Digital Commons Network^™

Intelligent, End-To-End Test Selection Using Artificial Intelligence, N/A

Defensive Publications Series

Software development relies on end-to-end (E2E) testing to ensure quality and reliability. However, E2E tests are often time-consuming and expensive. This disclosure describes automated techniques to comprehensively perform end-to-end (E2E) testing of software in an adaptive manner. Based on analyzing code changes, the techniques intelligently recommend a subset of tests to achieve targeted E2E testing. The tests enable developers to develop with more confidence. Employing machine learning (ML) models trained over a combined dataset of static code analysis, code coverage metrics from multiple system-under-test (SUT) configurations, historical bug data, and an integrated user feedback mechanism, the most relevant tests for …

Mastering Technology Audit Quality Assurance: A Framework For Auditors, Ramkumar Bharathan

Defensive Publications Series

Ramkumar's document presents a comprehensive framework (COBIT, NIST, FFIEC CAT, FFIEC IT Handbook) for establishing, implementing, and maturing a robust Technology Quality Assurance Review (QAR) program within internal audit, emphasizing its strategic imperative in today's complex digital landscape. Driven by escalating technology risks (cybersecurity, cloud, AI), evolving regulatory demands (FFIEC, COBIT, NIST CSF, GLBA, SOX, NYDFS, State Privacy Laws), and the need for reliable assurance, the framework details a systematic approach grounded in IIA Global Internal Audit Standards (specifically the QAIP) and integrated with key governance and control frameworks like COBIT and NIST CSF. It outlines core QAR principles, a …

Context-Aware Dynamic Diagram Generation For Enhanced Whiteboarding Sessions In An Online Meeting, Ram Mohan R, Rajarshee Dhar, Deepesh Arora

Defensive Publications Series

Proposed herein are techniques that provide for the ability to derive context for an online meeting from various contextual sources. During operation, the system proposed herein can predict a user's diagramming intent and then dynamically populate a whiteboarding palette with relevant diagrams based on the context of the online meeting.

Core Components Of Artificial Intelligence, Mahsa Shishegar Hamedani

Defensive Publications Series

Abstract:

This paper explores the core components of Artificial Intelligence (AI), focusing on the essential elements that allow AI systems to function autonomously. These components include Perception, Learning, Reasoning, Decision-Making, and Action, each playing a critical role in the development of intelligent systems. The paper discusses how these interconnected components work together to enable AI to process information, adapt to new situations, make informed decisions, and interact with the environment. Understanding these fundamental elements is crucial for advancing AI technology and its applications in various fields.

“Innovative Methodology For Full Human Head Transplantation: A Multidisciplinary Surgical And Neurological Approach”, George Karelias

Defensive Publications Series

This disclosure introduces an innovative concept for an implantable artificial intelligence (AI) neural interface designed to restore functional connectivity between a transplanted human head (or brain) and the recipient body by bridging the severed spinal cord. The system features a bi-directional neural implant equipped with AI-driven signal decoding, encoding, and adaptive learning capabilities, enabling both motor control and sensory feedback across the disconnection. Core components include neural sensor arrays, a machine-learning processor, stimulation modules, and optional regenerative and drug delivery systems. This approach addresses the longstanding barrier of neural reconnection in full head transplantation and holds promise for applications in …

Design Of An Artificial Intelligence System For Analyzing And Predicting User Behavior In Digital Spaces To Counter Soft Social Threats, Mahsa Mahsa Shishegar Hamedani

Defensive Publications Series

This invention introduces an AI-based system designed to analyze and predict user behavior in digital environments. By leveraging machine learning algorithms and behavioral data, the system identifies early signs of soft social threats such as misinformation and manipulation. It enables proactive threat detection, offering real-time analysis and predictive capabilities to enhance digital security and social stability across various platforms.

Protecting Security And Privacy In Mobile Ai Ecosystems, Zongyong (Vincent) Chen

Defensive Publications Series

This publication describes systems and methods for protecting security and privacy in mobile artificial intelligence (AI) ecosystems. The use of protected virtual machines (pVMs) serves to isolate the AI components of the ecosystem, ensuring both privacy and security for end-user data. By running the AI on a separate operating system (OS), even a compromise of the main OS does not threaten the integrity of the data and output of the AI. In addition, the AI can be implemented on separate hardware than that of the main OS, such as separate processor(s), memory, etc., making user data inaccessible to malicious actors.

