Digital Commons Network^™

Storage Compartment For A Cleaning Cloth In The Motor Vehicle, Daniel Hoppe

Defensive Publications Series

STORAGE COMPARTMENT FOR A CLEANING CLOTH IN THE MOTOR VEHICLE

Intelligent Windshield Cleaning System, Daniel Hoppe

Defensive Publications Series

INTELLIGENT WINDSHIELD CLEANING SYSTEM

Natural Language Video Processing (Machine Learning-Based Identification, Search, Extraction), Nathan Frey, Lloyd Thompson, Jeff Chien, Wade Norris, Brian Mulford

Defensive Publications Series

The systems and methods described herein provide for a natural-language to generative video process using language-classification and knowledge graph identifiers to find frame segments within a codex of videos that depict the scene described. The mechanics necessary for converting speech to meaningful entities that can be matched to non-text images/videos serve as an underlying basis for generative video.

Disambiguating Assistant Tasks Using Supplemental Data, Google Inc.

Defensive Publications Series

A virtual, intelligent, or computational assistant (e.g., also referred to simply as an “assistant”) is described that relies on supplemental data (e.g., contextual information, user information, etc.) to decipher, and determine actions to perform based on, otherwise ambiguous user commands. With explicit permission from a user, the assistant may access a user’s location history, calendar, e-mail, messages, past assistant interactions, contacts, photos, search history, sensor data, and other contextual or user information to interpret user statements. The supplemental data can be stored locally on a device that is executing the assistant or in a cloud computing environment that is accessible …

Reservation Model For Real-Time Bidding Based Advertising System, Renuka Chadha, Bernhard Suter, Reto Strobl, Beverly Wong, Shashank Gupta, Sujata Kosalge, Radu Jurca, Ivan Romanov, Razvan Constantin

Defensive Publications Series

An advertisement management system can enable an advertiser to switch between realtime bidding and reservation delivery modes. A request is received to transition a campaign from a real-time bidding mode to a reservation mode. A reservation engine determines proposed terms for a reservation contract. An acceptance of the proposed terms is received from the advertiser. The campaign is transitioned from the real-time bidding mode to the reservation mode and a reservation contract is established between the advertisement management system and the advertiser, with campaign parameters made immutable for the duration of the contract.

Using Object Detection To Extract Structured Content From Documents, Natalya Fridman Noy, Jeffrey Dudek

Defensive Publications Series

Structured content such as figures, tables, graphs, captions, and other graphical material often capture the essence of a document. Experienced readers often review the graphical material in a document first to quickly grasp the contents of the document. It is thus evident that identifying and extracting the structured content of a document, e.g., graphical components, is important in building a deeper semantic understanding of the document.

Techniques presented herein automatically extract the structured content of documents. Machine-learning techniques, e.g., object detection, computer vision, etc., are used to recognize and extract the structured content. The techniques work well regardless of the …

Competitor Identification Based On Organic Impressions, Yunting Sun, Aiyou Chen, Abinaya Selvam

Defensive Publications Series

This document describes a technique for identifying competitors of a brand based on organic impression data. The system identifies relevant search queries that lead to the brand website based on a selection of search queries from historical data. The system assigns a weight to each search query associated with the brand website. The system calculates an organic impression count based on the search query. The system calculates a score for each candidate competitor as it relates to the brand website. The system ranks/sorts each candidate brand based on its calculated score to determine the top competitors for the brand website. …

Displaying Web Pages Within A Software Keyboard, Dong Kim, John Arendt

Defensive Publications Series

Techniques described in this disclosure display entire web pages with a display area of a software keyboard. In response to user selection of a search result, an optimized version of a web page is retrieved and displayed in the display area. Use of optimized versions of web pages ensures that the software keyboard does not experience crashes or performance problems, e.g., due to inefficient web pages, web pages that include incompatible content, etc. Display of entire web pages in the keyboard display area saves the user the effort to switch to a browser application to view the web page to …

Low-Bandwidth, Client-Based, Rendering For Gaming Videos, Ian Fischer, Shumeet Baluja

Defensive Publications Series

A system for low-bandwidth, client-based, rendering for gaming videos is described. The system may include a gaming device, server device, and user devices. The gaming device may include a processing device and graphics processing unit (GPU). The processing device receives user input and generates rendering commands from the user input. A first rendering unit of the GPU generates gaming video from the rendering commands. The server device receives the gaming video and the rendering commands from the gaming device. The server device determines the first user device is not compatible with the rendering commands, compresses the gaming video, and transmits …

Assistant Generated Answers Using Supplemental Data, Google Inc.

Defensive Publications Series

A virtual, intelligent, or computational assistant (e.g., also referred to simply as an “assistant”) is described that relies on supplemental data (e.g., contextual information, user information, etc.) to decipher, and determine answers to, user queries. With explicit permission from a user, the assistant may access a user’s location history, calendar, e-mail, messages, past assistant interactions, contacts, photos, search history, sensor data, and other contextual or user information to answer a user query. The supplemental data can be stored locally on a device that is executing the assistant or in a cloud computing environment that is accessible to the assistant from …

Applying Machine Learning Techniques To Determine Product Risks

Defensive Publications Series

Generally, the present disclosure is directed to techniques to automatically determine risks associated with a product. In some implementations, the techniques of the present disclosure can include or otherwise leverage one or more machine-learned models to determine if the release and continued sales of product has legal, privacy and/or business vulnerabilities based on identifying sensitive keywords in product-related documentation and product areas.

