Open Access. Powered by Scholars. Published by Universities.®
- Publication Year
- File Type
Articles 3421 - 3450 of 11865
Full-Text Articles in Entire DC Network
Dynamic, Prompt-Aware Orchestration And Lazy Loading Of Ai Model Context Protocol (Mcp) Servers For Large Language Model (Llm) Applications, Hp Inc
Defensive Publications Series
This idea introduces an architecture for dynamically discovering, automatically installing, loading, and orchestrating Model Context Protocol (MCP) servers in response to user prompts in AI-driven applications. Unlike traditional MCP implementations that preload all available servers at startup, this approach enhances the user experience by automatically discovering missing MCP servers, auto-installing, and loading the MCP server with the AI application to accomplish the desired outcome for the user. This technique allows natural language and intent to determine the necessary MCP servers needed and abstracts the rest of the complexities for the user, allowing them to focus on prompting naturally.
Low-Cost Media Size Sensing For Printers, Hp Inc
Low-Cost Media Size Sensing For Printers, Hp Inc
Defensive Publications Series
This disclosure relates to media size detection in printers, particularly when the input tray is located at the bottom of the printer. Conventional implementations often detect media size by sensing the position of the width adjustor, sometimes through the angular position of a gear linking the left and right adjustors. In removable trays, adding electronics directly to the tray is costly due to the expense of providing a reliable connection to the printer’s main printed circuit assembly (MPCA).
Learning Temporal Embedding Of Telemetry Via Neural Odes, Hp Inc
Learning Temporal Embedding Of Telemetry Via Neural Odes, Hp Inc
Defensive Publications Series
The invention provides a modular machine learning pipeline that preprocesses structured telemetry logs, applies attention over tokenized key–value pairs, uses a hybrid Neural ODE/SDE-RNN architecture to evolve the embedding over time, employs self-supervised learning for training, supports both offline and online updates, and generates natural language summaries via a language model. Effectively each user’s environment is represented as an embedding which captures the time evolution
Supervisory Recovery Process For Windows Devices With Dynamic Error Handling, Hp Inc
Supervisory Recovery Process For Windows Devices With Dynamic Error Handling, Hp Inc
Defensive Publications Series
The supervisory recovery method addresses these challenges by offering flexible restoration options, including automatic retries, retry from, failure impact, alternative solution, failure dependency, skip functions, and repairs for unexpected issues after OOBE. This approach minimizes manual steps, streamlining the recovery process and increasing the likelihood of a successful outcome.
Platform Validation Framework Sequence Diagram Conformance, Hp Inc
Platform Validation Framework Sequence Diagram Conformance, Hp Inc
Defensive Publications Series
This disclosure describes a framework for clearly demonstrating that an SDK provider and Solution Clients in a distributed system conform to spec via exacting documentation utilizing sequence diagrams. The framework ensures that sequence diagrams which are shared with Solution Clients can be validated independent of release schedules.
Shielding With Dissipation Solution, Hp Inc
Shielding With Dissipation Solution, Hp Inc
Defensive Publications Series
Traditionally, shielding can is an effective way for blocking RF interference from circuits and chips on PCB to impact antenna. Ideally it should be well grounded, and the gaps that may cause leakage should be as small as possible. However, some cases have limited grounding points which resulted in relatively large gaps, the leakage of RF interference could possibly be bad, and thus further absorber is required to suppress RF interference. The idea here is to apply resistors between shielding can and PCB ground to act as absorber or damping to absorb unrequired RF energy.
Dual Radio Headset, Hp Inc
Dual Radio Headset, Hp Inc
Defensive Publications Series
When a user using a Bluetooth Headset (call/music) and moves away from the Bluetooth Host, he experiences the audio becoming choppy or completely lost depending upon the various factors which culminates into Bluetooth range/coverage. Some headsets can operate up to 100 feet/ 30 m from the host by switching from class 2 to class 1 operation (based on proximity to the host) this helps to optimize the battery life.
Cross-Vendor Print Quality Diagnostics Via Genai, Hp Inc
Cross-Vendor Print Quality Diagnostics Via Genai, Hp Inc
Defensive Publications Series
Our idea, Print Quality Diagnostics with GenAI, employs proprietary prompt engineering techniques and specialized large language model (LLM) integration methods that create a unique approach to print quality diagnostics. By utilizing carefully crafted LLM prompts and specialized system architecture, our product enables cross-vendor compatibility, going beyond existing solutions which tend to be vendor specific. The system integrates defect classification, reasoning explainability, and remediation generation into a unified architecture. Our solution dramatically reduces model development time while maintaining diagnostic accuracy comparable to traditional models.