Method To Increase Heat Dissipation Through Mda Heat Sink Metrology, Wei Leng Tee, Min Won Bae, Khin Yuupar Htaing, Thanakorn Sangrit

Defensive Publications Series

A hard disc drive (HDD) consists of many components such as a magnetic recording disc 103, base deck 102, and cover 101. In some embodiments, a cover 101 with a form-in-place gasket (FIPG) 104 and a base deck 102 are arranged to provide a sealed HDD. A final welding process is applied on final cover 100 to seal the drive hermetically. A magnetic recording disc 103 is rotatably supported within the housing. A data transducer is then controllably moved by an advance part to glide across the surface of the magnetic recording disc. Together with incorporated electronics linked to the …

Providing Dynamic Hello Timers And Adjacency Stability In Control Plane Protocols Using Extended Berkeley Packet Filter Techniques, Sreejith Sreekumaran, Navin Bhat, Jyothirmai Anekallu, Palak Juneja, Madhan Sankaranarayanan

Defensive Publications Series

A novel extended Berkeley Packet Filter (eBPF) -powered solution is proposed herein for managing dynamic hello packet timers in order to achieve seamless network dependability. The eBPF techniques proposed herein may provide for reducing convergence loss and improving protocol stability by automatically modifying timings in real-time based on network conditions, thereby assuring network resilience and responsiveness.

Method For Manufacturing Radial Brushes, Thomas Gyarmati

Defensive Publications Series

: The invention relates to a method for manufacturing radial brushes, in which the bristles, once inserted into the brush body, no longer require circumferential trimming. Nevertheless, the finished brush exhibits a diameter tolerance comparable to that of conventionally trimmed brushes.

By eliminating the cutting step, there is no bristle waste, which represents a significant advantage, particularly when using high-cost bristle materials.

The method enables the use of densely packed bristle bundles with an increased bristle content, thereby enhancing the brushing performance. Moreover, the method allows for high flexibility in the production of customer-specific brushes with variable characteristics.

A key …

Compliant Mechanism For Electronic Component Insertion, N/A

Defensive Publications Series

Repeatedly inserting electronic components into assemblies can cause substantial strain on human hands. This disclosure describes a simple, low-cost tool made of compliant material that provides dexterity and precise force control when inserting electronic components into assemblies. The described tool, a roughly two-dimensional structure, can be manufactured using low-cost extrusion processes. The flexibility of the compliant material enables the human technician or automatic mechanism to achieve granularity of control over forces applied to the extremes of the component through the torsional movement of the handle. The described tool reduces the ergonomic challenges of human insertion of components, improves the controllability …

Dual Inline Memory Module (Dimm) Insertion Tool, N/A

Defensive Publications Series

Repeatedly inserting electronic components into assemblies can cause substantial strain on human hands. Technicians who insert DIMM (dual inline memory modules) cards or similar components into motherboards hundreds-to-thousands of times daily can face a risk of fatigue or injury. This disclosure describes an ergonomic tool that enables safe and efficient insertion of electronic components such as DIMMs into their assemblies. The tool includes a groove to hold the DIMM or similar component. The obverse side of the groove is a flat, relatively wide surface that the technician can use to press down the DIMM. The wide surface area reduces the …

Liquid-Cooled Cold Plate Using Plastic Tubing, N/A

Defensive Publications Series

The OSFP (octal small form factor pluggable) transceiver is usable to connect a network device to optical fiber or copper cable. OSFP transceivers can occupy tight confines and consume substantial power, such that the heat they generate requires liquid (not air) cooling to dissipate. This disclosure describes techniques to use plastic (e.g., polyamide ) tubing to transport coolant to the cold plates of OSFP transceivers or other heat-generating equipment. A plastic-to-copper connector connects the plastic coolant tube to the copper pedestal. The supplanting of copper by plastic in coolant tubing increases flexibility and eliminates cold working of metal parts, resulting …

Estimating Wifi Coverage With Artificial Intelligence, N/A

Defensive Publications Series

Traditional techniques for estimating WiFi coverage and capacity rely on manual surveys and measurements, which are costly, time consuming, labor intensive, and do not scale with the region under coverage. This disclosure describes techniques that leverage machine learning to predict the coverage and capacity of a WiFi network given a building floor plan, placement and configuration of access points within the floor plan, and the distribution of users within the region of coverage. Existing floorplans (including placement of building materials, e.g., concrete, steel, wood, drywall, etc.), seating data, and coverage/capacity maps are used as data to train a machine learning …

Cache Efficient Assignment Of Serving Tasks, N/A

Defensive Publications Series

This disclosure describes techniques for cache efficient assignment of serving tasks (workers) to client requests. Per techniques of this disclosure, a distributed computing system is configured such that a particular client can connect only to a subset of workers of the distributed computing system. To further ensure cache efficiency, the selection of a particular worker of an assigned subset of workers for a client request is based on a hash value of a key generated from a worker identifier. In some implementations, worker selection may be performed based on a consistent hash, wherein the worker (service task) name or identifier, …

Wifi Access Point Placement And Configuration Using Reinforcement Learning Agents, N/A