This disclosure applies machine learning techniques to automate product audits and improve product review quality. Machine learning techniques are applied to proactively identify risks associated with a product, e.g., for a software product or service, the techniques are …

Smart Voicemail With Action Triggers Based On Semantic Understanding, Robin Dua, Sujith Ravi

Defensive Publications Series

Techniques for responding to voicemails based on language processing of voicemails to identify actions that can be performed in response to the voicemails. The actions can be presented as selectable suggestion elements that, when selected, cause certain actions to be performed in reply to the voicemail. For example, a voicemail that includes a question for the recipient can be processed in order to generate selectable suggestion elements for answering the question via text message. Some techniques identify requests (e.g., request for a document, contact information, payment, etc.) from the recipient and, in response, queries can be performed in order to …

A Method For Population Segmentation By Event Based Machine Learning, David Parkinson

Defensive Publications Series

The present disclosure provides systems and methods for automating the entire process of segmentation analytics for users of an application. More particularly, input data is received by a machine learning algorithm. The algorithm inspects the data to determine patterns. Upon discovering patterns of event information the algorithm generates a list of segmentations with applicable characteristics. An application’s users are grouped together, matching user characteristics with those shared by a particular segment. Over time the algorithm learns from an increased amount of data, generating more accurate and comprehensive user segmentations with time and increased use. Keywords associated with the present disclosure …

Smart State Management, Andy Lou, I-Chun Shih, Johnny Hung, Ricky Tang

Defensive Publications Series

A method of smart state management to make a software stateful is disclosed. This method involves determining where to define states within a software using a classification learning mechanism.

Placing Sponsored-Content Associated With An Image, Yushi Jing, Shumeet Baluja

Defensive Publications Series

Techniques are described for placing sponsored-content associated with an image. The techniques may include matching a first image for which a sponsored-content item is to be selected with a reference image. A sponsored-content item to be presented may be selected based on an association between the reference image and the sponsored-content item to be presented.

Actions On Landing Pages, John D. Lanza, John D. Lanza

Defensive Publications Series

Systems and methods described herein allow for using audio overlay to provide actionable audio extensions with primary audio content. A data processing system can receive a first input audio signal from a client device of a user, and identify a user request from the first input audio signal. The data processing system can generate an audio response to the user’s request, and one or more audio extensions for presenting with the audio response. The data processing can then transmit the audio response and the audio extensions to the client device. The data processing system may receive a second audio signal …

Pooling Engine, John D. Lanza, John D. Lanza

Defensive Publications Series

Systems and methods described herein allow for consolidation of requests from multiple users in a voice activated interaction computer environment. A data processing system can receive at least two audio input signals from at least two user devices associated with at least two corresponding users, and determine, for each received input audio signal, a respective user request. The data processing system can generate, for each received input audio signal, a corresponding action data structure, and compare the action data structures associated with separate user requests to determine a pooling parameter indicative of overlapping features or themes among the action data …

Automatic Playlist Creation For Group Events, Matthew Sharifi, Jakob Foerster

Defensive Publications Series

A playlist creation engine for automating the creation of a playlist which intelligently incorporates the interests, tastes, and preferences of every attendee of a group event which has indicated a preference to include their personalized consumption profiles based upon their consumption of media content through the content sharing platform is described. The playlist creation engine identifies a type of event and a group of users attending an event to provide the attending users with a request to build each user’s personalized consumption profile for use in generating a requested playlist. Playlist entries are determined from the personalized consumption profiles and …

Millimeter-Wave Radio System, Mark Felipe, Sean Keys

Defensive Publications Series

A millimeter-wave radio system enables a need-based communication between a set of radios to improve performance metrics, such as signal strength/power performance, that impact the quality of service. The system monitors performance of a radio receiver using a software engine present within the radio receiver. Based on the monitored power performance, the system makes adjustments internally as a first course of action and then provides an update to a radio transmitter. Based on this update, the radio transmitter transmits a message to the radio receiver which acts as a second course of action for the receiver to optimize the power …

Calibration-Free Estimation Of User-Specific Bending Of A Head-Mountable Device, Erfan Davami, Thad Starner

Defensive Publications Series

Disclosed is an approach for determining camera rotation relative to an individual’s eye, which could then be used for applications such as calibration-free estimation of user-specific bending of a device (e.g., a head-mountable device (HMD)), among others. Due to biological conditions, an average gaze vector of an eye most often corresponds to a straight-forward gaze by an individual. Therefore, the average gaze vector is often known with respect to a coordinate system of an individual’s eye. Additionally, when the HMD is worn by a user, the HMD uses an eye-facing camera to determine the average gaze vector of the user’s …

