Digital Commons Network^™

Optimizing User Experience By Continual Automated Assessment Of Usability Metrics Obtained Using A Large Language Model, Johannes Start, John Lunney

Defensive Publications Series

Current software testing methods and metrics do not provide a comprehensive, objective, and automated assessment of the UX in real time, thus making it challenging to be proactive about identifying and addressing usability issues that have a negative impact on the UX and business goals. This disclosure describes techniques that leverage the capabilities of large language models (LLMs) to provide a deep, comprehensive, and nuanced understanding of the UX of a software solution that takes into account the user’s intent and context. Preprocessed data about interaction analytics and contextual information obtained with user permission can be analyzed via an LLM …

Evaluating Artificial Intelligence (Ai) Agents In Multi-Human Environments, Na

Defensive Publications Series

Artificial intelligence (AI) agents typically serve a single human as a client, but such a single human focus by an AI agent can be sub-optimal in contexts where the AI agent performs tasks for teams or other groups of humans. This disclosure describes techniques to develop and evaluate AI agents capable of robust operation in multi-human environments. In a data collection phase, group chat or virtual meeting experiments are organized where participants roleplay conversations to accomplish predefined tasks such as scheduling a meeting, planning an event, etc. Communication transcripts from such experiments capture the complexities of group interactions. In a …

Generating Personalized Multimedia Content By Prompting Models With Location And Context Specific Inputs, Shiblee Hasan, Joseph Johnson Jr

Defensive Publications Series

Currently, generating content tied to specific locations via genAI models can often result in hallucinations of non-existent places or creation of generic content that is devoid of the target specifics. The difficulties in generating content for specific locations result from the genAI models being trained on massive datasets containing images from many locations without sufficient focus on a single location. This disclosure describes techniques to enhance the creation of location-specific personalized 2D or 3D immersive or non-immersive multimedia experiences (e.g., 2D or 3D panoramic video with or without sound) via multimodal genAI models. Base genAI models are boosted with multimedia …

Reducing Shot Cuts In Diffusion-Generated Video, Mitchell Mcintire, David Yao, Orly Liba

Defensive Publications Series

This publication describes techniques for reducing shot cuts in videos generated by diffusion models, particularly in image-to-video applications. Shot cuts are abrupt transitions between distinct scenes of video. Shot cuts may be unexpected and may be irrelevant to provided input. To reduce shot cuts, a diffusion model may begin a diffusion process from a static video, where the frames of the static video may be noised instances of an input image. The diffusion model may also insert the input image as a first frame of the generated video, anchoring the output to the input. By ‘anchoring’ the output video and …

Document-Level Understanding For Browser Contexts, Hp Inc

Defensive Publications Series

This article presents a browser-native assistant for advanced contextual understanding of web content. The system features real-time DOM analysis, semantic saliency scoring, and structured memory, enabling users to query and summarize web pages in natural language. By leveraging real-time browser events, the assistant dynamically adjusts context relevance to improve the accuracy and relevance of its responses.

A Motherboard Architecture For Memory Module Lighting Synchronization Via Embedded Controller Pwm Output, Hp Inc

Defensive Publications Series

This invention relates to lighting synchronization control technology among multiple memory modules on a computer motherboard, particularly through a single PWM signal generated by the embedded controller (EC) and transmitted to each memory module's lighting controller.

Smart Ac Plug Load Switcher “Smart Plug”, Hp Inc

Defensive Publications Series

Present public power generation systems are designed to deliver a set voltage level to their end loads (for example, in computer systems). The current draw (and thus the load Wattage) is determined by the load that the end device is designed to. Common consumer power outlets are sized to 120 VAC with either 15 and 20 Amp (USA/NEMA example) service, for example. Safety is attained via a breaker/fuse in the distribution system. If the load device exceeds the delivery system’s capacity, the safety device is tripped. This all works well when the load device’s power is fixed. However, new systems …

Iot Dock-Based Prox-Activate Workspace, Hp Inc

Defensive Publications Series

This invention discloses a system and method for automatically activating a workspace environment using a combination of GPS, mobile applications, cloud services, and an IoT-enabled docking station. When a user enters a predefined geographic area near their workspace, the system detects the user's proximity via GPS and triggers a wake-up signal to the user's notebook (NB) through a Type-C connection. This solution enhances productivity by reducing boot-up delays and enabling seamless transitions into the work environment.

