Open Access. Powered by Scholars. Published by Universities.®
- Publication Year
- File Type
Articles 4621 - 4650 of 11865
Full-Text Articles in Entire DC Network
On-Device Llm-Driven Application Function Calling For Natural Language Interaction, Milen Ferev, Yishuang Pang, Pulkit Bhuwalka
On-Device Llm-Driven Application Function Calling For Natural Language Interaction, Milen Ferev, Yishuang Pang, Pulkit Bhuwalka
Defensive Publications Series
Large language models can enable application programmers to translate user instructions in natural language into a sequence of corresponding function calls; however, currently LLMs that provide such capabilities are housed in the cloud. Use of remotely hosted LLMs can introduce operational latency, affect reliability, and can be infeasible in certain cases. This disclosure describes the use of an on-device LLM to enable app developers to support natural language interactions with users without having to implement code within their applications. Per the techniques, with user permission, a user command in natural language as received by an application is provided to the …
Waypoint-Based Key Performance Indicators For Head Tracking, Narges Noori, Sean Yuan, Luke Jia
Waypoint-Based Key Performance Indicators For Head Tracking, Narges Noori, Sean Yuan, Luke Jia
Defensive Publications Series
A benchmarking algorithm identifies waypoints along a virtual reality (VR) headset trajectory and extracts key performance indicators (KPIs) for the headset tracking system at the identified waypoints. Artificial intelligence models are trained to identify the waypoints along complex trajectories with up to six degrees of freedom.
On Demand Digital Card Display, Anup Tripathi, Imo Akpan, Haven Vu, Suresh Kalakrishnan
On Demand Digital Card Display, Anup Tripathi, Imo Akpan, Haven Vu, Suresh Kalakrishnan
Defensive Publications Series
Disclosed herein is an outline of processes for accessing account information linked to a payment card, free from dependence on an issuer application. In this process a computing device first acquires a deep link from the payment network and then receives input from the user. Subsequently, an application clip is invoked in response to the user input, prompting the user for identity verification. Upon successful verification, the account information is displayed through the application clip, deep link or with in the native application.
Reconciling Third-Party Data By Leveraging Automated And Human Signals, Sandeep Ranganathan, Nicholas Seckar
Reconciling Third-Party Data By Leveraging Automated And Human Signals, Sandeep Ranganathan, Nicholas Seckar
Defensive Publications Series
The need for reconciling data originating from disparate sources exists in multiple domains such as digital maps for products such as restaurant reservations, appointments, food ordering and pickup; and artist attribution for products such as event ticketing etc. While domain-specific reconciliation infrastructure exists, such infrastructure generally does not work outside the domain that it is designed for. This disclosure describes techniques to reconcile different types of entities in a scalable and reliable manner by leveraging signals from automation (software running on a server) and humans (e.g., crowdsourced, internal engineering or operations teams, external partners, etc.) so that multiple domains can …
Low Latency Conversational Agent Via Multistage Autoregressive Completion Networks, D Shin
Low Latency Conversational Agent Via Multistage Autoregressive Completion Networks, D Shin
Defensive Publications Series
In high traffic scenarios, the total query-to-response latency for a conversational agent (chatbot) can be as high as a few seconds. The high total latency includes latencies for sending a query to a server and for the server to generate and provide a response. The high latency can make the conversational experience suboptimal for a user, as the user perceives it as a half-duplex conversation, where the user provides a query and needs to wait for a response before providing a follow up. This disclosure describes techniques that enable low latency conversations with a conversational agent with a large language …
Correcting Positioning Errors Using Position Data From Proximal Devices, D Shin
Correcting Positioning Errors Using Position Data From Proximal Devices, D Shin
Defensive Publications Series
In indoor environments, global positioning systems can report inaccurate coordinates due to multipath, fading, lack of satellite visibility, etc., leading to a suboptimal user experience. This disclosure describes techniques that use positions of nearby mobile devices detected via short-range wireless communication to correct the location coordinates of a device. In such cases, each device obtains respective location coordinates via GPS, and the location coordinates are very similar. The short-range connectivity of such devices, referred to as tagged devices, serves as a strong prior that the engaged devices are proximal. The proximity is leveraged to reduce the error in their location …
Conversational Platform To Obtain Operationally Relevant Data Insights, Amit Mathur, Sourav Sahu, Gopal Singh, Gunalan Samsil
Conversational Platform To Obtain Operationally Relevant Data Insights, Amit Mathur, Sourav Sahu, Gopal Singh, Gunalan Samsil
Defensive Publications Series
Organizations can face significant challenges in assessing operational health due to siloed data sources and the absence of standardized metrics. Critical data, scattered across various departments with disparate databases, can hinder a comprehensive view of performance and cause delayed decision-making and inefficient resource allocation. This disclosure describes techniques that leverage generative artificial intelligence to ingest data from diverse data sources, generate key operational metrics, and provide operational insights via a user-friendly conversational interface. The techniques result in automated, data-driven decision-making; improved operational efficiency; data democratization; support for simplified data exploration; and powerful insight generation.