Defensive Publications Series

Reliable WiFi coverage hinges on the strategic placement and accurate configuration of access points (APs). Traditional, manual AP placements are time consuming, costly, and suboptimal. This disclosure describes techniques that leverage Q-learning to determine optimal access point (AP) placements and configurations in an indoor environment. A reinforcement learning (RL) agent is trained using a digital representation of the indoor environment and a reward function that maximizes network coverage while minimizing the number of APs. The digital representation of the indoor environment includes floorplans and coverage maps. The trained RL agent can be deployed to optimize the placement and configuration of …

Wifi Access Point Placement And Configuration Using A* Search, N/A

Defensive Publications Series

Reliable WiFi coverage hinges on the strategic placement and accurate configuration of access points (APs). Traditional, manual AP placements are time consuming, costly, and suboptimal. This disclosure describes techniques that leverage the A* search algorithm to determine optimal AP placements/configurations. Given a set of AP placements/configurations and a metric that captures the difference between the network coverage of a particular AP placement and a target network coverage, an AP placement/configuration is found that optimizes WiFi coverage while minimizing the number of APs and their configuration complexities. The techniques advantageously discretize the search space by leveraging the facts that only a …

Support Agent Powered By Artificial Intelligence, Kishore Garg, Piyush, Madhumita Chakraborty

Defensive Publications Series

For cloud service providers or social media services that receive support requests from users or creators, case triaging and routing require human agents to sift manually via multiple tools and forms. This can introduce errors and result in an unacceptably high total resolution time (TRT). This disclosure describes an artificial intelligence (AI) agent that automates the handling of support requests (cases) raised by content creators or users of large services such as video sharing, social media, or other cloud-based services. By automatically performing authorization checks, routing cases, and remediating issues raised by users or creators, the AI agent can reduce …

Saving Idle Power And Engineering Hours By Powering Off/On Idle Ai Accelerators, N/A

Defensive Publications Series

The technology described in this paper relates to an artificial intelligence (AI) accelerator chip's application-specific integrated circuits (ASICs)’ power down for idle power reduction, and low-latency peripheral component interconnect express (PCIe) relink made on AI accelerator chips. Complete turning on or off of the AI accelerator chips may be accomplished by leveraging various software and/or hardware components in the AI accelerator systems. The specialized software and/or hardware may be configured to turn off the entirety or only certain components of each AI accelerator chip, such as application-specific integrated circuits (ASICs) embedded in the AI accelerator chip, and remove individual PCIe …

Enhanced Image Retrieval Under Scene Changes For Indoor Localization Using 3d Gaussian Splatting And Object Masking, Jongwon Lee, Hongsheng Yu, Chao Guo, Fengtao Fan

Defensive Publications Series

We propose a method for image retrieval in indoor scenes that are subject to perceptual changes caused by the temporal replacement of common household objects. To address this challenge, we propose a method that masks out these dynamic regions in either the query or database images, allowing retrieval based on the static scene elements. Additionally, we augment the database by rendering images from unseen viewpoints using 3D Gaussian splatting, thereby making candidate matches more likely to be retrieved from an augmented database. Each design choice is supported by the corresponding supporting experiments. Through evaluation on a dataset collected in several …

Dynamic Frame Sampling For Multimodal Large Language Model Video Understanding, Beltrán Labrador, Georgi Stephanov, Yury Stuken, Fabio Sulser, Ilia Akolzin, Olivier Siegenthaler, Ágoston Weisz, Manuel Tragut

Defensive Publications Series

Video understanding using multimodal large language models (LLMs) includes recognizing objects, actions, scenes, and extracting meaningful insights from video streams. To reduce computational burden, a subset of frames sampled from the video are fed to the LLM. However, this can reduce the accuracy of LLM inference when the video includes salient information in a small set of frames and is wasteful when the video has slow-moving scenes. This disclosure describes a dynamic subsampling technique that can be used to select the most salient frames with a higher likelihood. Specifically, attention-guided frame selection, 3D convolutional feature extraction, and entropy-based subspace projection …

Audio-Video Synchronization Using Machine Learning, Avichal Rakesh, Mozart Alexander Louis

Defensive Publications Series

Audio and video components of a video can fall out of alignment for a variety of reasons, e.g., since they are recorded separately, due to encoding errors, etc. The viewing experience is degraded when audio and video are out of sync. Manual synchronization is expensive. This disclosure describes the use of a machine learning model to determine synchronization information for audio and video streams that are to be synchronized. The audio and video streams are provided as input to a multimodal model, e.g., a large language model that can handle both audio and video. The model is tasked with outputting …

Scalable Real-Time Bias Detection And Mitigation In Ai-Based Application Evaluation, Puneet Mahajan

