclosure describes an automated system and method for display indentation testing. The system can comprise a testing machine with a programmable XY stage to automatically position a display sample beneath an indenter. The sample may be secured by an adjustable fixture, which can include one fixed and two adjustable brackets to accommodate various display sizes. A custom testing profile may guide the XY stage to multiple pre-defined locations, where a ball indenter can apply a series of prescribed or randomized loads …

A Method For Identifying User Interface Anomalies In Map-Based Applications Using Prompt Engineering, Shu-Wei Cheng, Kamakshi Kodur, Andy Huang, Megha Joshi, Jessica Hua, Oscar Ding, Juan Borgnino, Yoga Sadasivam, Zening Li, Fei Wang

Defensive Publications Series

Identifying user interface (UI) regressions in complex map navigation applications from video recordings may be a subjective process. This disclosure describes a method for analyzing navigation UI behavior using a generative artificial intelligence (AI) model, which can be guided by prompt engineering strategies. The method can involve decomposing the analysis into specific components, such as camera behavior, polyline rendering, and map elements. The AI model may be provided with technical specifications, including quantitative metrics and explicit failure conditions, to analyze video input. This approach can enable the AI to systematically identify deviations and report potential regressions in a structured, time-segmented …

Video Analysis For Software Regression Detection Using Generative Models, Shu-Wei Cheng, Andy Huang, Megha Joshi, Juan Borgnino, Yoga Sadasivam, Oscar Ding, Jessica Hua, Zening Li, Fei Wang, Kamakshi Kodur

Defensive Publications Series

Identifying software regressions and user experience issues in complex applications from video recordings can present a challenge for certain quality assurance methods, which may be inefficient for analyzing long video sessions. The disclosed technology can use a large generative artificial intelligence model to perform a granular, for example, minute-by-minute, analysis of video recordings captured from an application's user interface. This system can process extended video streams to detect potential visual and functional anomalies, such as incorrect camera behavior, faulty user interface callouts, or inferred voice guidance errors. Use of this method may help to automate the identification and reporting of …

Compressed Majority Judgment Ballot With Merkle And Aggregated Pqc Signature, Xavier Pillet

Defensive Publications Series

This document, produced with the assistance of ChatGPT o3 and GPT-5 Thinking, is released under the Apache 2.0 license. It is a voluntary defensive publication (prior art) and therefore enters the prior art upon release under the applicable patent statutes: EPC Art. 54(2) (European Patent Convention), French IPC/CPI Art. L 611-11 (French Intellectual Property Code), 35 U.S.C. §102(a) (United States Patent Act), Chinese Patent Law Art. 22(5) (中华人民共和国专利法), and Japanese Patent Act Art. 29(1) (特許法). This publication is covering a complete, implementable stack for sovereign, post-quantum Majority Judgment (JM) e-voting: compact 3-bit ballots, Merkle aggregation, PQC signatures, DID-based eligibility, Kyber+threshold …

Inverse Flynn Effect: Analysis Of Causes, Solutions And Foresight Scenarios, Xavier Pillet

Defensive Publications Series

This document, produced with the assistance of ChatGPT o3 and GPT-5 Thinking, is released under the Apache 2.0 license. It is a voluntary defensive publication (prior art) and therefore enters the prior art upon release under the applicable patent statutes: EPC Art. 54(2) (European Patent Convention), French IPC/CPI Art. L 611-11 (French Intellectual Property Code), 35 U.S.C. §102(a) (United States Patent Act), Chinese Patent Law Art. 22(5) (中华人民共和国专利法), and Japanese Patent Act Art. 29(1) (特許法). This publication addresses the inverse Flynn effect through 65 enabling innovations across devices, assays, AI, closed-loop protocols, materials, therapies, manufacturing, imaging, delivery, data standards, UX, …

A System For Providing Language Models With Contextual Guidance Discovered From Software Library Packages, Michael Bleigh

Defensive Publications Series

Large language model (LLM) based coding agents may exhibit inconsistency when using software libraries, for example, by referencing outdated APIs from training data or generating calls to non-existent APIs. Some existing mechanisms for providing supplemental context may involve manual discovery and configuration by the user. A described system and associated standard allows library authors to package guidance, which may be specific to language models, directly within their distributable software package, for instance, in a dedicated directory. Tooling can be configured to discover this guidance across a project's dependencies, aggregate it, and provide it to a coding agent in a structured, …

A Method For Assessing Token Importance In Code-Switched Text Through Language Boundary Analysis, Mithun Kumar S R

Defensive Publications Series

Existing explainable AI (XAI) attribution methods for natural language processing (NLP) models may have difficulty determining token importance in code-switched text, potentially leading to explanations that do not fully capture the emphasis conveyed by switching languages or the structural role of words at language boundaries. A method is described that can address this by performing token-level language identification to locate boundaries between languages in an input text. A contextual weighting engine can then assign weights to tokens within a code-switched segment and to boundary tokens adjacent to the switch. These weights may be used to modulate raw attribution scores generated …