Digital Commons Network^™

Virtual Private Cloud Powered By Artificial Intelligence, Dhandapani Shanmugam

Defensive Publications Series

It is often the case that multiple service providers, e.g., e-commerce services, payment gateways, delivery services, etc., are hosted on a single cloud service provider but have interfaces only to the public internet. In such a case, a single transaction can involve multiple public calls or egresses from the private cloud (more secure) to the public internet (less secure), adding to the cost and latency of the transaction and reducing the efficiency and security of the transaction. This disclosure describes artificial intelligence (AI) techniques to analyze the origin and destination IP addresses of communications that arise from applications, virtual machines, …

Summarization Framework For Sharing And Contextual Previews, Thomas Nont Boonsiri, Siddhartha Sivakumar, Salvador Guerrero Ramos, Lan Julia Liang, Elizabeth Martinez

Defensive Publications Series

A system is disclosed that provides a summary of web-based content. Web-based content comprising text-based content is received by a browser. A portion of the text-based content is processed through a large language model to determine a summary of the portion of the text-based content. The portion is determined according to a threshold token limit. The summary is provided for display via a user interface.

Implementations of the present disclosure are generally directed to user device generation of a summary of web-based content (webpages, text, images, media files, etc.) via a generative model, such as a large language model (LLM). …

Method And System For Tap To Customer’S Own Device To Authenticate A Card, Esha Dutta, Thomas D. Rapkoch, Rhodora Enrijoy H. Martin, Mark Rigby, Geraldine Mitchley

Defensive Publications Series

Present disclosure provides a method and system for tap to customer’s own device to authenticate a card and this process is implemented in various use cases such as tap to activate a card, tap to confirm a card, tap to add/provision a card and the like. For instance, the tap to authenticate a card includes sending a request by the user to load new card credentials to wallet operator application that transmits a request to perform a financial activity (eg: provisioning card) including EMV cryptogram to issuer mobile gateway. The issuer mobile gateway requests for the issuance of a token …

Cross-Lingual Prosody Preservation In Speech Generation Applications, Daniel V. Klein, Paul Mccartney

Defensive Publications Series

This disclosure describes a data collection approach to obtain a high-quality, paired prosody dataset to support cross-lingual prosody translation in voice dubbing and other speech generation applications. Current datasets lack the nuanced prosodic features that are essential for natural and expressive cross-lingual speech synthesis, resulting in synthetic speech that often sounds flat or lacks emotional resonance. Per techniques of this disclosure, audio spoken by professional voice actors or other individuals proficient in a language is recorded in various languages. Paired recordings of matching content are available in the collected dataset. The recordings capture nuanced prosody such as rhythm, emotion, and …

Safe, Robust, And Graceful Handling Of Stale Or Offline Remote Sensors, Eric Burger, Ernie Yang

Defensive Publications Series

Remote sensors for a control device such as a thermostat enable it to adjust the temperature at the location of the occupants of the home or office. If the thermostat misses one or more remote temperature readings, temperature control can be unsatisfactory. This disclosure describes techniques that forestall thermostat problems such as stale sensors and outdated readings arising from lapses in connectivity between a thermostat and its remote sensors. A robust, fail-safe mechanism is described that detects and mitigates connectivity problems by reverting control to a local temperature sensor if the requested remote sensors provide invalid or stale data. A …

Rack Bus Bar Connector With Built-In Hotswap Features, N/A

Defensive Publications Series

The technology described in this paper relates to a hotswap method for high-voltage rack systems such as ±400 V or 0-800 V systems by leveraging controlled and sequenced insertions of different power connector pins to achieve safe and efficient power connector insertion and removal. The power connector may include multiple pins of various lengths, which allows each pin to establish an electrical connection to a bus bar in a pre-defined order. For example, ground pins and pre-charger pins are electrically connected to the bus bar before the main power pins are electrically connected to the bus bar. The pre-charger pins …

Wireless Broadcast Sharing In Multi-Controllermulti-User Environments, Henry Salvatore Savage

Defensive Publications Series

This publication describes a vehicle computing system (e.g., a head unit, in-vehicle entertainment system, infotainment system, etc.) that may automatically switch from outputting audio to a single computing device using a wireless point-to-point protocol to outputting audio to two or more computing devices using a wireless broadcast protocol. The vehicle computing system may include multiple modules for controlling wireless connections (e.g., multiple BLUETOOTH controllers, WIFI controllers, etc.). Each user in the vehicle may have wireless listening devices (e.g., wireless headphones, speakers, earbuds, etc.) connected to at least one of the wireless controllers. For example, a user may have wireless headphones …