Data Storage Using Artificial Intelligence Models, Na
Data Storage Using Artificial Intelligence Models, Na
Defensive Publications Series
Traditionally, data is stored in an addressable manner and is retrieved using database queries. Even when compression techniques are utilized, such data storage incurs a substantial cost and requires technical database skills to retrieve. This disclosure describes artificial intelligence (AI) techniques to efficiently store data. The techniques leverage the ability of AI models to ingest and compress data during a training phase and to reconstruct that data during a prediction phase. An AI model, thus repurposed for storage, automatically compresses the data by storing the data within model parameters. The stored data can be recovered using natural language queries that …
A Distributed Observability Architecture For Early Detection And Faster Resolution Of Various Data Incidents In Large Scale 4g/5g Deployments, Ravi Shekhar, Murat Alrabasi
A Distributed Observability Architecture For Early Detection And Faster Resolution Of Various Data Incidents In Large Scale 4g/5g Deployments, Ravi Shekhar, Murat Alrabasi
Defensive Publications Series
In the landscape of observability solutions, proposed herein is an adaptable architecture designed to effortlessly evolve alongside dynamic network environments. Specifically, a multi-layered, event-based architecture, streamlined data collection, processing, and analysis across various network layers are proposed herein, which may provide a comprehensive observability solution for 4G/5G mobile network infrastructure and applications.
Automatic Test Case Selection By Prompting A Large Language Model With Code Descriptions And Labeled Test Cases, Keun Soo Yim
Automatic Test Case Selection By Prompting A Large Language Model With Code Descriptions And Labeled Test Cases, Keun Soo Yim
Defensive Publications Series
Software testing requires selecting the tests to perform that are appropriate for a given event, such as code change, software release, solution deployment, and so on. Manual test selection is slow, expensive, and error-prone, such that relevant tests may be left out of the test set. This disclosure presents the use of a large language model (LLM) to select the optimal set of test cases relevant for a given software code change or update. Descriptions of source code changes and dependencies along with descriptions of labeled test cases are used to prompt the LLM to select relevant test cases. Multiple …
The Composite Projection Projecting Configuration That Contains 9 Projection Projecting Configurations Which Have The Capability To Collectively Project Multiple Color Combinations., Punarjeewa Abeysekera
The Composite Projection Projecting Configuration That Contains 9 Projection Projecting Configurations Which Have The Capability To Collectively Project Multiple Color Combinations., Punarjeewa Abeysekera
Defensive Publications Series
This paper describes a composite projection projecting configuration that contains 9 projection projecting configurations which have the capability to collectively project multiple color combinations. The composite projection projecting configuration will consists of 9 identical projection projecting configurations. The 9 identical projection projecting configurations will be arranged in a 2 dimensional array configuration. The composite projection projecting configuration, by utilizing the 9 identical projection projecting configurations, will have the capability to collectively and concurrently project 9 projections onto a common projection superimposing area. In addition to that, the composite projection projecting configuration, by utilizing the 9 identical projection projecting configurations, will …
Two Stage Back-To-Face Serial Compressor Wheel
Two Stage Back-To-Face Serial Compressor Wheel
Defensive Publications Series
Designing a two-stage centrifugal compressor involves several challenges, particularly concerning thrust and clearances. Efficient thrust balancing is essential to prevent excessive wear on bearings and ensure stable operation while optimizing tip clearance is essential to avoid performance losses. The proposed solution of two stage back-to-face serial compressor wheel with wheels in same direction and integrated labyrinth improves the interstage design, reduces the pressure difference across the stages, and improves overall efficiency.