Automating Geospatial Analysis Using Large Language Models And Code Execution In A Persistent Environment, Patrik Blohmé, Ananya Gupta, Gearoid Murphy, Sameera Ponda, Danusch Merrikh-Yazdi, Gia Jung, Renee Johnston, Marcin Sieniek

Defensive Publications Series

Sustainability professionals, urban planners, clean energy developers, and other professionals encounter substantial challenges in efficiently analyzing and utilizing geospatial data for decision-making related to location opportunities, accessibility, and asset placement. Traditional approaches involve complex, multi-step workflows that require specialized expertise and time-consuming analysis across disparate data sources, resulting in substantial delays before actionable insights can be derived. This disclosure describes techniques for automating geospatial workflows over multiple datasets from varied sources, facilitated by large language model (LLM) agents. The techniques enable users to efficiently query geospatial data and receive insights through an AI-driven process that minimizes complexity and reduces time …

P250307e_Uss0068_Twip Defpub Saws As Acf, Matched Filter, And Phase Detection, Christian A. Mohr, Mark Rode, Marco Dannemann

Defensive Publications Series

This invention presents a method and device for real-time analog signal processing using Surface Acoustic Wave (SAW) technology. The system performs three core functions without digital conversion: (1) Autocorrelation Function (ACF) for time-frequency transformation, (2) Matched Filtering for pattern recognition, and (3) Phase Detection of modulated signals.

A SAW chip with interdigital transducers (IDTs) on a piezoelectric substrate decomposes input signals into parallel frequency channels ("frequency bins"). Envelope detection of each channel’s output generates an analog spectrogram. Weighting these outputs with predefined coded patterns (via analog operational circuits) enables instantaneous cross-correlation, acting as a matched filter. For phase-sensitive detection, in-phase/quadrature …

P250306e_Uss0073_Twip Defpub Freeform- Phase-Value - Parametermatrix For Sending/Receiving Array, Christian A. Mohr, Mark Rode, Marco Dannemann

Defensive Publications Series

The invention relates to an ultrasonic measurement system comprising: a phased-array transmit antenna (1) comprising a plurality of ultrasonic transducers (PMUTs), a phased-array receive antenna (2), wherein the receive antenna (2) is distinct from the 5 transmit antenna (1), a control unit (3) configured to assign to each transducer of the transmit antenna (1) a signal with an individually adjustable phase value, wherein the phase values are defined by a transmit parameter matrix (P_TX), wherein the control unit (3) is further configured to transmit multiple orthogonal codes 10 (C₁, C₂, ..., Cₙ) simultaneously superimposed on the same frequency band, with …

Fictitious Data-Based Watermarking System

Defensive Publications Series

Systems and methods are provided for watermarking via unique synthetic datasets. A candidate model may be trained with a poison dataset including synthetic information. The synthetic information may relate to a fictitious place, person, or scenario, for example. A set of one or more inference questions associated with the poison dataset may be determined. The candidate model may be tested based on the set of one or more inference questions. A confidence may also be generated. The confidence may be associated with one or more identification metrics associated with the candidate model.

Method, System, And Computer Program Product For Large Language Model (Llm)-Enabled Searching, Jia Yi Kuo, Jianyang Huang, Karthik Gopalakrishna, Bhuvanjeet Singh Gandhi, Paromita Guha, Ashutosh Kumar Chandel, Arpita Agrawal, Ritu Saxena

Defensive Publications Series

A method includes: chunking each resource into chunks; inputting the chunks into a model; for each chunk, generating an embedding vector; receiving a query message from a user device; inputting the query message into the model; generating a query embedding vector for the query message; executing a semantic search on the query embedding vector by: inputting the query embedding vector and the embedding vector for each chunk to the model; for each chunk, generating a chunk similarity score; determining a subset of chunks based on the chunk similarity score for each chunk of the resource; inputting the subset of chunks …