Vision Ai-Based Audio Transcription Enhancement, Hp Inc
Vision Ai-Based Audio Transcription Enhancement, Hp Inc
Defensive Publications Series
This article describes an AI-based enhancement for audio transcription that utilizes video data to improve accuracy and context during conference calls. By analyzing participant gestures, facial expressions, and non-verbal actions, the technology aims to create a more engaging communication experience for both in-person and remote participants with an enhanced audio transcription that captures deeper levels of communication and a detailed collaboration context in a meeting.
System For Generating Immersive Experiences Via An Ai Pipeline And Sensory Protocol, Gaurav Umeshkumar Gupta
System For Generating Immersive Experiences Via An Ai Pipeline And Sensory Protocol, Gaurav Umeshkumar Gupta
Defensive Publications Series
Current virtual reality (VR) content creation is complex, and the resulting sensory feedback from haptic devices relies on proprietary, pre-programmed effects. This limits the availability of personalized, deeply immersive experiences and creates a fragmented hardware ecosystem. The disclosed technology describes a method for generating multi-sensory VR experiences from user prompts. It uses an artificial intelligence pipeline to generate a narrative, visual assets, and a corresponding synchronized sensory track. A key component is an open communication protocol that enables standardized communication between any VR content and various haptic hardware devices. This approach facilitates the on-demand creation of cohesive and immersive experiences, …
Foldable Device Support Plate Configured To Reduce Inner Display Failure From Impact, Pj Wang
Foldable Device Support Plate Configured To Reduce Inner Display Failure From Impact, Pj Wang
Defensive Publications Series
This publication describes a novel support plate bending structure for foldable devices that addresses the problem of inner display damage due to Z-axis movement during impacts. The inner display may be damaged from hitting internal structures upon being dropped. This invention introduces a bending structure engineered into the support plate which automatically inserts into the gap between the arm and bracket when the device is in the closed configuration. Acting as a critical stopper, this feature mitigates system Z-axis movement, thereby reducing the chances of inner display failures caused by relative motion during Z-direction impacts. This technique requires no extra …
Identifying Conversation Segments Of An Audiovisual Stream, Tongzhou Chen, Kartik Audhkhasi
Identifying Conversation Segments Of An Audiovisual Stream, Tongzhou Chen, Kartik Audhkhasi
Defensive Publications Series
This disclosure describes lightweight techniques to determine segments of an audio or audiovisual stream that include conversations. Per the techniques, a voice activity detector (VAD) isolates super-segments, e.g., segments of relatively long duration. Super-segments are split into smaller segments and mapped into an embeddings space. The embeddings obtained from the smaller segments are clustered. A chunk of video is determined to include a conversation if its length exceeds a certain threshold T and the number of segments it includes exceeds a certain threshold S; it has at least a certain number N of major clusters; and at least a certain …
Agentic Workflow For Root-Cause Analysis And Fix Suggestions Based On Log Differencing, Shubham Khurana, Shreya Rana, Aastha Jain, Aneesha Jain
Agentic Workflow For Root-Cause Analysis And Fix Suggestions Based On Log Differencing, Shubham Khurana, Shreya Rana, Aastha Jain, Aneesha Jain
Defensive Publications Series
Integration testing in continuous integration/development environments produces large volumes of logs containing many errors and warnings, many of which are benign or unrelated. Identifying the specific error causing a test failure amidst this noise is a major challenge for developers seeking efficient and accurate debugging. This disclosure describes techniques that leverage large language models (LLMs) to automate and abstract root-cause analyses for test failures. A sequential, multi-step pipeline comprising artificial intelligence (AI) agents accepts as input a reference to a failing test run and produces a high-value, actionable output, e.g., a summary of the root cause and a suggested code …
Interactive Intelligent Agent For Gaming Assistance, Christopher James Tegethoff
Interactive Intelligent Agent For Gaming Assistance, Christopher James Tegethoff
Defensive Publications Series
This document describes techniques that enable a computing device (e.g., a personal computer, laptop, smartphone, smart watch, server system, etc.) to provide assistance and/or companionship to a player of a game application using an intelligent software agent (e.g., artificial intelligence (AI), machine learning model, etc.). Gaming environments may present challenges such as difficult learning curves, slow progression, complicated/unintuitive mechanics, and/or hostile social interactions. The intelligent agent may mitigate these challenges by acquiring game knowledge (e.g., via observing and learning from the player, “playing” the game on its own, training on tutorials and/or other instructional material, etc.) and offering in-game guidance …