Enriching Application Programming Interface (Api) Documentation Using A Large Language Model (Llm) Architecture To Improve Llm Api Interpretability, Gal Bashan, Gafnit Amiga, Nadav Lotan, Raz Lotan, Tal Vintrob
Enriching Application Programming Interface (Api) Documentation Using A Large Language Model (Llm) Architecture To Improve Llm Api Interpretability, Gal Bashan, Gafnit Amiga, Nadav Lotan, Raz Lotan, Tal Vintrob
Defensive Publications Series
Proposed herein is an innovative system designed to address the critical challenge of maintaining accurate and detailed Application Programming Interface (API) documentation in the fast-past environment of software development. Broadly, the system streamlines the creation and updating an of API documentation by leveraging a Large Language Model (LLM) that intelligently analyzes code changes and generates documentation patches/updates based on any detected code changes.
Motion Capture Based Signature Authorization In A Mixed Reality Environment, Samuel Wu
Motion Capture Based Signature Authorization In A Mixed Reality Environment, Samuel Wu
Defensive Publications Series
The subject disclosure provides computer-implemented methods that include generating, in a mixed-reality environment, a signature capture area in a written document; detecting, using a gesture sensor, a series of movements of a user within the signature capture area; converting the series of movements into an image; and inserting, into a signature portion of the written document as a signature of the user, the image.
Dynamic, Context-Aware Computing Device Initialization For Personalized User Experiences, Sudeep Chauhan, Eusebius Lai
Dynamic, Context-Aware Computing Device Initialization For Personalized User Experiences, Sudeep Chauhan, Eusebius Lai
Defensive Publications Series
This document describes techniques for tailoring the initial setup and default configuration of computing devices such as smartphones and smartwatches, to individual users. The proposed computing device personalization engine utilizes a multi-faceted analysis of user data, environmental context, and device usage patterns, collectively referred to as contextual signals, to derive a set of abstract dimensions and scores that encapsulate user preferences. In some instances, the personalization engine uses machine learning models to derive the abstract dimensions and scores. The set of abstract dimensions and scores are translated into a predefined set of default Enum values that are indicative of a …
Streamlining Data Engineering With Llm-Backed Self-Service Tools, Sathya Anurag Siruguppa
Streamlining Data Engineering With Llm-Backed Self-Service Tools, Sathya Anurag Siruguppa
Defensive Publications Series
Existing automation solutions for data processing pipeline, including those that use machine learning techniques, lack customization and flexibility and fail to address specific needs or account for unique data requirements such as privacy requirements or stringent validation requirements for certain types of data. This disclosure describes techniques that support multiple use cases in data processing pipelines through self-service tools and provide a framework for various sections of data pipeline creation. This benefits stakeholders such as data scientists, machine learning (ML) engineers, and data engineers. A large language model (LLM)-backed engine powers the chatbot and automatically generates certain files that provide …
Anticipated Amount Verification, Pinesh Roy, Barbara Patterson
Anticipated Amount Verification, Pinesh Roy, Barbara Patterson
Defensive Publications Series
The present disclosure provides a system and a method for verifying an anticipated amount of a digital financial transaction. The disclosure proposes including a request for an anticipated amount verification (AAV) associated with the transaction, while forwarding a transaction request from a merchant to an issuer via an acquirer and a payment processing platform. Thereafter, the issuer generates an AAV response indicating whether the transaction is eligible for the anticipated amount. Further, the issuer communicates the AAV response to the merchant, via the payment processing platform and the acquirer. Based on the AAV response the merchant determines whether the transaction …
Redirection Of User Devices To Alternative Vplmns During Capacity Overload, Vimal Srivastava, Sri Gundavelli, Ravi Kiran Guntupalli
Redirection Of User Devices To Alternative Vplmns During Capacity Overload, Vimal Srivastava, Sri Gundavelli, Ravi Kiran Guntupalli
Defensive Publications Series
Techniques described herein provide for temporarily redirecting a user device to a partner VPLMN when a current serving VPLMN is congested. In particular, the serving network indicates to a Home Public Land Mobile Network (HPLMN) that a congestion condition has occurred so that the HPLMN can trigger redirection of UEs to a different VPLMN. In addition, a UE may inform the HPLMN about poor QoS so that the HPLMN can redirect the UE to different VPLMN. Furthermore, the HPLMN may redirect the UE to a desired VPLMN temporarily without modifying its Roaming Priority List (RPL) information in the Universal Subscriber …
One-Click Exportable Feature Set File For Gui-Based Network Controllers Or Systems, Mihir Prakash Vyas
One-Click Exportable Feature Set File For Gui-Based Network Controllers Or Systems, Mihir Prakash Vyas
Defensive Publications Series
For graphical user interface (GUI) -based network management systems, it can be difficult to obtain a complete list of features, knobs, options, etc. currently in-use by a network controller/management system. Proposed herein is an easy, quick, and reliable technique to facilitate capturing and sharing feature sets, including minor features, for GUI controller-based network management systems/devices that may enhance customer experience and provide accurate information regarding any applicable bugs.