Assessing Road Lighting Conditions Using Speed Variation Analysis, Priyank Garg

Defensive Publications Series

Currently, digital maps and navigation systems do not provide sufficient information about lighting conditions for roads and lanes. The lack of reliable information can negatively impact route selection and driving speed if a road is poorly lit). This disclosure describes techniques to automatically determine the lighting level of a road or road segment by analyzing speed variation of vehicles that traverse it at late night (when it is dark) and early morning (when there is adequate light), under otherwise similar conditions (traffic levels, road surface quality, etc.). A significant difference in speed between these periods suggests poor road lighting, as …

Automatically Surfacing Navigation Assistance In Driving Contexts, Ilya Levin, Yarin Aidelman, Arie Orlovsky, Or Abramov, Itamar Besor

Defensive Publications Series

While digital map and navigation applications provide useful assistance when a user is driving, the user needs to explicitly seek guidance from these applications. If the user chooses to drive without assistance, the user may miss out on useful information such as congestion alerts, faster alternative routes, etc. This disclosure describes techniques, implemented with user permission, to automatically detect a driving context and proactively surface useful notifications. The driving context can be detected based on analyzing connection type and name associated with Bluetooth connections established by a mobile device that runs the navigation application. Additional user-permitted factors such as current …

High Density Interconnect Chip On Board, N/A

Defensive Publications Series

This paper is generally directed to systems, methods, and devices that facilitate direct attachment of a high density silicon or semiconductor chip to an embedded coreless printed circuit board (“PCB”). The technology disclosed provides an assembly and method for manufacturing directed to a new coreless PCB (“CPCB”) with an embedded coreless substrate, as well as systems incorporating CPCBs. The CPCB enables direct bonding of a high density silicon chip (“HDSC”) to the CPCB as the last step in the system manufacturing process. This is in contrast to the traditional manufacturing process of first bonding an HDSC (with or without an …

Securing The Future Of Ai: A Deep Compliance Review Of Anthropic, Google Deepmind, And Openai Under Soc 2, Iso 27001, And Nist, Ramkumar Bharathan

Defensive Publications Series

As foundation models like Claude, Gemini, and OpenAI become integral to enterprise AI solutions, the need for rigorous, standards-aligned security evaluation grows. This paper presents an expert-level comparative audit of the SOC 2 controls implemented by Anthropic, Google DeepMind, and OpenAI, drawing from public attestations, model cards, and technical documentation. Going beyond traditional compliance, the assessment incorporates frameworks from ISO 27001, NIST, and AI-specific risk models to benchmark each provider’s security maturity across five domains: organizational governance, infrastructure controls, operational security, data protection, and AI-specific safeguards.

Key findings indicate that Google demonstrates enterprise-grade maturity with advanced automation, zero-trust architecture, and …

Method To Adapt Different Downhole Drilling Tool Bus Systems, Baker Hughes Company

Defensive Publications Series

The invention relates to a constructive method for adapting a downhole tool to different downhole bus systems in order to be used in a bottom hole assembly with the respective downhole bus systems used. The adaptation relates to a possibly necessary mechanical adaptation to the different connector systems of the bus systems and to the adaptation of the downhole tool to the respective power and communication system used.

Filtering Low Quality Query Recommendations, Jap Leen Kaur Jolly, Karan Rai, Paritosh Shroff, Stephanie Saade

Defensive Publications Series

Some search engines provide query recommendations that suggest additional queries related to a user’s original query. Low-quality query recommendations can be irrelevant to the user’s information need and degrade the user experience. This disclosure describes techniques to filter low-quality query recommendations such that they are not displayed to users. The filtering is performed by evaluating a set of candidate query recommendations based on query recommendability scores along three quality-related dimensions - content, formulation, and flavor. Additionally, updateworthiness for a query recommendation can also be used. Implementation of the techniques can raise the quality, utility, and trustworthiness of query recommendations and …

Service Affecting System Awareness For Distributed Decision Making, Anonymous

Defensive Publications Series

The disclosure presents a comprehensive framework designed to manage algorithmic updates in optical Interface Modules (IM), ensuring network stability during software upgrades by delaying potentially disruptive changes until a coordinated recalibration occurs. This method introduces robust communication between IM cards and network element controllers, facilitating precise control over calibration processes, including direction-specific distinctions. It addresses multiple operational scenarios—such as Optical Channel Monitoring (OCM) data handling, Amplified Spontaneous Emission (ASE) adjustments, Raman pump calibration, and band-specific updates in multi-band deployments. Furthermore, the framework supports opportunistic service-affecting upgrades, like FPGA changes, by aligning them with suitable maintenance windows, while historical calibration data …