Correlating Cross-Service Behavioral Signals To Identify Coordinated Inauthentic Activity, Mithun Kumar S R
Correlating Cross-Service Behavioral Signals To Identify Coordinated Inauthentic Activity, Mithun Kumar S R
Defensive Publications Series
Some systems for detecting fraudulent user-generated content (UGC) can be siloed and reactive, analyzing individual pieces of content or account metrics in isolation, which may leave them susceptible to coordinated campaigns spanning multiple services. A system and method are described that can construct a unified cross-service behavioral graph to address such challenges. This graph can ingest and correlate metadata from user activities across disparate services, for example, productivity, payment, and communication platforms, to identify precursor signals that may be indicative of fraudulent intent. By analyzing patterns such as co-reviewing, temporal pulsing, and suspicious cross-service coordination (e.g., template creation), the system …
Generating Financial Risk Tiers For Geo-Located Entities Using Monetary Claims From User-Generated Content And Transactional Signals, Mithun Kumar S R
Generating Financial Risk Tiers For Geo-Located Entities Using Monetary Claims From User-Generated Content And Transactional Signals, Mithun Kumar S R
Defensive Publications Series
Moderation systems for local business platforms may have difficulty identifying certain types of financial fraud, such as overcharging or fake deposit scams, because of challenges in validating monetary claims found in user-generated content (UGC). A disclosed method can address this potential limitation by synthesizing two different data streams. A system can use technologies, for example natural language processing and computer vision, to extract unstructured financial claims and pricing information from UGC, which may include content such as text reviews and photos of menus or receipts. The system can then correlate these claims with structured, verified transactional signals from integrated payment …
Battery Enclosure Panel And Associated Means Of Manufacture, Thomas J. Beck, Rahul Vilas Kulkarni, Xavier Varone
Battery Enclosure Panel And Associated Means Of Manufacture, Thomas J. Beck, Rahul Vilas Kulkarni, Xavier Varone
Defensive Publications Series
A battery enclosure panel may include a formable inner sheet having a generally planar outer surface and at least one recessed portion formed in the generally planar outer surface that defines a receiving area for a fire suppressant material, and an outer sheet at least partially joined to the generally planar outer surface of the inner formable sheet.
Predictive Capacity Planning: Estimating Hardware Requirements For Customer Deployments, Mahesh Chidrawar, Sanjeev Panem Jaya, Umesh Prabhakar Gaikwad
Predictive Capacity Planning: Estimating Hardware Requirements For Customer Deployments, Mahesh Chidrawar, Sanjeev Panem Jaya, Umesh Prabhakar Gaikwad
Defensive Publications Series
Achieving suitable dimensioning for accurately determining the hardware specifications necessary for new mobile gateway deployments is a complex challenge, involving numerous variables whose interrelationships must be understood in the context of specific use cases. This research aims to develop a model for estimating the necessary capacity for product dimensioning and deployments on existing hardware infrastructure. By leveraging data from current customer labs, the study seeks to facilitate effective capacity planning for clients. The model will be constructed through a comprehensive analysis of various customer datasets, enabling the prediction of the hardware requisites tailored to specific deployment needs. The objective is …
Simulating Network Security Incidents Using Large Language Models, Achuth Narayan Rajagopal, Naveena Mohan
Simulating Network Security Incidents Using Large Language Models, Achuth Narayan Rajagopal, Naveena Mohan
Defensive Publications Series
Simulated replays of cyberattacks can ensure the effectiveness of defensive measures taken against the cyberattacks. However, there are no reliable automated techniques to recreate or replay a cyberattack. This disclosure describes techniques to synthetically generate network traffic to recreate, replay, or simulate a cyberattack or other malicious actions on the network. The techniques enable a network administrator to textually describe a malicious scenario, which can be fictitious or can be based on real cyberattacks that occurred in the recent past. A large language model (LLM) is leveraged to generate synthetic traffic that matches the textually described scenario to test and …
Automated Self-Improving And User-Responsive Video Chapter Generation Using Generative Artificial Intelligence, Ilia Labzovsky, Danny Karmon
Automated Self-Improving And User-Responsive Video Chapter Generation Using Generative Artificial Intelligence, Ilia Labzovsky, Danny Karmon
Defensive Publications Series
