Digital Commons Network^™

Interactive Tool For Researching Large Unstructured Document Collections, Niv Efron, Daniel Chirpich, Jim Albrecht, Roni Rabin, Guy Ronen, Amos Yoffe, Shlomo Urbach, Tali Rosen Shoham

Defensive Publications Series

Reviewing large document collections is an activity that arises commonly in certain professional contexts such as investigative journalism. Such document collections can arise in many use contexts such as investigative journalism; academic research; litigation, arbitration or other legal context; audit; research using document archives; etc. The collections may include a large number of documents, including scanned images of documents or handwritten documents, and are often devoid of structure or organization. This makes it difficult to sift through such collections and identify important pieces of information. This disclosure describes a tool that enables easier access to such collections and features that …

End-To-End Encrypted Video With Adaptive Bit Rates, Anonymous

Defensive Publications Series

Server-side adaptive bit rate (ABR) encoding is a technique used to match video bandwidth to a user’s channel bandwidth. In ABR, an uploader streams in the video to a server at a high quality, while the downloader streams from the server the highest quality compatible with its bandwidth. As currently practiced, ABR is incompatible with end-to-end encryption since the server needs to decrypt the uploaded video to generate multiple lower bandwidth versions of the video.

This disclosure describes techniques for end-to-end encryption of video while making the video available to a receiver at multiple bit rates. The sender uploads to …

Simplified Ai Powered Dialing Plan Handler In Unified Communications Manager, Anupam Mukherjee

Defensive Publications Series

In a unified communications (UC) environment, the Call Manager (UCM) application contains a call control layer in which a digit analysis module supports the transformation of called and calling numbers and the determination of a routing decision. Such activities are typically complex and computationally intensive. To address such challenges techniques are presented herein that leverage artificial intelligence (AI) models (such as Artificial Neural Network (ANN) models) that are, among other things, customized and trained (based on, for example, customer requirements) in a cloud to, among other things, augment the number transformation and routing decision processes. Under aspects of the techniques …

Wireless Link Aggregation To Provide An Efficient And Smooth Multi-Protocol Data Link, Vincent Cuissard, Amine Choukir, Domenico Ficara, Arun Khanna

Defensive Publications Series

When connected to a cellular (e.g., 5G) network a device can receive/transmit Internet Protocol (IP) traffic via the packet data protocol (PDP). When connected to a Wi-Fi network a device can receive/transmit IP traffic via the Institute of Electrical and Electronics Engineers (IEEE) 802.11 protocol. This proposal provides for combining the capabilities of both cellular and Wi-Fi via a Wireless Link Aggregation (WLA) technique that allows for seamless roaming, no IP address change and no gateway change for a mobile device that is switched between access links, and enhanced experiences for users.

Password-Less Continuous Multifactor Authentication (Cfma) For Wireless Networks, Vinay Saini, Jerome Henry, Tim Szigeti, Robert Barton

Defensive Publications Series

The new generation of wireless networks can involve a mix of radio technologies, especially for industrial environments. As devices roam back and forth between different radio networks, it is very difficult to continually monitor the security posture and identity of devices connected to the network. Presented herein are techniques that involve the combination of a Device-Generated Trust Card and a Network-Generated Trust Card that can be used to validate device identity (e.g., an Internet of Things (IoT) device identity) and behavior using a continuous Multi-Factor Authentication (cMFA) structure.

Enterprise Private 5g Self-Healing Network, Abhishek Dhammawat, Rajesh I V, Sri Gundavelli

Defensive Publications Series

With the advent of Internet-of-Things (IoT) devices being utilized in enterprise deployments, industry is heading towards the deployment of Fifth Generation (5G) technologies in enterprise networks with regulators around the globe opening private as well as shared spectrum to facilitate 5G usage. In enterprise private 5G deployments with use-cases such as factory automation, etc. there is need for minimal service disruptions. Additionally, the UE/subscriber session scale will be increased for 5G deployments and the number of applications having critical use-cases with low latency and high bandwidth will also increase. With all these niches and for supporting URLLC use-cases comes the …

Password-Less Continuous Multifactor Authentication (Cfma) For Wireless Networks, Vinay Saini, Jerome Henry, Tim Szigeti, Robert Barton

Defensive Publications Series

The new generation of wireless networks can involve a mix of radio technologies, especially for industrial environments. As devices roam back and forth between different radio networks, it is very difficult to continually monitor the security posture and identity of devices connected to the network. Presented herein are techniques that involve the combination of a Device-Generated Trust Card and a Network-Generated Trust Card that can be used to validate device identity (e.g., an Internet of Things (IoT) device identity) and behavior using a continuous Multi-Factor Authentication (cMFA) structure.

