Open Access. Powered by Scholars. Published by Universities.®
Articles 2731 - 2760 of 4131
Full-Text Articles in Entire DC Network
Automated Text Clustering And Classification Of Feedback Data Using Llms, Zhaoying Han, Sathya Anurag Siruguppa
Automated Text Clustering And Classification Of Feedback Data Using Llms, Zhaoying Han, Sathya Anurag Siruguppa
Defensive Publications Series
Feedback from stakeholders connected to a software product is crucial for product development. Feedback can be used for a variety of purposes, such as continuous improvement, rapid prototyping, experimentation, resource allocation, improving user experience, etc. However, for products with a large set of features and a large volume of feedback (which may be in the form of freeform text), analyzing the feedback to generate useful insight is difficult. This disclosure describes the use of a vector database formed based on unsupervised learning to cluster a set of seed feedback data, tested by partitioning the seed data into training and test …
Accurate Path Tracing And Discrepancy Detection In Control And Data Plane Operations Using A Controller, Jaganbabu Rajamanickam, Darren R Dukes, Madhan Sankaranarayanan
Accurate Path Tracing And Discrepancy Detection In Control And Data Plane Operations Using A Controller, Jaganbabu Rajamanickam, Darren R Dukes, Madhan Sankaranarayanan
Defensive Publications Series
When an SRv6 (Segment Routing (SR) over Internet Protocol version 6 (IPv6)) node processes a packet, the destination address of the IP packet may be changed. This poses a challenge for end-to-end observability. Techniques described herein provide for validating segment identifier behavior at each node in a path and validating the path without requiring that intermediate nodes be aware of an operation performed. According to the techniques, SRv6 data plane and control plane validation is performed by the central controller.
Partial Models For Building Adaptive Model-Based Reinforcement Learning Agents
Partial Models For Building Adaptive Model-Based Reinforcement Learning Agents
Defensive Publications Series
In neuroscience, one of the key behavioral tests for determining whether a subject of study exhibits model-based behavior is to study its adaptiveness to local changes in the environment. In reinforcement learning, however, recent studies have shown that modern model-based agents display poor adaptivity to such changes. The main reason for this is that modern agents are typically designed to improve sample efficiency in single task settings and thus do not take into account the challenges that can arise in other settings. In local adaptation settings, one particularly important challenge is in quickly building and maintaining a sufficiently accurate model …
Transfer Entropy In Graph Convolutional Neural Networks
Transfer Entropy In Graph Convolutional Neural Networks
Defensive Publications Series
Graph Convolutional Networks (GCN) are Graph Neural Networks where the convolutions are applied over a graph. In contrast to Convolutional Neural Networks, GCN’s are designed to perform inference on graphs, where the number of nodes can vary, and the nodes are unordered. In this study, we address two important challenges related to GCNs: i) oversmoothing; and ii) the utilization of node relational properties (i.e., heterophily and homophily). Oversmoothing is the degradation of the discriminative capacity of nodes as a result of repeated aggregations. Heterophily is the tendency for nodes of different classes to connect, whereas homophily is the tendency of …
Graph Neural Networks Do Not Always Oversmooth
Graph Neural Networks Do Not Always Oversmooth
Defensive Publications Series
Graph neural networks (GNNs) have emerged as powerful tools for processing relational data in applications. However, GNNs suffer from the problem of oversmoothing, the property that the features of all nodes exponentially converge to the same vector over layers, prohibiting the design of deep GNNs. In this work we study oversmoothing in graph convolutional networks (GCNs) by using their Gaussian process (GP) equivalence in the limit of infinitely many hidden features. By generalizing methods from conventional deep neural networks (DNNs), we can describe the distribution of features at the output layer of deep GCNs in terms of a GP: as …
Operating System With Automatic End-To-End Encryption (E2ee) For User Content, Dave Kleidermacher
Operating System With Automatic End-To-End Encryption (E2ee) For User Content, Dave Kleidermacher
Defensive Publications Series
While end-to-end encryption (E2EE) can prevent access to user data by unauthorized parties, E2EE is supported by a limited set of applications. Retrofitting existing applications and services to support E2EE is expensive. This disclosure describes techniques that provide E2EE for user personal data across applications and services at the level of the operating system of a user device by automatic detection and/or manual specification of such content. With user permission, personal data is detected automatically by the operating system, e.g., using on-device machine learning or heuristics. When such data is detected, prior to sharing the data with another user, the …
Integration Of Dual Neural Networks In Game Engines Using Python Api And Socket Communication, Renee Gagnon
Integration Of Dual Neural Networks In Game Engines Using Python Api And Socket Communication, Renee Gagnon
Defensive Publications Series
This defensive publication describes a method for integrating dual neural networks, specifically a Spiking Neural Network (SNN) and a Long Short-Term Memory (LSTM) model, into game engines using Python API and socket communication. This approach enhances AI capabilities in first-person shooter (FPS) games by enabling real-time, dynamic behaviors in non-player characters (NPCs). The document outlines the architecture, implementation, and potential applications of this method, providing valuable insights for researchers and developers in the field of game AI.