Traditional video segmentation methods that offer limited granularity, static chapter structures. Such solutions lack deep semantic understanding and do not incorporate user interaction data for video segmentation, leading to inefficient information discovery and a poor user experience. This disclosure describes the use of generative artificial intelligence (genAI) to automatically create and dynamically refine smart chapters for long-form videos. These chapters are rich in detail, including titles, summaries, key concepts, and preview clips. With user permission, user interaction data can be used to dynamically adapt the chapter structure for individual viewers or the overall population. The techniques transform static videos into …
Efficient Photo Compression And Reconstruction Using Generative Artificial Intelligence, Amir Kamali, Hossein Talebi, Prachi Panpalia, Nick Torres
Efficient Photo Compression And Reconstruction Using Generative Artificial Intelligence, Amir Kamali, Hossein Talebi, Prachi Panpalia, Nick Torres
Defensive Publications Series
Storing high-resolution digital photos and video requires substantial storage space on a user device and/or a cloud-based image library. This disclosure describes the use of generative artificial intelligence to achieve storage savings via image downscaling with little or no perceptible impact. With user permission, a customized model such as a LoRA patch is trained based on images in the user’s library. Images in the library are stored as downscaled versions. When the user requires a high-resolution version, a generative artificial intelligence model, optionally with the customized LoRA patch, is used to generate the target resolution version from the downscaled image. …
Augmenting Photo Reminiscing Experiences With Llm-Generated Narrations, Ji Hun Kim, Abdulrahman Khan, Tommy Meaney, Tim Zhu, Zhou Zhou
Augmenting Photo Reminiscing Experiences With Llm-Generated Narrations, Ji Hun Kim, Abdulrahman Khan, Tommy Meaney, Tim Zhu, Zhou Zhou
Defensive Publications Series
This disclosure describes techniques that leverage a large language model (LLM) to generate custom narrations when arranging sets of photos in the form of a visual experience for users to reminisce on thereby personalizing and adding depth to the reminiscing experience. Per the techniques, a set of photos from a user’s photo library to be featured in a visual experience is identified. Photo pixels, metadata, and other relevant data are used to populate a prompt that is to an LLM. The response from the LLM is parsed and associated with photos to be featured in the visual experience. The LLM-generated …
User Intent-Conditioned Image Captioning Using Llm For Generating Personalized Photo Descriptions, D Shin
User Intent-Conditioned Image Captioning Using Llm For Generating Personalized Photo Descriptions, D Shin
Defensive Publications Series
Image captioning models can receive an image as input and generate captions that describe objects, scenes, and other visual aspects in the image. However, such models are not personalized to a user and lack the ability to generate captions that are tailored to specific user intent or contexts. This disclosure describes techniques to generate personalized image captions using a large language model (LLM). With user permission, user intent data such as context, preferences, notes, or tags from users (e.g., that are stored in association with user images) are used as input to a LLM to refine image captions generated by …
Real-Time Feedback On User-Written Generative Artificial Intelligence Model Prompts, Gulmohar Khan, Ramprasad Sedouram, Karthik Srinivas
Real-Time Feedback On User-Written Generative Artificial Intelligence Model Prompts, Gulmohar Khan, Ramprasad Sedouram, Karthik Srinivas
Defensive Publications Series
A common user interface for consumer use of generative artificial intelligence (genAI) tools is a chat interface, where a user can input their request as a text prompt within an input field and where the genAI model generates a response that answers the prompt. When users are insufficiently skilled at prompting, they may require multiple interactions with the genAI tool before a satisfactory result is generated. There are no mechanisms that guide the users to provide a prompt that uses good prompting strategies to express their intent. Also, user-provided prompts may include errors, gaps, or other issues that impact the …
System And Method For Thermochromic Indicator Payment Card, Mohit Koul, Rohit Mukherjee
System And Method For Thermochromic Indicator Payment Card, Mohit Koul, Rohit Mukherjee
Defensive Publications Series
The present disclosure related to a system and method for thermochromic indicator payment card. The system and method more specifically implements thermochromic technique. The payment card is embedded with thermochromic material that can change colour in response to temperature fluctuation, providing a passive and dynamic visual indicator for transaction verification and security alerts. Further, the present disclosure enables real time transaction awareness and fraud detection without relying on electronic alerts and external power sources.