Application Specific 5g Ran Split Options In A Single Radio Unit, Shailender Potharaju, Fred Anderson, Clark Carty, Indermeet Gandhi

Defensive Publications Series

Various technical specifications, such as Third Generation Partnership Project (3GPP) and Open Radio Access Network (O-RAN) specifications, define various functional splits that can be utilized for a Fifth Generation RAN (5G-RAN). For example, Split option 2 and Split option 7.2 are the typical implementations, and Split option 6 (Physical layer/femto application platform interface (PHY/FAPI)) is likely to be prevalent soon. Split option 7.2 is fronthaul heavy and involves strict latency budgets. For example, Split option 7.2 needs roughly four times the bandwidth of Split option 2/6 for the same amount of user plane throughput. Presented herein is a system in …

Component Conscious Thermal Imaging System, Min Zhu, Junnan Lin

Defensive Publications Series

Hardware components can experience failures, such as component shorts, during different manufacturing stages as well as in the field once such products are deployed. Presented herein is a new component conscious thermal imaging system that can increase hardware debug effectiveness for determining component failures related to thermal events. In one instance, the system can be used during manufacturing in order to setup a component thermal database that can be used to identify thermal related quality events in the early stages of component manufacturing. In another instance, the system can be used in the field to debug component failures for deployed …

An Intent-Based Caching System For Constrained Internet-Of-Things Environments, Dima Dababneh, Sean Ceballos-Mcgee, Connie Sidberry, Robert Barton

Defensive Publications Series

Internet-of-Things (IoT) environments often have very constrained edge devices, with limited memory, storage, and compute power. Additionally, these environments can be highly distributed. In many cases, edge devices/applications need to access content repeatedly from a cloud service or higher-tiered application. Because device and network connections can become constrained, it is desirable to cache objects, files, etc. local to the edge devices. However, caching systems and cache replacement algorithms may not consider the criticality or requirements of one application over another. Techniques presented herein provide for adapting well-known caching mechanisms for utilization within environments involving constrained networks. Further, intent-based classification techniques …

Onboarding And Provisioning Automation For Mesh Access Points, Vincent Cuissard, Samuel Pasquier, Alessandro Erta, Luca Bisti, Amine Choukir, Domenico Ficara

Defensive Publications Series

Mesh networks can be implemented in many different scenarios in which direct cabling of access points may not be practical or may not make financial sense. For example, in Internet of Things (IoT) deployments, access points may be located far from each other with very limited physical access. Presented herein are techniques to leverage hardware and software features available on mesh access points to avoid manual configuration in the field.

Nozzle Assembly With Single Inlet And Dual Outlets For Roll Cooling Spray Headers, Tudor Piroteala, Mitchell G. Roux, Louis M. Nazro, Frank Su, Luiz Porto, David A. Gaensbauer

Defensive Publications Series

No abstract provided.

Training A Ranking Model, Donald Arthur Metzler Jr, Xuanhui Wang, Marc Alexander Najork, Michael Bendersky

Defensive Publications Series

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a ranking machine learning model. In one aspect, a method includes the actions of receiving training data for a ranking machine learning model, the training data including training examples, and each training example including data identifying: a search query, result documents from a result list for the search query, and a result document that was selected by a user from the result list, receiving position data for each training example in the training data, the position data identifying a respective position of the selected result …

Networkless Mobile Payments With Minimal Changes In Trusted Execution Environments, Arun Samudrala, Dilip Padmanabhan, Abhijith Bera, Pankaj Gupta, Hemanth Sambrani, Will Drewry

Defensive Publications Series

Peer-to-peer mobile digital payments can be made in the absence of a network as follows: the receiver of funds verifies the availability of on-device balance by examining the prior, authenticated, transaction records of the sender. A new transaction record including the transaction amount is created, made immutable and secure using cryptographic techniques, and is stored at both sender and receiver. When either sender or receiver regains network connectivity, the transaction is settled with the original provider of the on-device balance, e.g., a financial institution. The integrity of the records of offline transactions, e.g., made in the absence of a mobile …

Method And System For Accessible Data Visualization On A Web Platform, Frank Elavsky

Defensive Publications Series

The present disclosure relates to a method and system for accessible data visualization on a web platform. At present keyboard limitations exist in a browser that do not permit full access to a data visualization. Therefore, in the present disclosure when the user is focusing a node, all of the previously mentioned interactivity rules apply to it, based on the kind of chart element it represents. When the user takes a keyboard action on that node, it is simulated as though that action took place in the chart space. So, the controller of the present disclosure renders what is necessary …

