Artificial Intelligence Powered Brand Protection Tool

Defensive Publications Series

Methods, systems, or apparatuses for analyzing threats to a brand that enables the use of artificial intelligence to analyze one or more electronic publications for unauthorized product listing, or generate a report of product listings associated with information provided in a query. The disclosed system may receive one or more queries, search one or more electronic publications, obtain data from the one or more electronic publications, create a database of the obtained data from the one or more electronic publications, analyze the obtained data, or generate one or more reports with information related to the obtained data.

Artificial Intelligence To Quantify Employee Communication And Personality Dynamics

Defensive Publications Series

Systems and methods for evaluating an employee’s professional personality are provided. The systems and methods may enable the use of artificial intelligence to perform sentiment analysis on one or more employee communications, and quantify one or more personality characteristics based on the one or more employee communications. The systems and methods may compare the quantified personality characteristics to one or more employees, and may generate one or more employee scores for employees for each of the one or more personality characteristics.

Multimodal Based Television Recommendation, Shravan Nayak, Tamojit Chatterjee, Sundaramoorthy Murugesan, Venkata Gangadhar Kanchu, Kopal Niranjan, Priyanshi Sharma, Kanishka Mishra

Defensive Publications Series

Systems and techniques are proposed that provide an improved experience for a user when interacting with a television (TV) application running on a media player. The TV application may interface with media content providers to fetch media content for streaming to and viewing on the media player by way of the internet. The media player may be a computing device (e.g., a Smart TV, a tablet computer, a smart phone, a laptop computer, a mobile computing device) connected to a network and interfaced to the internet.

The TV application may receive media content recommendations from a recommendation engine. The recommendation …

Network-Based Distributed Low Rank Adaptation Adapter System, Nadhem J. Al-Fardan

Defensive Publications Series

Presented herein are novel techniques for integrating low-rank adaptation (LoRA) adapters in a distributed Graphics Processing Unit (GPU) environment while addressing privacy concerns. By hosting a base model and the adapters on separate GPUs, the techniques provide for balancing computational efficiency with data security. Through the techniques proposed herein, a dynamic system for AI model inference can be provided that separates the base model and the LoRA adapter onto different GPUs or two groups of clustered GPUs connected via a network. A cluster is defined as a set of GPUs that run a similar binary such as a base model …

Auditing The Future: Ensuring Ethical And Secure Ai And Quantum Computing Itgc And Business Controls, Ramkumar Bharathan

Defensive Publications Series

Artificial intelligence (AI) and quantum computing are rapidly transforming the business landscape, demanding a parallel evolution in auditing frameworks to ensure their ethical, secure, and reliable deployment. This paper presents a comprehensive auditing approach tailored to the unique challenges posed by these technologies, aligning with Sarbanes-Oxley (SOX), International Financial Reporting Standards (IFRS), and Generally Accepted Accounting Principles (GAAP). It addresses key control areas, including access controls, change management, data integrity, model validation, and incident management, providing detailed audit steps with walkthroughs and testing examples. The framework incorporates EU guidance on algorithms and AI governance, emphasizing algorithmic bias mitigation, data privacy, …

Providing Enhanced Learning In Llms By Associating Data With Role-Based Source Information, Laxmi Mukund, Shraddha S K

Defensive Publications Series

In today's fast-paced world, the rapid introduction of new product features necessitates accurate and timely documentation. Traditional manual documentation methods are inefficient, error-prone, and often result in outdated information. Techniques are presented herein to address these challenges by leveraging Large Language Models (LLMs) to automate the generation of precise and reliable documents by associating data with well-defined role-based source information. The impact of this solution goes beyond generating documentation for new features; it can also be extended to enable targeted stakeholder interactions, to identify speakers in recorded conversations, and to classify unclassified dialogues.

Design Method For Power Efficiency Optimization Of Zonal Illuminated Projector, Ted Cym

Defensive Publications Series

A method is provided for optimizing a power efficiency of an illuminator including a plurality of illumination zones that are individually addressable for providing a zonal illumination at a non-emissive display panel. The method includes calculating a plurality of illumination powers of the illuminator for a plurality of zone layouts of the illuminator, each zone layout including a number of the zones included in the illuminator, a size of the illuminator, and a color sub-zone configuration for each of the plurality of illumination zones. The method also includes selecting a zone layout with a lowest illumination power of the illuminator.

Design Method For Power Efficiency Optimization Of Zonal Illuminated Projector

Defensive Publications Series

A method is provided for optimizing a power efficiency of an illuminator including a plurality of illumination zones that are individually addressable for providing a zonal illumination at a non-emissive display panel. The method includes calculating a plurality of illumination powers of the illuminator for a plurality of zone layouts of the illuminator, each zone layout including a number of the zones included in the illuminator, a size of the illuminator, and a color sub-zone configuration for each of the plurality of illumination zones. The method also includes selecting a zone layout with a lowest illumination power of the illuminator.