Artificial Intelligence Based Inter Image Design Recommendation, Prajna Shetty, Alok Roy, Satyam Raj, Kandikatla Varshini
Artificial Intelligence Based Inter Image Design Recommendation, Prajna Shetty, Alok Roy, Satyam Raj, Kandikatla Varshini
Defensive Publications Series
The present disclosure discloses a system and method for providing design recommendations to a user. The method includes parsing the one or more design images of SVG format and analyzing the text and shape parameters associated with the one or more design images. Further the method includes generating python data frames with text identifiers, shape identifiers and connecting component identifiers and generating, by the connecting shape generator, component graph from the trained design images. Further the method includes loading trained component graph from the connecting shape generator and calculating shortest paths from node to be predicted to all other nodes …
“Dynamic Schema Evolution”, Harshwardhan Tripathi
“Dynamic Schema Evolution”, Harshwardhan Tripathi
Defensive Publications Series
The present disclosure provides a solution for seamless integration of data from diverse sources by dynamically managing schema changes. The solution comprises establishing a base schema during data processing and comparing it to incoming data layouts. Any detected changes, such as new tags or fields, are automatically managed within a staging layer, preventing disruptions to the main dataset. These changes are reviewed by Business and Engineering teams for relevance and compatibility. Only after thorough validation, approved changes are processed into the final dataset. This approach ensures that only validated and reliable data is integrated, maintaining data integrity and minimizing disruptions. …
Large Language Model (Llm) Business Helper For Cloud Databases, Ahmet Salih Aksakal
Large Language Model (Llm) Business Helper For Cloud Databases, Ahmet Salih Aksakal
Defensive Publications Series
Cloud database customers can inspect the data they store across multiple databases (potentially spanning across multiple database products) and find details such as schema and nature of stored data from individual database products. This requires technical knowledge of database querying and programming, which complicates business analytics and planning. This disclosure describes techniques to leverage data cataloging services and large language model (LLM) tools to analyze database schema information and to generate contextual information about the structures of the databases owned by a customer. By combining the power of an LLM chatbot with the schema gathering abilities of data cataloging services, …
Caching Of Retrieval Augmented Generation Model Trained With Dynamic Data And Dynamically Updating Of Cache, Mustafa Kadioglu, Paul C. Stojanovski, Sunil Karthik Kota
Caching Of Retrieval Augmented Generation Model Trained With Dynamic Data And Dynamically Updating Of Cache, Mustafa Kadioglu, Paul C. Stojanovski, Sunil Karthik Kota
Defensive Publications Series
Techniques are presented herein provide for an intelligent caching mechanism of a Retrieval Augmented Generation (RAG) model trained with dynamic data which is subject to get deleted or modified periodically. Based on the feedback from the customer, Cache Vector DB is updated. This method is capable of processing positive and negative feedback from customers to update the Cache Vector DB while simultaneously protecting from negative feedback spam. The method leverages a framework that not only caches an original question/query (primary) and the corresponding response, but also secondary question/queries which either uses GenAI to create synthetic queries of the original question/query …
System And Method For Configurable Vehicle Metrics Collection And Reporting, Selim Gurun, Umang Bhatia, Rui Qiu, Sergey Grizan
System And Method For Configurable Vehicle Metrics Collection And Reporting, Selim Gurun, Umang Bhatia, Rui Qiu, Sergey Grizan
Defensive Publications Series
This publication describes systems and techniques for implementing a vehicle communications language for describing vehicle system data sources, data collection rules, data processing, and data results using Protobuf for serializing data and establishing an interface description language for communications among various vehicle systems and vehicle data sources. Described techniques may enable vehicle data collection, vehicle data configuration and report generation. Aspects of the disclosure may include end-to-end cloud integration for vehicle systems and data sources utilizing a digital twin of a subject vehicle within the infrastructure of a cloud services provider and defining within the digital twin, available vehicle services, …