Thermal Drift Compensation And Margin In Autofocus Cameras For Streaming Devices

Defensive Publications Series

The present disclosure generally relates to autofocusing camera technology and, more particularly, focusing in response to thermal expansion

Enhanced Offline Emergency Sos Using Distributed Channeling, Sudeep Chauhan, Sumukh Ashok Shevde

Defensive Publications Series

This document describes techniques that enhance the reliability of mobile device safety services by leveraging existing peer-to-peer connectivity protocols, such as BLUETOOTH Low Energy (BLE)-based proximity pairing. When a mobile device (e.g., a smartphone, smart watch, etc.) detects a safety trigger (e.g. SOS, hard fall, etc.) and is unable to connect to the Internet, the device uses peer-to-peer connectivity protocols, such as proximity pairing (e.g., Fast Pair), which work across different device manufacturers, to broadcast a “distress signal” to other devices in the vicinity. The signal may include only a few bits corresponding to specific status indicators and, depending on …

Providing Residence Time Over Srv6 Transport For Precision Time Protocol, Anonymous

Defensive Publications Series

This disclosure outlines a streamlined method for precise timing synchronization using Precision Time Protocol (PTP) over Segment Routing IPv6 (SRv6) networks lacking native PTP support. Specifically designed for timing-sensitive 5G applications, the approach leverages SRv6 headers to efficiently manage timing information without extensive packet-level processing. Provider Edge (PE) devices identify and flag PTP packets, intermediate nodes calculate and embed packet residence times into the Segment Routing Header (SRH), and egress nodes update this information into the PTP payload. This method significantly reduces complexity and enhances scalability in SRv6-enabled transport networks.

Method And System For Performing Electronic Battery Assembly Sensing, Baker Hughes Company

Defensive Publications Series

The present disclosure is directed at methods, systems, and techniques for performing electronic battery assembly sensing.

Computing Device Network Configuration For Power-Saving, Jean-Philippe Cormier

Defensive Publications Series

This document describes techniques that enable a mobile computing device (e.g., a smartphone, smart watch, wearable device, augmented reality glasses, etc.) connected to a network to adjust or disable features (e.g., data for certain applications, power-hungry radios, wake-up cycles from sleep, etc.) based on battery status. The device presents the features to the cellular network (e.g., 5G, LTE, 3G, etc.) using feature group indicators (FGIs). In some cases, such as when the device is not charging, the device will not share certain measurements to the cellular network such as SON/ANR (extra network measurements to improve network quality) and will instead …

Synthetic Data Generation For Training An Audio-Video Translation Model, Daniel V. Klein, Paul Mccartney

Defensive Publications Series

Synthetic data is generated to train an audio-video translation model that aims to achieve synchronized audiovisual outputs. Existing translation systems struggle to maintain natural synchronization between translated speech and corresponding visual elements, resulting in awkward timing and reduced content quality. To address this, audio, video, and onscreen text inputs are processed by dedicated models responsible for translation and adaptation tasks. These inputs are translated and adjusted to maintain consistency in timing, tone, and visual synchronization. Manual corrections are applied through human-in-the-loop (HITL) checkpoints, ensuring quality control and addressing nuanced issues that automated processes might miss. The architecture incorporates feedback loops, …

Intelligent Generation Of Shortcuts And Interactive Components, Tim Wantland, Aiko Nakano, Abraham Wallin

Defensive Publications Series

A computing device may dynamically generate a graphical user interface that includes content (e.g., actions, shortcuts, widgets, etc.) related to the current context (e.g., historical user patterns, location, date and time, etc.). The computing device (e.g., a wearable device such as a watch, ring, glasses, etc., a mobile device such as a smart phone, laptop, etc. or other type of computing device) may use the current context to predict what a user of the computing device may want to do (e.g., predicted actions) so that, rather than requiring a user to navigate through several levels of a user interface, the …

Personalized Semantics In Queries, Tim Whalin, Nick Krasney, Gaurav Gupta

Defensive Publications Series

The technology described in this paper relates to providing personalized query results to users based on inferred semantics associated with the query and the user submitting the query. The query may be a search query, a navigation query, an agentic query, or the like. The semantics include, for example, personally meaningful concepts associated with the user submitting the query. The inferred semantics are determined using an artificial intelligence (“AI”) model. The queries can be input in a natural language form. The queries may include user-specific context, such as references to user-specific data related to historical queries, calendar entries, contact information, …

Expanding Data Classification For Regulatory Compliance Using Natural Language Processing, Assaf Namer