System And Method To Authorize A Payment Transaction Using A Digital Glove, Sri Sai Krishna Kiran Atmuri

Defensive Publications Series

The present disclosure relates to a system and method to authorize a payment transaction using a digital glove. The present disclosure suggests tapping a cardholder’s digital glove on a merchant payment system without entering a Personal Identification Number (PIN) to initiate a payment transaction. The merchant payment system transmits the payment transaction data to an issuer system through a network via an acquirer server for authorization. Upon notification, the cardholder enters the PIN by performing a sequence of pinches between a thumb finger distal phalange and the other finger phalanges of the digital glove. Thereafter, encrypt the chronological order of …

Benefits Marketplace, Ana Luisa Wu, Jenny Ramirez, Jose Ricardo Tafur Jr, Ricardo Collada, Maikel Zuniga, Manuel Ferreiro Caballero, Salvador Ricardo, Liset Iturriaga Chavez, Francisco Alberdi, Alexander Dale, Orlando Quesada, Ernesto Garcia Vazquez, Daniel Astudillo, Laura Figueroa, Sebastian Hernandez, Cristina Villarroel, Yuset Rosello Moya, Rachel Quijano Vizoso

Defensive Publications Series

A method of provisioning a benefits marketplace platform is disclosed. According to the method, a payment network server computer presents payment network benefits to a consumer and receives feedback from the consumer based on the payment network benefits presented to the consumer. A profiling module profiles the payment network benefits based on the feedback received from the consumer and provides a benefits profile to the payment network server computer. The payment network server provides a recommendation of benefits to the consumer based on the benefits profile. A benefits module receives benefits selected by the consumer from a checkout function, applies …

User Interface For Efficient Control Of Autonomous Agent Tasks, Diego Rivas, D Shin

Defensive Publications Series

This disclosure describes techniques to manage tasks that are executed by software agents in graphical environments such as extended reality environments displayed via an extended reality headset and/or other graphical user interface. Per the techniques, tasks may be performed by the user or an agent. A visual indication is provided for a task that indicates whether the task is controlled and performed by the user or by an agent. Panels are displayed in the graphical environment that indicate the task being performed, inputs and/or outputs related to the tasks, and progress in the task. A task performed by an agent …

Automatic Generation Of Search Dimensions For Query Refinement, Manasa Babu, Sanika Barve, Ramprasad Sedouram

Defensive Publications Series

This disclosure describes techniques to seamlessly refine and update search queries, resulting in an enhanced user experience and convenience. Per the techniques, in addition to traditional textual refining of queries, users can fine-tune the output results for specific requirements pertaining to information size, personalization, multi-device display of search results, etc. A user can adjust the granularity of the search results, from a gist to in-depth or personalized knowledge, by simple gestures such as tilting the device towards or away from themselves. Information size in the search results can be automatically adjusted to suit the present display size. Search results can …

Predictive Flow Capture Controls For Multi-Domain Network Assurance Tools, Robert Barton, Jerome Henry, Abhishek Kumar, Gino Corleto

Defensive Publications Series

Advanced network assurance tools provide for the ability to manage network assurance across a variety of service domains. Such tools may facilitate synthetic network monitoring through active monitoring of live flow data. While this opens new opportunities for network assurance, architecturally it also presents challenges related to the orchestration of how and where flow reporting is established and managed. Techniques are proposed herein that include providing an AI/ML mechanism to predict network problems such that flow collection can be automated to be performed only where necessary, when necessary, and for the protocols and/or hosts that are experiencing issues. The techniques …

Thermal Imaging Based Solutions For Implementing Platform Management In Networking/Server Equipment, Hari Pillai, Phong Ho, Stephen A Scearce, Lauren A Marino

Defensive Publications Series

Existing systems for internal monitoring of networking/server equipment are not scaling to meet the needs of the next generation of electronic system issues. Current solutions for internal monitoring only provide a limited amount of information that leaves gaps in the understanding of the condition and integrity of the networking products. To overcome these issues, proposed herein are techniques that leverage real time infrared (IR) monitoring to quickly detect intrusive events, thermal complications, corrosion, and dust accumulation before damage to the networking products or customers occur.

Method And System For Mitigating Cognitive Driver Distraction From In-Vehicle Conversations, Anonymous

Defensive Publications Series

Driver distraction is a significant safety concern, with passenger conversations contributing significantly to cognitive load. While physical distractions are often addressed, in-vehicle conversational dynamics and their impact on driver attention remain largely unmanaged. Intense or emotionally charged conversations can significantly reduce driver focus, regardless of the number of passengers. Current systems lack the capability to monitor and mitigate the cognitive distraction arising from in-car interactions. This invention addresses this critical gap by leveraging existing vehicle systems to manage in-car conversational tone and enhance driver focus.

