Open Access. Powered by Scholars. Published by Universities.®
Articles 4081 - 4110 of 4606
Full-Text Articles in Entire DC Network
Identifying Impaired State For A Driver, Catherine Cardinal, Cliff Sze
Identifying Impaired State For A Driver, Catherine Cardinal, Cliff Sze
Defensive Publications Series
Generally, the present disclosure is directed to improving driver safety by identifying impaired drivers. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a driver’s impairment state based on driving data.
Determining High-Level Topical Annotations For A Conversation, N/A
Determining High-Level Topical Annotations For A Conversation, N/A
Defensive Publications Series
Generally, the present disclosure is directed to annotating a conversation with high-level topical annotations. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict topical annotations for a conversation based on audio data from the conversation.
Determining Projection Format For A Video, N/A
Determining Projection Format For A Video, N/A
Defensive Publications Series
Generally, the present disclosure is directed to determining one or more projection formats for a video. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a projection format for a segment of a video based on image data extracted from the video.
Recommending Items For Deletion Based On Item Quality And Usage, N/A
Recommending Items For Deletion Based On Item Quality And Usage, N/A
Defensive Publications Series
Generally, the present disclosure is directed to recommending items stored in a data storage system for deletion. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a likelihood of deletion for an item based on item data and user interaction data for one or more items.
Using Customer Support Interaction Data To Estimate Customer Satisfaction, N/A
Using Customer Support Interaction Data To Estimate Customer Satisfaction, N/A
Defensive Publications Series
Generally, the present disclosure is directed to using customer support interaction data to estimate customer satisfaction. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a customer satisfaction score based on customer interaction data.
Determining Priority Value Of Processes Based On Usage History, Neil Dhillon, Tanmay Wadhwa
Determining Priority Value Of Processes Based On Usage History, Neil Dhillon, Tanmay Wadhwa
Defensive Publications Series
Generally, the present disclosure is directed to determining optimal priority values for one or more processes in a computing system. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict an optimal priority value for a process based on system data and/or process data.
Semi-Supervised Classification Using Object Metadata, Dave Feltenberger
Semi-Supervised Classification Using Object Metadata, Dave Feltenberger
Defensive Publications Series
Generally, the present disclosure is directed to classification of data objects (e.g. documents, images, graphs, etc.). In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to classify a data object based on object data and/or metadata associated with the object.
Optimizing Route Guidance To Preserve User Loyalty, Thomas Price
Optimizing Route Guidance To Preserve User Loyalty, Thomas Price
Defensive Publications Series
Generally, the present disclosure is directed to optimizing route guidance in a navigation or ride share service to preserve user loyalty. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict likelihood of user abandonment based on user data and/or route data.
Classifying A User As Driver Or Passenger In A Vehicle Using Machine Learning, Thomas Price
Classifying A User As Driver Or Passenger In A Vehicle Using Machine Learning, Thomas Price
Defensive Publications Series
Generally, the present disclosure is directed to classifying a user as driver or passenger in a vehicle using machine learning. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a user type based on device usage data and/or device sensor data.
Determining Optimal Dimming Of Displays, N/A
Determining Optimal Dimming Of Displays, N/A
Defensive Publications Series
Generally, the present disclosure is directed to dimming a display to reduce power consumption and/or improve security of a device. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predictively dim a display connected to a device based on usage data and/or sensor data available from the device and/or signals derived from the usage data and/or sensor data.
Weighting Knowledge Sources To Facilitate User Input, N/A
Weighting Knowledge Sources To Facilitate User Input, N/A
Defensive Publications Series
Generally, the present disclosure is directed to weighting one or more knowledge sources used to determine user input. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a weight for a knowledge source based on user input data and/or output of one or more knowledge sources.
Learning Weights For Smart Navigation Planning, N/A
Learning Weights For Smart Navigation Planning, N/A
Defensive Publications Series
Generally, the present disclosure is directed to learning costs or weights associated with elements used in navigation planning. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a travel cost or weight associated with an element based on properties of the element and/or time data and/or weather data.
Routing Audio Output To Provide Inadvertently Muted Audio Output To A User, N/A
Routing Audio Output To Provide Inadvertently Muted Audio Output To A User, N/A
Defensive Publications Series
Generally, the present disclosure is directed to routing audio output to increase the likelihood that a user hears the audio output. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict an output channel for audio output based on audio output data such as connection data, audio data, and/or user interaction data.
Context-Dependent Account Selection, N/A
Context-Dependent Account Selection, N/A
Defensive Publications Series
Generally, the present disclosure is directed to selecting a user account out of one or more user accounts for a user. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a user account for a user based on context data.
Electronic Document Navigation Assistance Using Markings And/Or Non-Uniform Scrolling, N/A
Electronic Document Navigation Assistance Using Markings And/Or Non-Uniform Scrolling, N/A
Defensive Publications Series
Generally, the present disclosure is directed to assisting in navigation within an electronic document. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a location within an electronic document to be marked based on user interaction with the electronic document.
Retrospective User Input Inference And Correction, N/A
Retrospective User Input Inference And Correction, N/A
Defensive Publications Series
Generally, the present disclosure is directed to retrospective user input inference and/or correction. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict intended user input based on actual user input.