Multi-Function Obd-Ii Device With Corrosion Protection, Gps Tracking & Anti-Theft Features, Vincent Gallo

Defensive Publications Series

An OBD-II plug-in device unifies active impressed-current corrosion protection, GPS-based tracking and geofencing, anti-theft immobilization, environmental sensing, and cellular/Wi-Fi connectivity for telematics and over-the-air firmware updates. The integrated system draws power and data from the vehicle’s OBD-II port and employs a galvanically isolated current regulation circuit to deliver adaptive cathodic protection to the chassis. A central microcontroller coordinates location monitoring, starter-relay control, tamper detection, and cloud communication while enabling secure firmware upgrades. Environmental sensors log temperature, humidity, and impact events to refine corrosion control and trigger security responses. This multi-function design simplifies installation, eliminates redundant hardware, and ensures comprehensive vehicle …

Automating Tasks In Video Games Via Intelligent Agent, Ilia Labzovsky, Danny Karmon

Defensive Publications Series

This document describes techniques that enable a computing device (e.g., a personal computer, laptop, smartphone, smart watch, server system, gaming console, tablet, wearable device, virtual reality (VR) headset, augmented reality (AR) glasses, etc.) to automate repetitive tasks in video games using an intelligent software agent (e.g., artificial intelligence (AI), machine learning model, etc.). Video games often necessitate repetitive, time-consuming tasks for player progression, leading to user fatigue and diminished enjoyment (e.g., loot farming, resource gathering, experience point (XP) grinding, repetitive combat encounters). In some instances, the player may have to complete the same tedious tasks multiple times to reach engaging …

Augmenting Conversational Artificial Intelligence Using Contextual Signals, Dongeek Shin

Defensive Publications Series

This document describes techniques that enable a mobile computing device (e.g., a smartphone, smart watch, wearable device, laptop computer, etc.) and an audio device (e.g., earbuds, headset, augmented reality glasses, etc.) to enhance conversational artificial intelligence (AI) with contextual signals (e.g., location data, ambient light, motion signals, etc.). For instance, the mobile computing device may be paired with the audio device, to which a user provides voice input to engage with the AI. However, the AI’s responses often rely on spoken commands and prior conversation history, which may reduce the AI’s understanding of the user's full context. The techniques of …

Custom Thin-Coat Designs For Reflective Waveguides In Augmented Reality Wearables, Jun Yang, Alexander Koshelev, Christophe Peroz

Defensive Publications Series

The use of reflective waveguides in augmented reality (AR) displays can reduce see-through transmission, such that a person facing the user can perceive an uneven brightness of the user’s face near the eye region. Uneven face brightness can cause some social and cosmetic discomfort. This disclosure describes a family of graded thin-film coats for reflective waveguides with multiple SKUs, each tailored to serve users of specific skin tone, color, and/or texture. The coating for the eye-pupil expander (EPE) and output coupler (OC) regions of the lens are tuned to achieve certain reflectance and wavelength-dependence criteria that render the prisms on …

Semantic Policy Summaries And Semantic Explanations For Explainable Ai In Networking, Dennis Lanov

Defensive Publications Series

Proposed herein is an explainability technique, referred to herein as a "Semantic Policy Summaries and Semantic Explanations" technique, for artificial intelligence (AI) -driven networking that translates complex model decisions into clear, human-readable explanations grounded in real-world network concepts. Instead of abstract outputs like feature scores, the proposed technique delivers goal-oriented justifications for AI network-related decisions/actions. The proposed technique integrates directly into a control loop, supports policy-based triggering, and logs explanations with decision metadata for audit and regulatory review. Further, the proposed technique aligns with requirements from the European Union (EU) AI Act and the proposed United States (U.S.) SAFE Innovation …

Gain Movie Insights Using Generative Artificial Intelligence, Akash Verma, Vineet Kumar, Abhay Kumar Gupta, Vivek Yadav, Szu-An Lu, Meng-Ting Tsai, Ya Hsien Liu, Chu-Feng Lien