Existing methods primarily focus on physical distractions, neglecting the cognitive impact of in-vehicle conversations. Adding …

Establishing A Secure Sharing Environment Using Partial Cryptographic Keys, Salmaan Rashid

Defensive Publications Series

A system for establishing a secure sharing environment using partial cryptographic keys. The proposed system uses a complete cryptographic key divided into multiple partial keys and provided to two or more participants. The participants attest each TEE of the secure sharing environment desired for use and send, to each TEE, their respective partial key. Each TEE combines the received partial keys to generate the complete cryptographic key. Using the complete cryptographic key, the TEEs can securely communicate and share data with each other.

Automated Regression Testing Framework For Large Language Models, Claudio Criscione, Sebastian Lekies

Defensive Publications Series

This disclosure describes a system for automated regression testing of large language models (LLMs). LLMs can be trained or fine-tuned with safety-oriented data and similar techniques to prevent undesirable outputs. However, additional training and fine-tuning can result in an LLM regressing and producing undesirable outputs. These regressions can be difficult to detect and track given the fast pace of research and development in this field. This system provides automated infrastructure and techniques for testing new versions of LLMs for regression against stored tests. The system includes integrator modules that provide a uniform testing interface for various types of LLMs and …

Defined-Trust Transport Protocol For Limited Domains, Kathleen Nichols

Defensive Publications Series

The Defined-trust Transport (DeftT) is designed to provide secure, default-deny networking for closed communities with dynamic membership and a collection-based construct that is efficient on broadcast media. Such closed communities, also known as limited domains [LDCCR], are frequently used in Operational Technology (OT) networking, in particular , in Critical Infrastructure uses. DeftT is designed to express and enforce application- and deployment-specific integrity, authentication, access control and behavior constraints directly in its protocol modules. It enables secure and completely self-contained (e.g., no external identity servers or certificate authorities) overlay networks where credentialed members can join and leave at any time. Security …

Action Item Management Across Virtual Collaboration Meetings And Spaces, Chidambaram Arunachalam, Gonzalo A Salgueiro, Paul Anholt, Hazim Dahir

Defensive Publications Series

The concept of sharing key points and identifying action items based on a discussion in a virtual collaboration meeting is the most effective way of documenting the decisions made and tasks assigned during the meeting. However, meeting summaries including AI based summaries, does not provide information on whether the action items were already completed during the meeting. Proposed herein are techniques to determine completion of action items related to a specific topic based on conversations, real-time media, file share events across related virtual collaboration spaces and virtual meetings.

Quantum And Bio-Inspired Self-Healing Networks, Na

Defensive Publications Series

This disclosure describes techniques to reduce the costs and inefficiencies associated with providing redundancy in networks by using self-healing network architectures. The techniques feature quantum-inspired routing algorithms (QIRA) and bio-inspired self-healing (BISHA). QIRA leverages quantum principles to concurrently explore multiple paths and to share information about network conditions through quantum entanglement. BISHA leverages cellular automata to mediate inter-device communication, such that when a failure is detected, the device is isolated and information about the failure is communicated to neighboring devices. The neighboring devices collaborate to reconfigure the network topology and reroute data traffic around the failed component. Machine learning algorithms, …

Integrated Circuit Emulation Based On Scan Mode Signal Flow, Na

Defensive Publications Series

This disclosure describes optimized emulation of integrated circuits (ICs) based on migration of information from scan test signal flows (DFT signals). Per techniques of this disclosure, DFT signal flows such as scan mode signals and scan clocks are automatically migrated to emulation signal flows and tied off during emulation such that they are excluded during emulation of the IC. Tie-off of DFT signals can enable faster emulation as well as relieve capacity constraints on the emulation platform. In some implementations, automatic scripts may be utilized to generate emulation code based on the functional tests to be performed and the DFT …

Data-Driven Design For Thermal Management And Control In Datacenters, Na

Defensive Publications Series

Design of thermal controllers for datacenter machines has traditionally relied on heuristics and trial-and-error testing, resulting in lengthy design times and possibly suboptimal controller parameters. This disclosure describes techniques that leverage control and linear-system theory to build models that enable the software simulation of thermal controllers used in datacenters. Per the techniques, test data is captured from a real machine and used to create a reduced-order model of the machine and its thermal controller. A transfer function of the reduced-order model is obtained using linear system theory. Control theory is leveraged to determine the appropriate controller parameters for the machine …

Contextual Conversational Advertisements In Agents, Oded Elyada, Ilia Labzovsky, Danny Karmon