Rapid Video Prototyping And Content Creation Using Generative Artificial Intelligence, Krishna Chytanya Ayyagari
Rapid Video Prototyping And Content Creation Using Generative Artificial Intelligence, Krishna Chytanya Ayyagari
Defensive Publications Series
Traditional video production is a resource intensive process, requiring significant time and large financial investment. A common practice in entertainment and gaming industries is to generate prototypes or previews of video content before final video creation. This disclosure describes the use of generative artificial intelligence to automatically generate short video content that can be used directly as short-form video content (e.g., advertisements, short-form videos, etc.) or can be used as part of a larger video that may also include human-created content. Video generation requests are received as prompts, which are refined and processed to ensure compliance. Generative video libraries are …
Federated Learning-Enhanced Retrieval Augmented Generation (Rag), Eugenia Kim, Amanda L. Holst Ph.D., Myungjin Lee
Federated Learning-Enhanced Retrieval Augmented Generation (Rag), Eugenia Kim, Amanda L. Holst Ph.D., Myungjin Lee
Defensive Publications Series
Techniques are provided to enhance the functions of a retrieval augmented generation (RAG) mechanism for a large language model (LLM). A Federated Learning (FL)-enhanced RAG (FLERAG) mechanism is provided that can account for relevant context-enhancing data from the retrieval process, as well as most recent data from the FL on which a large language model (LLM) may not have been trained. Using FLERAG, the output generation is determined through a scoring or ranking method that indicates whether the response from the LLM or the FL model is most accurate and relevant. This generated response is then provided back to a …
Heatsink For High-Power Dissipating Electrical Components With Exposed Die, Anonymous
Heatsink For High-Power Dissipating Electrical Components With Exposed Die, Anonymous
Defensive Publications Series
The present disclosure relates heatsinks, namely a heatsink for a very high-power dissipating electrical component with an exposed die from thermal and vibration/shock perspective. The present disclosure addresses two different but co-existing challenges for the high-power components (e.g., ~500W), namely 1) Vibrational/Shock safety of Exposed/Lidless die type chip, and 2) High heat dissipation. To address these challenges, 1) a non-conductive foam gasket is added to the design to support the load over the die during the vibrations and 2) U-shaped heatpipes are added to heatsink to reduce the temperature variation throughout the heatsink. Benefits include heatsink size and weight reduction, …
High-Resolution Database Monitoring: Metric Streaming Protocols For Mission-Critical Applications, Shanu Sirothiya, Sonali Mahajan, Sumit Kumar Saha, Akshit Grover, Nityanand Vishwakarma
High-Resolution Database Monitoring: Metric Streaming Protocols For Mission-Critical Applications, Shanu Sirothiya, Sonali Mahajan, Sumit Kumar Saha, Akshit Grover, Nityanand Vishwakarma
Defensive Publications Series
Proposed herein is a Metrics Transfer Protocol (MTP), which is a highly efficient and specialized protocol engineered for the rapid and streamlined transfer of metric data between clients and servers. By utilizing dual connections for control and data, MTP ensures that metrics are initialized accurately and that subsequent updates, reflecting only the changes or deltas, are communicated with minimal overhead. This design is ideal for applications requiring real-time performance monitoring, as it minimizes network latency and resource usage, providing a lightweight alternative to more general-purpose protocols, such as Hypertext Transfer Protocol (HTTP) and the like.
Advanced Metrology For Display Uniformity Performance Judgement, Mike Li, Vicky Wen
Advanced Metrology For Display Uniformity Performance Judgement, Mike Li, Vicky Wen
Defensive Publications Series
This publication describes an advanced metrology for display device uniformity, which may better represent and reproduce on a display what a human eye naturally observes.