Data Sharing Between Devices Without Server Device, Hp Inc
Data Sharing Between Devices Without Server Device, Hp Inc
Defensive Publications Series
This is about the method to share data between devices without the need for a server. This approach eliminates the costs associated with setting up and maintaining a server, offering a more efficient alternative to traditional methods. Devices are categorized as a Center, Sub-Center, or General device for data sharing purposes, with these roles pre-assigned by an administrator. This method is particularly well-suited for environments with a relatively small number of devices, such as in small to medium-sized businesses, where customers prefer not to manage additional servers.
Dynamic Employee Emissions Management Controller, Tiziano Canovai, Esther Roure Vila, Marisol Palmero, Diane Mievis, Michael King
Dynamic Employee Emissions Management Controller, Tiziano Canovai, Esther Roure Vila, Marisol Palmero, Diane Mievis, Michael King
Defensive Publications Series
Efforts to reduce greenhouse gas emissions from the transport industry are critical in addressing global emissions. Employee commuting contributes significantly to emissions in the transportation industry (Scope 3 emissions). To overcome this issue, proposed herein is Dynamic Employee Emissions Management (DEEM), a system and method for dynamic urban mobility management in transportation networks, calculating emissions from employee commuting (under Scope 3, Category 7).
Efficient Network Automation Through Enhanced Retrieval Augemented Generation (Rag) Architecture, Koushik Padavalli Gowrish, Supratim Chakraborty, Somaraju Kaki, Nithin Chitta, Mahesh Chittur Kalyanakrishnan
Efficient Network Automation Through Enhanced Retrieval Augemented Generation (Rag) Architecture, Koushik Padavalli Gowrish, Supratim Chakraborty, Somaraju Kaki, Nithin Chitta, Mahesh Chittur Kalyanakrishnan
Defensive Publications Series
Proposed herein is an enhanced retrieval augmented generation (E-RAG) architecture for use with network automation applications. The enhanced RAG architecture aims to increase RAG performance by reducing the hallucinations of a Large Language Model (LLM) and by decreasing the context length of messages sent to the LLM.
Coordinated Radar Detection For Unii-2 Channel Access, Anirban Karmakar, Mathieu Monney, Domenico Ficara
Coordinated Radar Detection For Unii-2 Channel Access, Anirban Karmakar, Mathieu Monney, Domenico Ficara
Defensive Publications Series
A channel switch to UNII-2 (Unlicensed National Information Infrastructure) channels includes observing a channel availability check (CAC) before accessing the channel. Only certain access points (APs) in a network may have the ability to observe CAC using radio based spare radar detectors. Another challenging issue is when radar is heard on a channel for which CAC has already been observed, causing the AP to switch the channel. To overcome these issues, proposed herein is a technique using coordinated radar detection amongst co-located APs in a network for CAC free UNII-2 channel access.