Defensive Publications Series

The technology described in this paper relates to identifying relevant advertisements (“ads”) based on a textual conversation, e.g., chat, between a user and an agent and generating a text-based natural language advertisement that fits the context of the conversation. A natural language query is generated based on the conversation between the user and the agent to identify relevant ads. The query is transmitted to an advertising system that can identify the relevant ads based on the context of the conversation and/or additional signals. An artificial intelligence (“AI”) model, such as a large language model (“LLM”), converts the relevant ad into …

Flowchart Orchestration For Complex Task Performance Using A Large Language Model, Na

Defensive Publications Series

While large language models (LLMs) can be used to perform complex tasks, e.g., rating search results, training and maintaining LLMs runs into several challenges. These include the significant expense to obtain training data, difficulty in identifying weaknesses, lack of explainability, and the lack of transferability to new foundation models or versions. This disclosure describes the use of flowcharts for orchestration of large language models (LLMs) to perform complex tasks. A complex task is broken down into a predefined set of simpler tasks. These simpler tasks are represented in a traditional flowchart format. The LLM is prompted to make the individual …

A Method And A System For Clustering Similar Transaction Disputes, Rupali Shah

Defensive Publications Series

The present disclosure provides a method and a dispute clustering system for Clustering similar transaction disputes. In the present disclosure, during a training phase, a dispute clustering system may receive historical transaction disputes to form sets of historical transaction disputes with similar causes. Each set of historical transaction disputes may be resolved using the same resolving method. In real time, the dispute clustering system may receive one or more transaction disputes. Then, the dispute clustering system may determine a related set of historical transaction disputes for each of the one or more transaction disputes by comparing the causes and rules …

Using Pattern Recognition To Dynamically Dampen User Control Inputs, Hp Inc

Defensive Publications Series

Many games and applications use human machine input devices. The types of operations of the input device fall into two primary categories: large gross movements and small movements for fine adjustments. The sensitivity settings for the input devices work to dampen the movements generated by the input device. The goal of this disclosure is to outline a method of dynamically dampening the inputs from an human input device based on the recognition of the type of movement through pattern matching. The process would allow large movements to be faster through lessened dampening, and would allow small fine adjustments to be …

User Interface Personalization Based On User Preferences And Practices, Ajay Prasad

Defensive Publications Series

While some websites and apps support manually configurable user interface personalization, such personalization approaches do not evolve dynamically with changing user preferences and do not effectively leverage the collective intelligence across the user community as a whole. Non-personalized and difficult to use user interfaces can lead to user frustration and/or reduced engagement. This disclosure describes techniques that, with user permission, leverage individual and community user preferences and practices for dynamic real-time personalization of a UI including one or more of layout, content, and functionality. User interaction with the personalized layout as well as explicit user input is used to fine-tune …

Web Security Scanner With Accelerate Detection And Verification Of Vulnerabilities, Yash Bhatnagar, Athira M.

Defensive Publications Series

A full scan is performed in response to receiving a request to perform a full scan. A plurality of findings is generated from the full scan. Each finding includes a URL, security vulnerability of the URL, fuzzy payload(s) used to interact with the URL, algorithm of the vulnerability detector used to identify the security vulnerability, and a current status. A verification of a finding is performed using the URL, fuzzy payload(s), and algorithm of the finding. The current status of the finding is updated based on the results of the verification of the finding. A quick scan is performed on …

System And Method For Checkpointing And State Synchronization For Fault Tolerance In Long-Running Map Reduce Jobs, Alok Roy

Defensive Publications Series

The present disclosure relates to a method and system for performing checkpointing and state synchronization for fault tolerance in long-running MapReduce jobs. The method involves integrating checkpointing mechanism into both the Map and Reduce phases of the MapReduce jobs, capturing and storing critical data and metadata at key intervals. Additionally, the method includes replicating these checkpoints across all clusters in the active-active setup, ensuring that any cluster can access the most recent checkpoint and resume the MapReduce job in case of failure or re-routing. Finally, the method ensures that the checkpoints are synchronized across clusters before the job proceeds, providing …

System And Method For Ensuring Sequential Event Integrity In Distributed Computing Systems, Manav Shah, Mansi Singh, Debdeep Bandyopadhyay, Shivam Mohan

Defensive Publications Series

The present disclosure discloses a system and method for ensuring sequential event integrity in distributed computing systems. The present disclosure aims to streamline event management and improve overall system performance. The present disclosure ensures high sequential integrity of interdependent events across distributed computing systems. The present disclosure is also proposed to solve resource contentions within context switching. Further, the system and method may communicate event processing results to all participating systems in batch and real-time mode to avoid redundant processing of the same event.