Workload Identity Management Using Agent And Controller, Ram Mohan R, Rajesh I V
Workload Identity Management Using Agent And Controller, Ram Mohan R, Rajesh I V
Defensive Publications Series
Techniques described herein provide for an approach in which an agent provisions a workload with initial credentials based on a workload profile that is learned using a controller. During runtime, when the workload requests a new token using its initial token to interact with peer workloads, an authorization server coordinates with the controller to learn the upstream/downstream peer workloads that will be in the path for the flows the workload will handle. Based on the security posture of those workloads, an appropriate scoped token is returned. This ensures malicious / untrusted workloads in the path cannot re-use or proxy the …
User Interface And Api For Curated Large Language Model Prompt Registry, Andrew Tomlinson, Kelsey Kraus, Margaret Kroll, Rupa Maiya, Anwaar Waddah Al Zireeni, Leslie Engler
User Interface And Api For Curated Large Language Model Prompt Registry, Andrew Tomlinson, Kelsey Kraus, Margaret Kroll, Rupa Maiya, Anwaar Waddah Al Zireeni, Leslie Engler
Defensive Publications Series
Proposed herein is a Large Language Model (LLM) user interface and backend web-based application, referred to as LLM Blocks or LLM User Interface (LLMUI), that shows users what an ideal prompt looks like for a selected task, lets them experiment with the prompt, see sample evaluation metric outputs, and permits finding and saving/sharing prompts across an organization.
Compact Compression Waterblock Lid With Die Inspection Opening For Die Failure Analysis, Thomas Chen, Danny H. Cheng, Galor Wenyi Zhang
Compact Compression Waterblock Lid With Die Inspection Opening For Die Failure Analysis, Thomas Chen, Danny H. Cheng, Galor Wenyi Zhang
Defensive Publications Series
A compact compression waterblock lid is provided with an integrated inspection cutout, designed for inspecting the die surface of chips. This solution enables active water cooling during failure analysis or operational testing while applying even distribution of the required force onto the chip in the socket. One of the benefits is that, by being in a socket, the chip does not need to be soldered to the board. The microchip is thereby allowed to function under typical or stress conditions while providing direct access to observe and analyze the chip's die for defects.
One-Click Analytics, Binayak Biswas, Afroze Ahmed, Bhuvanjeet Singh Gandhi, Karthik Gopalakrishna
One-Click Analytics, Binayak Biswas, Afroze Ahmed, Bhuvanjeet Singh Gandhi, Karthik Gopalakrishna
Defensive Publications Series
Present disclosure discloses a method and a system for providing one-click analytics for business users. In some embodiments, the method 400 includes at step 402, receiving user context information and a predefined prompt format. At step 404, the method 400 includes pre-processing the user context information. At step 406, the method 400 includes generating a response from the preprocessed data using Generative AI. At step 408, the method 400 includes formatting the generated response to obtain a standardized response based on the predefined prompt format and display the standardized response as an actionable insight.
Realistic Meeting Simulator, Daniel Macak, Jan Podmajerský, Andrea Pryglova, Michal Staškovan
Realistic Meeting Simulator, Daniel Macak, Jan Podmajerský, Andrea Pryglova, Michal Staškovan
Defensive Publications Series
A system is provided that simulates authentic presentation scenarios, whether in-person or virtual, to equip individuals with the skills and confidence needed to navigate complex interactions and challenging situations. This system allows individuals to enhance their professional effectiveness with a comprehensive approach to communication training.
Personalized Karaoke Experiences With Multiple Views And Ai-Based Coaching, Rakesh Kushwaha
Personalized Karaoke Experiences With Multiple Views And Ai-Based Coaching, Rakesh Kushwaha
Defensive Publications Series
Home and professional karaoke setups currently offer little to no immersive experiences, are not personalized to the user’s skills and preferences, and do not provide effective feedback on the user’s singing performance. This disclosure describes techniques that employ multiple views and utilize artificial intelligence (AI) features to provide a personalized immersive interactive karaoke experience that users can enjoy at home. The multiple views facilitate enhanced immersion via virtual reality/ augmented reality (VR/AR) features, such as virtual backgrounds, props, costumes, immersive settings, etc., along with pertinent information, such as lyrics synchronized to the song audio. The multiple views can facilitate enhanced …
Biometric Eye Scanner With Wearable Form Factor
Biometric Eye Scanner With Wearable Form Factor
Defensive Publications Series
Systems and methods for generating eye models. One aspect of the present disclosure relates to a method for generating an eye model. The method may include scanning, via a wearable device, an eye of a user. The method may include capturing biometric information based on the scanning. The method may include determining a cornea tomography based on the biometric information. The method may include generating an eye model based on the cornea tomography. [
A System And A Method For Generating Dynamic, Personalized Credit Score Predictions, Alok Roy, Prajna Shetty, Kandikatla Varshini, Satyam Raj
A System And A Method For Generating Dynamic, Personalized Credit Score Predictions, Alok Roy, Prajna Shetty, Kandikatla Varshini, Satyam Raj
Defensive Publications Series
The present disclosure provides a system and a method for generating dynamic, personalized credit score predictions. The system proposes an intelligent interactive tool that provides users with real-time insights as to how different financial decisions and activities of a user can influence his/her credit scores. In particular, the system receives user input on their financial data and decisions, fetches and analyses user’s credit information based on the user inputs and generates dynamic, personalized credit score predictions. Thus, the system empowers the users to explore hypothetical scenarios, enhancing their financial literacy and facilitating better informed decision making.