Defensive Publications Series

The terms of use (ToS) and end-user license agreements (EULA) can vary across software as a service (SaaS) providers used by an enterprise, posing challenges in maintaining consistent data privacy, security, and residency standards. This disclosure describes techniques that leverage natural language processing (NLP) to continually evaluate and classify the ToS/EULA of SaaS providers to classify data that is owned by the enterprise and handled by a SaaS provider. NLP-based techniques are used to automatically evaluate processing, usage, residency, handling, etc. of data by SaaS providers to ensure compliance. The described data governance/ protection techniques recognize that artificial intelligence (AI) …

Large Reasoning Models For Route Planning In A Digital Map, Victor Carbune, Sandro Feuz

Defensive Publications Series

This disclosure describes route-planning techniques that leverage large reasoning models to predict route traversal costs based on user-generated content (UGC) and data available in digital map applications. Large reinforcement models, which can advantageously be trained using reinforcement learning (RL), can use sparse and hard-to-interpret signals arising from UGC to extend and to customize routing algorithms for best-path or travel-time queries. Per the techniques, large reasoning models steer the routing algorithm to more effectively estimate the optimal route by introducing user-generated data (e.g., reviews, comments, photos, etc.) into the reasoning chains of a large reasoning model; by causing the large reasoning …

A Method And A Product Design System For User Focus Traps, Emmanuel Addo Narh, Thirumalla Srinivas, Alicia Mae Frausto

Defensive Publications Series

The present disclosure provides a method and a product design system for user focus traps. The processor of product design system uses react functionality hook useFocusTrap in the product design system Nova React library, providing comprehensive insights into the code references and pseudo code used for the implementation of these hooks. The product design system is an all-encompassing toolkit designed to revolutionize the product design and development process. The product design system provides a centralized library of assets and guidelines to ensure consistency and efficiency.

Thermoelectric Cooling Of High Temperature Memory Flash Devices In Multi-Chip Modules, Baker Hughes Company

Defensive Publications Series

The amount of data being recorded by downhole drilling and measurement services is constantly increasing, because customers request more often images, comprehensive sensor data as well as determined raw values. Therefore, it is essential ensure huge data storage capabilities while operating within this high temperature (>150°C) and high vibration environment. Smaller silicon process structures of DIEs are necessary to fulfill the storage capability as well as the corresponding high-speed requirements. This leads to challenges, especially for high temperature applications with memory NAND flash technology that active cooling becomes imperative – here realized by using thermoelectric cooling.

Intelligent Connector Cap For Self-Testing Of Equipment, Baker Hughes Company

Defensive Publications Series

Many connectors of special equipment have a complex standard interfaces on its pins (USB, CAN, ETH, etc). For verification of functionality interfaces is necessary a communication partner (an additional test box or target device/tool), what is expensive and not overall available.

Device Information As Context Input For Custom Response Generation By A Virtual Assistant Or Chatbot, Karen Kaushansky, Tony Sheeder, Junius Gunaratne, Jon Bloom, Ilya Pyatigorskiy, Kinda Akash, Ryan Pham

Defensive Publications Series

Chatbots and virtual assistants powered by a multimodal large language model (LLM) can respond to a wide range of queries and can process input in text, audio, video, or other formats. However, when deployed on a user device such as a smartphone, it is difficult to determine what data to capture and analyze, and the modality in which to provide the response. This disclosure describes the use of contextual signals such as device capability, device settings and state, connections with other devices, etc., obtained and used with user permission, to detect the input to be analyzed by the LLM and …

System And Method For Maintaining Global Ordering Of Co-Related Events In A Distributed Event Streaming Platform, Alok Roy

Defensive Publications Series

The present disclosure provides a method and a system for maintaining global ordering of co-related events in a distributed event streaming platform. The method comprises generating a group identifier for co-related events at a producer system and appending a sequence number to each event based on their order within the group. The message sequences are then transmitted to a distributed event streaming platform. A consumer system monitors the metadata of received message sequences to detect the group identifier and creates a temporary lookup table to track received sequence numbers. If a non-consecutive sequence number is detected, the consumer system pauses …

Context-Aware Autonomous Interaction Using Generative Artificial Intelligence, Ajay Prasad

Defensive Publications Series

Users are often overwhelmed by a steady stream of notifications from their personal devices. Solutions such as a do not disturb mode, canned automated responses, etc. suffer from limited scope and lack of integration with other apps. Such solutions are not seamless and are not sufficiently customizable. This disclosure describes an artificial intelligence (AI) powered virtual assistant mode for user devices such as smartphones. With user permission, an AI-powered virtual assistant for a device in this mode is configured to automatically receive and analyze incoming communication, initiate actions, interact with other applications, and to provide the user a summary of …