Defensive Publications Series

A television (TV) application may present various types of media content of interest to a user. The media content may have different formats such as streaming video and audio. The types of media content may include, but are not limited to, movies, television shows, live sporting events, news items, short form videos, and music. In addition, or in the alternative, a variety of media content providers may deliver various types of media content for viewing by the user. The TV application may deliver a customized viewing experience to a user that spans the diverse types of media content provided by …

Enhancing 5g Non-Terrestrial Network (Ntn) Communications With Machine Learning-Driven Non-Line-Of-Sight (Nlos) Prediction And Mitigation, Alwin Xavier, Wojciech Grobel, Snezana Mitrovic, Vinay Saini

Defensive Publications Series

The innovation proposed herein leverages machine learning to predict Non-Line-of-Sight (NLOS) conditions in a 5G Non-Terrestrial Network (NTN), significantly enhancing communication reliability in such networks. By dynamically encoding these predictions into user equipment (UE) Route Selection Policies (URSPs), the innovation proposed herein optimizes data transfers, ensuring efficient use of satellite coverage and improving overall network performance with minimal changes to the existing 5G core infrastructure.

Multi-Agent Architecture With Central Intelligent Agent, Shashwat Anand, Rohan Chaudhury

Defensive Publications Series

This disclosure describes a multi-agent architecture featuring a centralized intelligent agent (orchestrator agent) for complex enterprise environments, e.g., partner interfaces for a cloud computing solutions provider. While agent-driven interfaces are feasible, there are challenges in agent discovery and coordination, leading to fragmented solutions and slower pace of development. By providing a central agent that coordinates with various domain-specific functional agents, the techniques of this disclosure provide a cohesive mechanism for managing and orchestrating agents. The agent is built to service various user personas and uses this context to provide high-fidelity assistance. Agents are built with awareness of access control lists …

Real-Time Context-Aware Assistance For Non-Native Language Speakers, Jaime Guajardo, Sriratana (Daowz) Sutasirisap

Defensive Publications Series

When two or more non-native language speakers or learners converse, they may often encounter words or grammatical structures they do not understand. In the absence of a native speaker, learners tend to pause their conversation to look up unfamiliar words in a dictionary, causing disruption to their conversation flow. This disclosure describes techniques that assist non-native language speakers or learners during real-time conversations without disrupting conversational flow. With user permission, an ongoing conversation is analyzed in real time by a large language model (LLM) based language assistant to identify non-idiomatic speech and grammar, vocabulary, and/or pronunciation errors. When a user …

Location-Specific Artificial Intelligence Models, Ajay Prasad

Defensive Publications Series

Virtual assistants powered by large language models (LLMs) are capable of interacting with users via chat, voice, or other interfaces and can answer arbitrary queries from the user. However, LLMs may not have access to updated location-specific information. This disclosure describes techniques for integrating location-specific AI models into virtual assistants. This allows virtual assistants to provide hyperlocal information and personalized responses by switching between general-purpose and specialized models with user permission. Location-specific models are trained with data relevant to particular places or businesses. The described techniques enhance user experience by providing accurate and up to date location-specific information appropriate to …

Simple And Cost-Effective Quantum Safe Anonymous Communications, Niranjan M M

Defensive Publications Series

Techniques presented herein propose method to establish quantum secure anonymous communication (e.g., in TOR networks) without having dedicated quantum secure hardware, also considering simple and cost-effective method. As per the presented method, Client<->Entry node, Client<->Middle node, Client<->Exit node exchanges secret seed value and PQPSK ID using McEliece algorithm (using Public-Private Key pair) over already established authenticated secure communication channels using 1W-AKE or hybridOR. The secret seed value and PQPSK ID shall be used to generate unique pair of PQPSK (Key, ID) using Quantum Key Generation service. At the end of all negotiation between the Client and the …

Interface To Control Gps Search While A Device Is In Satellite Communication Mode, Ashwin Pandith, Andy (Jibing) Wang, Nidhi Gulia, Sharath Ananth, Chinmay Dhodapkar