Integrating Artificial Intelligence (Ai) In Software Development, Bo Lin, Jens Michael Schueppert
Integrating Artificial Intelligence (Ai) In Software Development, Bo Lin, Jens Michael Schueppert
Defensive Publications Series
Building vertical applications for digital map applications, search engines, or other software requires several months of manual effort and often, repetitions of similar work across vertical applications. The time consuming and inefficient current processes cannot easily scale across different vertical applications, locales, or use cases. This disclosure systematically and dynamically integrates generative artificial intelligence (Gen AI) models within software development workflows. The AI-augmented software development process leverages AI models to perform tasks as an agent, to analyze data, to generate code, and/or to evaluate results. AI models can be integrated in a dynamic manner in major iterative software development activities …
Subscription Manager Sdk, Sangeeta Patnaik Visa, Suresh Kalakrishnan Visa, Hari Krishna Annam Visa, Debopam Das Visa
Subscription Manager Sdk, Sangeeta Patnaik Visa, Suresh Kalakrishnan Visa, Hari Krishna Annam Visa, Debopam Das Visa
Defensive Publications Series
The present disclosure relates to a subscription management solution, which boost customer retention and decrease churn while significantly lowering call center and dispute-related operational expenses. The Subscription SDK is built by integrating three different systems i.e., Card-On-File API (AMS), Token Inquiry API (TIS) and Stop Payment Service into DE SDK called as DE Subscription Management Bundle. This combined solution helps in i) quick and easy implementation through the Digital Enablement (DE) SDK and the Developer Platforms, ii) enables users to create stop instructions to protect against payments that are no longer necessary or authorized, iii) increases online subscription transparency by …
Deep-Drawn Mass-Spring Component With Sustainable Post-Consumer Filler, Konrad Hirschbeck
Deep-Drawn Mass-Spring Component With Sustainable Post-Consumer Filler, Konrad Hirschbeck
Defensive Publications Series
No abstract provided.
Dynamic Data Access Framework Using Symbolic Links And Context-Aware Caching, Puneet Mahajan
Dynamic Data Access Framework Using Symbolic Links And Context-Aware Caching, Puneet Mahajan
Defensive Publications Series
Managing and accessing data in distributed computing environments presents several challenges that traditional techniques often fail to address efficiently and adaptively. In distributed computing environments, data is replicated across various locations to ensure availability and performance. However, this has several drawbacks including high costs, complexity in data management, inefficient resource utilization, use of static data access strategies, high latency, etc. Data virtualization, data federation, distributed file systems, etc. do not provide responsive data storage that is dynamically updated based on real-time usage patterns and that can help reduce data storage costs. This disclosure describes a dynamic framework that integrates context-aware …
Ai Based Space Categorization And Intelligent Summaries, Fergal Keating
Ai Based Space Categorization And Intelligent Summaries, Fergal Keating
Defensive Publications Series
Collaboration technologies allow users to create spaces/channels for discussions and task creations. Recent advances in artificial intelligence (AI) facilitate assisting users dealing with data overload issues associated with spaces/channels, such as summarization, task extraction, and assistance with wording. However, users are largely left to organizing and managing spaces themselves. To overcome this issue, techniques are proposed herein that provide for the automatic categorization of spaces in conjunction with a more advanced summarization system that combines personalized profiles with space categorization to provide unique and concise user notifications and also remove unwanted noise.