Intelligent Generation Of Physical Products, Luke Mulcahy

Defensive Publications Series

Users search for products via a search engine by submitting search queries. Users are presented with search results based on the search queries. The search results may include offers for products that are relevant to a search query or may include no relevant offers. The search system logs search query data including logging whether offers related to the search query were presented in the search result and whether users submitting the search queries selected any offers associated with the search results. The search system designates search queries in which users selected offers in the search results as converted search queries …

Producing A Sharable Link For Replaying Online User Actions, Reinaldo Aguiar

Defensive Publications Series

A mechanism is provided for a Web browser application that is initiated by a first user to automatically produce a single sharable link that, when opened by a second user, re-plays the actions performed by the first user who shared the link. For example, the second user can use the sharable link to automatically play back the actions on a different computer system. The actions can include, for example, search terms that were used, web sites that were accessed, and other steps that were performed.

Nfv Dynamic And Adaptative Contextual Execution, Ignacio Aldama, Ruben Sevilla Giron, Javier Garcã­A Lopez

Defensive Publications Series

Network function virtualization is an emerging technology that is changing the paradigm of operations within the telco industry. Associated with the virtualization benefits it also appear a growing complexity to manage and operate complex and dynamic scenarios.

In a traditional network all the elements are physical and so they are limited and very rigid. You may have N and only N elements and it is not easy, neither cheap and neither fast, to either increase or decrease that number.

In an NFV world you got one additional layer as the “physical elements” become virtual and run as software on top …

Client-Side Session-Based Contextual User Model Builder, Arthur Asuncion, Kang Li

Defensive Publications Series

A client-side user model is created and maintained for use in selecting content. For example, a user model builder creates and updates a client-side user model. The client-side user model is populated with information from a video-specific user model received from a server and updated at the client side using information from a user activity history, including a video watching history. When requesting a video from the server, the client device can send a user profile derived from the client-side user model. The server can use information from the user profile to personalize content provided to the client device.

Displaying Closed Captioning Using Speech-To-Text, Brant Candelore, Mike Nejat, Peter Shintani

Defensive Publications Series

A system and method for providing synchronized Closed Captioning (CC) for live programming is disclosed herein. The method includes providing CC by speech-to-text conversion of the live content followed by synchronization of the CC with the dialogue to be displayed in a device. The speech-to-text conversion may be run in real-time or non-real time. The system provides annotations for different speakers and background music. The live content may also be stored on a video recorder, such as a set-top box, and processed offline. Similarly, a web stream may be recorded and processed offline. Alternatively, it may also be possible to …

Algorithm To Enrich Display Of A Link In A Web Application, Erwan Gereec

Defensive Publications Series

This document describes a universal algorithm to enhance display of a link tag in web applications. The purpose is to display at least the title of the website, a little description and if exist, some preview of images which are displayed on the website. Thus, the user who receive a link will have relevant information to know what is the content of the website.

Modifying Metadata Based On A Received Gesture, Ricardo Marquez, David Tabb, Amit Chandak, Ella Musina, Melissa Nikolic, Reagan Kim, Jennifer Low

Defensive Publications Series

A metadata modification system can be used to modify metadata associated with one or more products displayed at a display screen. The system presents the metadata associated with one or more products at the display screen of the electronic device. The system receives a gesture from a user at the display screen. The system modifies the metadata associated with one or more products based on the received gesture from the user.

A Novel Method To Discover Datacenter Hardware Instantly Using Software Defined Grid Model, Umesh M L, Naveena Kedlaya

Defensive Publications Series

Hardware (Servers, Storage, Switches) in Research & Development (R&D) datacenters may be used for ongoing research, development, testing, build, qualification and certification, performance, regression testing, resolving customer issues, and the like. Engineers may move the hardware frequently without involving the IT department to meet project requirements and may shutdown the hardware whenever not in use. Currently, there are no solutions available to locate the hardware when it is powered off. It may be useful to detect datacenter hardware instantly, even if the hardware is shutdown, powered off, or exists in an isolated private network. Further, a reporting module capable of …

Crowd Opinion Mining And Scoring, Hassan Abolhassani

Defensive Publications Series

A system and method are disclosed for mining and rating one or more crowd opinions. The system uses a machine learning approach for crowd opinion mining and scoring. The machine learning algorithm creates and updates concepts of the target query in the server by simultaneously mining the web to update opinion scores. A search interface is provided to find concepts and opinions on the targets. Based on the search, the system retrieves the target-related opinions from the server. The system sends crowd-sourced opinions or answers to users. Crowd knowledge is utilized to find opinions and the scores are displayed in …

Cross-Device Private Search, Iyad Assad, Matthias Heiler

Defensive Publications Series

A system is described that facilitates cross-device private search. The cross-device private search may facilitate using a first device to search content on a second remote device without the second remote device uploading the content to a cloud server. A query received on the first device may be forwarded to a third-party server of a service provider. Then the service provider can forward the query to the second remote device, receive search results from the second remote device based on the query, and forward the search results to the first device. Alternatively, in some implementations where the first device and …