Defensive Publications Series

Employing artificial intelligence (AI) for evaluating AI applications poses a number of challenges, including the detection and mitigation of various biases. This disclosure describes a multi-faceted approach for dynamically identifying and mitigating biases in AI-based evaluation. Potential biases can be detected by analyzing inconsistencies and discrepancies between the outputs of the primary and auxiliary evaluator models via an ensemble learning framework, optionally assisted by human evaluators. Specific adversarial examples can be generated and used to identify hidden biases in the primary AI evaluator. Detected biases can be mitigated by employing reinforcement learning to make continuous real-time adjustments to counteract the …

Generating Content Recommendations For User Groups Using A Large Language Model, Sathya Anurag Siruguppa

Defensive Publications Series

The large number of available content options from streaming audio and video service providers poses a challenge for providing useful recommendations to users, especially when a group of users with diverse viewing interests want to collectively engage with content. This disclosure describes techniques that leverage a large language model to generate a feature dataset based on user-permitted metadata for individual users of a group of users and content metadata for available content. A composite rating for individual pieces of content is determined and suitable content for the group is identified.

Enhanced Text Entry Assistance On Smart Devices Using Generative Artificial Intelligence, Yuanbo Zhang, Yuqing Zhong, Yang Lu, Xu Liu, Shumin Zhai, Renjie Liu, Lan Wei, James Su, Haoyue Tong, Haicheng Sun, Fan Yang

Defensive Publications Series

Users entering text on smart devices can receive artificial intelligence (AI) assistance for spelling and grammar corrections, content generation, content refinement, exploration, etc. However, providing a coherent user interface (UI) for AI assistance across different applications and device form factors can be challenging. The disclosure describes techniques of text input assisted by generative AI that enable frictionless content creation, revision, and exploration using a software keyboard-based user interface that is consistent across applications and devices. A user experience (UX) flow that is cognizant of the ambiguity inherent in content creation and open-ended content revision enables human-AI interactivity. A suitable large …

Adaptive Multimodal Speech Compression With Generative Priors, Roger William Graves

Defensive Publications Series

This disclosure describes techniques that leverage generative artificial intelligence to efficiently compress and accurately synthesize speech. Speech compression is achieved with personalized adaptation and with low perceptual loss. Per the techniques, a raw audio recording of a user is transcribed to text and is also used by a generative artificial intelligence (AI) speech synthesis model to create a fine-tuned personalized model of the user’s speech. A low-fidelity quantized representation is generated of information such as prosody, amplitude, inflection, etc. The transcript, the quantized audio data, and the personalized, fine-tuned model are compressed and saved. The compressed data is provided to …

Simplified Digital Payments Without The Use Of A Payment App, Manasa Babu, Shay Seng

Defensive Publications Series

Making a digital payment requires users to have a smartphone or another device with a payment app and with internet connectivity. This can present barriers to adoption for some users such as those with limited device storage, weak or no internet connectivity, users of feature phones or the use of feature phones. This document describes the use of alternative authentication techniques such as biometrics, voice identifiers, one-time passwords, or QR codes that can be used to initiate and authorize digital payment. A merchant can request a paying user to provide any of these factors to initiate a payment request. The …

Electrodeposited Conductive Polymer Coatings And Methods Of Use, Wenyang Pan, Pinghung Wei, Zixuan Deng, Li Yao, Vanessa Tolosa, Zidong Li, Ru He

Defensive Publications Series

Provided herein are methods and compositions for producing a conductive polymer coating on hard and soft electrode surfaces. The conductive polymer materials include 3,4-ethylenedioxythiophene monomer or a derivative thereof, and a dopant. The conductive polymer can include additives including a surfactant or crosslinker. Also described herein are methods of making conductive polymer coatings. Further described herein are molded products including the conductive polymer material as described herein and wearable devices including the molded products.

Multimode Link Optimization Based On Return Loss Detection At The Transmitter Side, Fabio Bottoni, Marco Mattarei, Carlo Mariotti

Defensive Publications Series

High-speed multimode fiber (MMF) links provide high bandwidth and a cost-effective solution for short to medium-range data transmission. However, optical return loss (ORL) can be a significant factor degrading the performance of MMF links. More specifically, the “ORL issue” is dynamic: a link might initially be operational but degrades after a number of insertions, thus causing the link to go down. The techniques presented herein leverage a return loss (RL) detector to optimize high-speed MMF links, thus enhancing the capability of switches to continuously monitor reflections.

Improved Process For The Preparation Of 4-Substituted 1, 4- Dihydropyridine, Anonymous

Defensive Publications Series

The present publication relates to preparation of 4-substituted 1, 4-dihydropyrine of Formula (I), more particularly to the preparation of Felodipine and Nisoldipine.