Method, System, And Computer Program Product For Automated Metadata Tagging, Vijaya Das Cr, Harsimran Singh, Jayachandra Babu Adusumalli, Venugopal K
Method, System, And Computer Program Product For Automated Metadata Tagging, Vijaya Das Cr, Harsimran Singh, Jayachandra Babu Adusumalli, Venugopal K
Defensive Publications Series
A computer-implemented method includes: receiving a first data set and lineage data associated with formation thereof, the first data set derived from a second data set; determining that a first field of the first data set was generated by applying a transformation of a first type of transformation function to a first field from the second data set based on the lineage data; identifying a metadata tag associated with the first field from the second data set or a field from a third data set from which the first field from the second data set was generated by applying the …
Traction Drive Rotor Cooling And Weight Reduction With Additional Light Weight Inner Solid Shaft
Traction Drive Rotor Cooling And Weight Reduction With Additional Light Weight Inner Solid Shaft
Defensive Publications Series
A hollow-shaft rotor cooling system is an effective cooling system often used for automotive traction motors. Flow and heat transfer of the hollow shaft rotor cooling system are governed by the shaft inner bore diameter and speed. There is usually a trade-off between rotor inner diameter and windage losses to optimize rotor cooling. The proposed solution discloses a lightweight rotating inner plastic shaft inside the metal shaft that not only reduces the rotor weight but also reduces the windage losses.
Integrated Method For Malignant Cancer Detection And Metastasis Prediction Using Fractal Vein Analysis, Deep Learning, And Multi-Modal Imaging, Renee Gagnon
Defensive Publications Series
This publication presents an innovative method for analyzing, detecting, and predicting malignant cancer and metastasis through the integration of fractal vein pattern analysis, Chaos Game Representation (CGR), state-space models, and a hybrid Spiking Neural Network (SNN) / Convolutional Neural Network (CNN) architecture. This comprehensive approach is designed to address the challenges associated with detecting malignancies and predicting metastasis in medical images across various imaging modalities, focusing on the unique combination of advanced mathematical techniques and deep learning to refine and improve accuracy in cancer detection and prognosis
Dbts Design For Hdr True Black, Hp Inc
Dbts Design For Hdr True Black, Hp Inc
Defensive Publications Series
The manuscript reveals a new direct bond touch sensor + cover glass design, that can help to mitigate the light leakage induced by the touch sensor. By adding the surface plasmon layer on one side of the touch sensor, the large angle light would be trapped in the surface, and difficult to emit out as light leakage.
Augmenting Images With Contextual Texture Cues To Assist Viewers With Color Vision Deficiency, Shumeet Baluja, David Marwood
Augmenting Images With Contextual Texture Cues To Assist Viewers With Color Vision Deficiency, Shumeet Baluja, David Marwood
Defensive Publications Series
This disclosure describes techniques that enable a viewer with color vision deficiency to better understand color images by automatically mapping problematic colors to distinguishable textures - distinctive black-and-white patterns corresponding to particular colors. A deep learning model selects appropriate textures for each color such that textures are consistent across images. Textures inform the viewer of similar and dissimilar colors and are only applied where needed. The original color image is augmented with textures such that both color-deficient users and users with regular vision can understand the image. The techniques can be used to display images on any device and can …
Method, System, And Computer Program Product For Capturing Temporal Dynamics, Liang Wang, Shubham Jain, Yingtong Dou, Junpeng Wang, Michael Yeh, Yan Zheng, Prince Osei Aboagye, Wei Zhang
Method, System, And Computer Program Product For Capturing Temporal Dynamics, Liang Wang, Shubham Jain, Yingtong Dou, Junpeng Wang, Michael Yeh, Yan Zheng, Prince Osei Aboagye, Wei Zhang
Defensive Publications Series
Methods, systems, and computer program products may obtain ratings by users for products at time points; for each user and product pair: generate, for a time point, a user neighborhood that includes products rated by that user at or after that time point and a product neighborhood that includes users that rated that product at or after that time point, an individual-based rating of that user for that product at or after that time point being generated based on the user neighborhood, and a crowd-based rating of that product for that user at or after that time point being generated …
A Micronized Crystalline Form-R Of 2-(3,5-Dichloro Phenyl) -1,3-Benzoxazole-6- Carboxylic Acid, Msn Laboratories Private Limited, R&D Center; Srinivasan Thirumalai Rajan, Sagyam Rajeshwar Reddy, Kammari Balraju
A Micronized Crystalline Form-R Of 2-(3,5-Dichloro Phenyl) -1,3-Benzoxazole-6- Carboxylic Acid, Msn Laboratories Private Limited, R&D Center; Srinivasan Thirumalai Rajan, Sagyam Rajeshwar Reddy, Kammari Balraju
Defensive Publications Series
The present invention provides Micronized Crystalline form-R of 2-(3,5- dichloro phenyl)-1,3-benzoxazole-6-carboxylic acid of formula (I) and process for preparation thereof