Application Launching During Application Installation On A Mobile Device, Vitor Baccetti, Matthew Patterson, Lee Gao, Phillip Oertel, Marcel Folaron, Jamal Eason, Yurii Zubrytskyi, Oleksiy Buynytskyy, Parul Soi, Todd Kennedy

Defensive Publications Series

This publication describes techniques to allow an application to launch on a mobile device, for example, a smartphone, once the device downloads a portion of a package file associated with the application. The package file includes the initial data necessary to launch the application. The initial data can be referred to as the “idle nugget” and can include executables, objects, images, audio, and/or geometry files. The mobile device is not required to download the remaining package file data before a user can initially launch the application. As a result, the idle nugget can be prioritized by an application distribution service …

Thermal Mitigation At User Equipment Based On Ambient Temperature, Madhusudan Kinthada Venkata, Shivank Nayak, Siddharth Ray

Defensive Publications Series

User equipment (UE) or other user devices employ thermal mitigation schemes that are responsive to the ambient temperature of the device. Such schemes can employ different levels or modes of thermal mitigation techniques based on ambient temperature, the thresholds that trigger the switch between levels or modes may be scaled based on ambient temperature, or a combination thereof.

Predictive Modeling Techniques For Adaptive Face Retouching, Wei Hong, Shengyi Lin, Li Wei

Defensive Publications Series

This publication describes predictive modeling techniques for adaptive face retouching according to a predictive model on a computing device. The adaptive face retouching enables preservation of important features that may depict diversity, persona, and uniqueness in images where retouching is applied by a computer. In the techniques, an algorithm may be implemented to balance image attributes from a variety of sources and categories. The algorithm may further reduce unnecessary, redundant processing on the computing device by storing computed attributes within a facial attribute database. As such, diverse facial features may be maintained while allowing adequate face retouching without requiring overuse …

Exercise Mode For Mobile Devices, Michael Depasquale

Defensive Publications Series

This publication describes methods and techniques for placing a mobile device (e.g., a smartphone, a tablet) in an Exercise Mode. The Exercise Mode is an On-Off device setting enabling users to focus on their exercises by temporarily suppressing incoming notifications (e.g., declining incoming phone calls, silencing text messages, ignoring alerts) when activated. In aspects, the Exercise Mode is activated either manually by a user or automatically by an Exercise Manager. The Exercise Manager is an algorithm (e.g., machine-learned algorithm) configured to analyze a plurality of signals generated by sensors (e.g., inertial sensors, location sensors, heart rate monitors) from one or …

Integration Of Non-Driving Activities Into Piloted Time Windows, Axel Unger

Defensive Publications Series

No abstract provided.

Preemptive Edge Application Instantiation, Mehmet Ozkan

Defensive Publications Series

Cloud-edge computing is a paradigm that brings servers physically closer to end-users, e.g., by co-locating servers and cellular towers. Cloud-edge computing can reduce latency and improve performance for mobile applications. For cloud-edge computing to deliver a superior user experience, the location of application instantiation is optimally selected to be close to the user. However, as a user moves, an initially optimal location becomes unsuitable. This disclosure describes techniques to predict the next possible edge zones for a cloud-edge application to maintain performance and service continuity as the mobile device moves. The techniques leverage factors such as mobility history, network characteristics, …

Intent Prediction Based On Contextual Factors For Better Automatic Speech Recognition, Diamantino Caseiro, Zelin Wu, Petar Aleksic, Era Jain

Defensive Publications Series

Automatic speech recognition (ASR) machine learning models are used to recognize spoken commands or queries from users. End-to-end ASR models, which directly map a sequence of input acoustic features into a sequence of words, greatly simplify ASR system building and maintenance. This disclosure describes techniques to improve the performance of end-to-end ASR models by providing predicted user intents as additional inputs. Intent prediction vectors or intent embedding is generated based on user-permitted contextual features using a trained intent prediction network (IPN). The IPN can be trained independently from the ASR model or jointly with the ASR model. Training of the …

Configuration-Adaptive Edge-Assisted Industrial Wireless Cameras With Deep Reinforcement Learning, Hp Inc

Defensive Publications Series

Embodiments of the present disclosure provide a method and system for designing and implementing

an industrial smart camera that uses low-power wireless communications and edge computing to

achieve cordless and energy-efficient visual sensing. The smart camera applies deep reinforcement

learning to adapt the camera configuration to maintain the desired visual sensing performance with the

minimum energy consumption under dynamic variations of application requirement and wireless

channel conditions.