Tax Rule Validation Using Large Language Models, Sandeep Chandna, Misha Narinsky
Tax Rule Validation Using Large Language Models, Sandeep Chandna, Misha Narinsky
Defensive Publications Series
The process of mapping products to jurisdiction-specific tax categories and validating rates against evolving legislation can involve complex, manual effort. A system is described that may leverage Large Language Models (LLMs) to assist with these tasks. The system can be configured to perform functions such as interpreting natural-language text from tax laws to classify products from an internal catalog. Additionally, the system can source and ingest public tax data to compare a company's internal tax rule configurations against external information sources. By automating the cross-referencing of internal product data with external legal and rate information, such a system can facilitate …
Establishing A Debug Channel Using Conductive Components Of An Electronic Device To Create Switchable Communication Paths, Averill Chang, Zoro Chang
Establishing A Debug Channel Using Conductive Components Of An Electronic Device To Create Switchable Communication Paths, Averill Chang, Zoro Chang
Defensive Publications Series
This disclosure describes a technique for establishing a debug channel over a wired connection by repurposing hardware components of an electronic device to create switchable communication paths for data transfer. In particular, internal debug signals can be selectively routed to non-dedicated external hardware contacts, for example, by using switching mechanisms. This approach may provide a stable and reliable debugging channel, potentially reducing the need for dedicated, corrosion-prone debug pins and mitigating certain performance and cost considerations associated with some wireless communication techniques.
A System For Recommending Advertising Campaign Settings Based On Performance Data Analysis, Ankit Aggarwal
A System For Recommending Advertising Campaign Settings Based On Performance Data Analysis, Ankit Aggarwal
Defensive Publications Series
Advertisers may experience difficulty in defining effective targeting and configuration settings for digital promotional campaigns, which can result in performance that does not meet certain goals for scale and return on investment (ROI). A system can utilize an artificial intelligence agent to analyze historical campaign data. The system may correlate past campaign configuration changes, such as adjustments to bids, budgets, and targeting criteria, with corresponding performance metrics including, for example, impressions, redemptions, and cost-per-action. Based on this correlation analysis, the system can formulate tailored recommendations for campaign settings to help meet specific advertiser goals. This process can contribute to improved …
Securing Llm Applications Using Intent-Based Feedback, Assaf Namer, Anil Kumar Sirikande, Christine Sizemore, Vinesh Prasanna M, Chris Diya, Prashant Kulkarni
Securing Llm Applications Using Intent-Based Feedback, Assaf Namer, Anil Kumar Sirikande, Christine Sizemore, Vinesh Prasanna M, Chris Diya, Prashant Kulkarni
Defensive Publications Series
Security threats such as prompt injection and jailbreaking pose significant risks to applications powered by large language models. Existing security toolkits often themselves leverage LLMs for malicious input detection. These security toolkits can produce false positives or negatives due to the non-deterministic nature of LLMs. This disclosure describes techniques that harden the security of LLM applications by incorporating an intent-based feedback loop that analyzes both user prompts and LLM responses and classifies adversarial intent using natural language processing (NLP). LLM system instructions are dynamically hardened to preemptively address identified vulnerabilities. Recognizing the importance of a layered security strategy, the techniques …
Expansion-Classification Prompt With “Thinking” Module For The Interpretation Of Highly Ambiguous Inputs, Chris Stewart
Expansion-Classification Prompt With “Thinking” Module For The Interpretation Of Highly Ambiguous Inputs, Chris Stewart
Defensive Publications Series
The technology described herein is directed to a method for accurately classifying user intent from highly ambiguous text strings. This can be used to classify online advertising deal names that require Special Category Data (SCD) compliance. As deal names inherently lack context, frequently use acronyms, and have minimal metadata associated with them, it is difficult to interpret the deal names for SCD relevance. This is addressed by using artificial intelligence (“AI”) models that can expand/interpret the deal names in a first phase and use the expansion/interpretation in a second phase to classify the deal name relevance for SCD purposes.