Defensive Publications Series

This disclosure describes techniques that enable explicit control over a user equipment (UE) modem to determine the start and the stop of GPS acquisition. An indication is provided that GPS acquisition is about to start, which is useful under circumstances such as: the UE is pointing to a non-terrestrial network (NTN) satellite awaiting a response; the UE is idle but in clear view of the sky; the user equipment is in a location with poor GPS satellite visibility; etc. The techniques employ user-provided information about the UE status (pointing to NTN satellites; in clear view of sky or not; etc.) …

Scope 4 Impact And Reporting Methodology, Marisol Palmero, Snezana Mitrovic, Suresh Krishnan

Defensive Publications Series

The concept of Scope 4 emissions, though not officially recognized in the Greenhouse Gas Protocol, presents a strategic opportunity to enhance sustainable operations. By focusing on "avoided emissions", based on good or best practices, organizations can influence and reduce Scope 2, and even Scope 3 emissions across their value chain. Techniques are proposed herein that seek to reduce energy consumption from networking devices. The proposed techniques could be extended to other practices within organizations. By reducing energy consumption, there is a direct link to improve energy efficiency, optimize resource usage, and ultimately contribute to an organization's overall sustainability goals. The …

A Hybrid Ai-Powered Voice-Enabled Medical Practice Assistant With Dynamic Triage Routing And Multi-Modal Patient Interaction, Patrick Proctor

Defensive Publications Series

We present a novel implementation of a voice-enabled medical practice virtual assistant that combines real-time speech processing, intelligent medical triage routing, and multi-modal patient interaction capabilities. Our system employs a hybrid AI architecture that dynamically selects between different language models based on query complexity classification, reducing operational costs while maintaining high-quality responses for complex medical inquiries. The system incorporates a deterministic medical triage decision tree based on established nursing protocols, enabling automated routing of urgent cases while maintaining HIPAA compliance. Additionally, we introduce an innovative multi-step interaction pattern for location-aware services, seamlessly transitioning between voice input, location detection, and turn-by-turn …

Low-Bandwidth, High-Sensitivity Telemetry Signal To Optimize Osc Launch Power During Deployment And Regular Operation, Anonymous

Defensive Publications Series

The present disclosure relates to methods and systems employing a low-bandwidth, high-sensitivity telemetry signal integrated into optical line systems to dynamically optimize the launch power of an Optical Supervisory Channel (OSC) during both deployment and normal operation. Traditionally, OSC power management faces challenges such as receiver overload on short spans and insufficient received power on longer spans, requiring manual provisioning. The disclosed telemetry signal independently and reliably measures optical span loss across an extended range (up to approximately 50 dB), significantly exceeding the measurement capabilities of conventional OSC signals. Using telemetry-derived span-loss data, OSC launch power can be automatically and …

Artificial Intelligence Assisted Disaster Preparation And Response, Mel Bass, Daniel V. Klein

Defensive Publications Series

During natural disasters, entire regions of communication networks can be disabled quickly. Impacted areas can be struck again shortly thereafter. Wildfires, blizzards, floods, and earthquakes present unique preparation and response challenges. A rapidly changing emergency event makes it difficult to communicate, provide accurate and timely status updates, triage responses, etc. This disclosure describes artificial intelligence (AI) techniques that aid in disaster preparation and response. Implemented with specific user permission and on devices that are configured in emergency mode, the techniques can enable tracking of people displaced by disasters and keeping families up-to-date with the status of their loved ones. The …

A System For Predictive Crowd Management Using Coupled Analysis Of Crowd State And Environmental Information, Mithun Kumar S R

Defensive Publications Series

Some crowd management systems analyze crowd dynamics, such as density and flow, in isolation from the physical environment. The disclosed system can couple a Dynamic Crowd State Vector, which may quantify real-time crowd metrics, with a Hyper-Local Condition Vector, which may quantify environmental factors such as surface integrity, passage width, and lighting. A machine learning model can process the real-time interaction between these two vectors to generate a unified, predictive risk score for a specific geographic zone. The resulting coupled analysis may facilitate the proactive identification of context-specific hazards, such as moderate crowds in some conditions, which may aid in …