Image Deblurring According To Facially Recognized Locations Within The Image, Wei Hong, Li Wei, Xuhui Jia, Marius Renn, Yukun Zhu

Defensive Publications Series

This publication describes techniques for image deblurring according to a facially recognized locations within the image. An algorithm may use facial detection and recognition to selectively sharpen aspects of faces within an image and the surrounding area associated with the facial detection. In one or more aspects, the selectivity of sharpening improves the computational load and other aspects of image provision to improve overall computer function, power consumption, and user experience. Individual faces within the image may be cropped or thumbnailed, providing portions of the image that include the faces. Counterpart images associated with the individual faces may be found …

Personalized Permissions For Content Recording And Streaming, Eric Cosky

Defensive Publications Series

People in shared online environments, e.g., video conferences, shared multiplayer environments, live streaming, etc. may have concerns that others can record their transmitted content without their knowledge or stream an online event to unknown destinations without permission. Not only is this a privacy risk for the individuals being recorded, the streaming of background music and effects with licensing restrictions can violate property rights of the holders of the content rights.

This disclosure describes techniques that enable participants that transmit content online in a shared environment to mitigate the risks of surreptitious recording and/or streaming of their content. With user permission …

Using Crash Detection And Alert Beacons On An Electronic Device To Increase Survival Of Snow Immersion Victims, Wei Wu, Xiangjun Zhao

Defensive Publications Series

This publication describes systems and techniques for crash detection on an electronic device to trigger transmitting alert beacons with the electronic device. The alert beacons may be received by nearby electronic devices to alert the users of those electronic devices of a snow immersion victim. A crash is detected by monitoring one or more sensors on an electronic device for a tumbling or inversion of the electronic device that indicates a snow immersion accident. Additionally, location and weather may also be monitored to detect a snow immersion accident and support a crash detection. After a crash is detected, a user …

The System To Exchange Information Confidentially Over Networks, Punarjeewa Abeysekera

Defensive Publications Series

This disclosure describes a method to exchange information confidentially over the networks between two computers. Two geographically stationary computers communicate with each other. Both of the computers have static IP addresses. Also, both of them have each other’s IP address before the communication session. Therefore, any one of them can start communication. Once the communication has started, they interact one after the other – in turns. In each interaction, one of the computers will send a data unit in an encrypted form to the other computer. This interaction between these two computers will continue until one of the computers terminates …

Intuitive Modeling Of Voxel-Based 3d Objects, Luis Quesada Torres

Defensive Publications Series

Voxels (volumetric pixels) are virtual cubes that can be used to construct, model, or analyze virtual three-dimensional objects. Voxels find extensive use in video game design, animation, three-dimensional printing, etc. Software modeling of voxel objects is difficult for a designer working on a two-dimensional computer screen because voxel placement in the third dimension (perpendicular to the computer screen) can be ambiguous or obscured.

This disclosure describes a simple, low-cost, and expandable hardware interface and accompanying software to intuitively model or define voxel-based 3D objects. The base has a matrix of connectors. A number of cubic pieces or physical voxels of …

Computing-Device Clipboards Accessible Via Audible Input, Dana Dahlstrom

Defensive Publications Series

This publication describes methods and techniques, implemented on computing devices, directed at enhancing a user experience associated with copying and pasting items, including text, images, and other media. In aspects, a computing device includes a multi-item clipboard and a Clipboard Manager configured to utilize a machine-learned algorithm (“Sound Recognition Module”) to translate sound data to a categorical key (e.g., a value used to identify a certain component within a data-structure). While a user performs actions directed at copying an item (e.g., keystroking a keyboard combination), the Clipboard Manager can accept audible input from a user (e.g., a voice command). Once …

On-Device Ranking For Displaying Suggestions, Ankur Aggarwal, Bálint Tillman, Sriram Natarajan

Defensive Publications Series

An important capability of mobile device operating systems and/or virtual assistants is to provide contextually appropriate suggestions that enable users to perform quick actions such as obtain navigation guidance, start media playback, listen and respond to messages, etc. Devices such as vehicle infotainment consoles include additional contextual information, e.g., fuel gauge level, driving stage, etc. that can be important when determining suggestions that are provided via such consoles. This disclosure describes on-device ranking of candidate suggestions to deliver the top ranking suggestions via the vehicle infotainment console. With user permission, the ranking can be based on relevant contextual factors, e.g., …