Machine Learning To Dissipate A Cyclone, N/A
Machine Learning To Dissipate A Cyclone, N/A
Defensive Publications Series
Generally, the present disclosure is directed to using machine learning to dissipate or otherwise combat a cyclone. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict characteristics of a cyclone (or infant cyclone which can be referred to as a cyclet) such as location, path, strength, number of anti-cyclone devices needed to combat the cyclone, positioning of the anti-cyclone devices, etc. based on information descriptive of the cyclone including imagery of the cyclone or other collected data such as, for example, wind speed and …
Configuring An Application Based On Settings In Similar Applications, N/A
Configuring An Application Based On Settings In Similar Applications, N/A
Defensive Publications Series
Generally, the present disclosure is directed to configuring settings in an application, such as an application on a device. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict a value for a setting in an application based on values of one or more settings in one or more similar applications.
Generating A User-Specific Reduced Content Webpage, N/A
Generating A User-Specific Reduced Content Webpage, N/A
Defensive Publications Series
Generally, the present disclosure is directed to generating a user-specific reduced content webpage that is tailored for a user. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict content on a webpage that is relevant to and/or irrelevant to a user based on the content of the webpage.
Identifying And Determining Trustworthiness Of A Machine-Learned Model, N/A
Identifying And Determining Trustworthiness Of A Machine-Learned Model, N/A
Defensive Publications Series
Generally, the present disclosure is directed to identifying and determining trustworthiness of a production machine-learned model. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to identify a production machine-learned model output within a production system and/or trustworthiness of the production machine-learned model based on the context in which the production machine-learned model is provided.
Predicting User Intent In A Recently Changed Interface, N/A
Predicting User Intent In A Recently Changed Interface, N/A
Defensive Publications Series
Generally, the present disclosure is directed to predicting an intent of an action performed by a user in a recently changed interface. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict the intent of a user in performing an action in a previous state of an interface based on user interaction data and/or interface data.
Context-Aware Switching Of Input Systems, N/A
Context-Aware Switching Of Input Systems, N/A
Defensive Publications Series
Generally, the present disclosure is directed to switching between one or more input systems based on a context of a user. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to predict one or more input systems to provide to a user based on context data relating to a context in which the one or more input systems are used.
Changing Watch Style And Functionality To Match Fashion Products, Matthew Tait
Changing Watch Style And Functionality To Match Fashion Products, Matthew Tait
Defensive Publications Series
The disclosed techniques enable automatic configuration of a wearable computing device, e.g., a smartwatch, to match style and/or functionality to fashion product or accessories. With user permission, the smartwatch obtains information about the fashion product via short-range communication technology, e.g., by reading a near field communication (NFC) tag in fashion products, e.g., jackets, handbags, shoes, etc. Information from the tag is used to identify characteristics of the fashion product such as function, brand, color, style, etc. The smartwatch is automatically configured to update style and/or function to match the identified characteristics.
Location Of Writing Instruments Using Position Sensing Diodes In Stereo Configuration, N/A
Location Of Writing Instruments Using Position Sensing Diodes In Stereo Configuration, N/A
Defensive Publications Series
The present disclosure relates to a position determination system and related method for accurately and rapidly capturing the position of an input device (e.g., a stylus). The position determination system can include an input device such as a writing instrument that includes a light emitting diode (LED) that can be detected by two or more position sensing diodes. The position sensing diodes (PSDs) can generate a bias voltage proportional to the position of a point (e.g., a brightest point) of light incident to the surface of the PSDs. As a result, based on the position of the LEDs relative to …
Automatic Context-Based Alerts For Content Sharing, Matthew Sharifi, Jakob Foerster
Automatic Context-Based Alerts For Content Sharing, Matthew Sharifi, Jakob Foerster
Defensive Publications Series
This disclosure describes techniques for automatic detection of unintentional sharing of content. A user attempt to share content, e.g., from a clipboard via a paste operation, is detected. With user permission, the clipboard contents and the context in which the content is being shared are analyzed to determine suitability of the content to the context. An alert is provided to the user upon detection that sharing the content is likely unintentional.
Ranking And Automatic Selection Of Machine Learning Models Abstract, Sandro Feuz, Victor Carbune
Ranking And Automatic Selection Of Machine Learning Models Abstract, Sandro Feuz, Victor Carbune
Defensive Publications Series
Generally, the present disclosure is directed to an API for ranking and automatic selection from competing machine learning models that can perform a particular task. In particular, in some implementations, the systems and methods of the present disclosure can include or otherwise leverage one or more machine-learned models to provide to a software application one or more machine learning models from different providers. The trained models are suited to a task or data type specified by the developer. The one or more models are selected from a registry of machine learning models, their task specialties, cost, and performance, such that …
Visual Search Using Automatically Annotated Images, Alexander Haubold
Visual Search Using Automatically Annotated Images, Alexander Haubold
Defensive Publications Series
Users interested in obtaining content related to certain concepts sometimes struggle with identifying suitable textual search terms. Users may be unable to formulate a query due to limitations of vocabulary, lack of language skills, lack of knowledge of the specific topic, etc. This disclosure describes techniques to enable a user to conduct a visual search and exploration using images with textual annotations that correspond to objects and concepts in the images. The textual annotations are analogous to hyperlinks, and are used by a search engine to obtain and display search results to the user. The search results can be text …
Automatic Generation Of Affiliate Links, Abhijit Chandgadkar, Eugenio Marchiori
Automatic Generation Of Affiliate Links, Abhijit Chandgadkar, Eugenio Marchiori
Defensive Publications Series
Online publishers that create new content are dissatisfied with currently available advertising formats, e.g., due to a belief that some current advertising formats spoil customer experience or drive customers away from their websites. This disclosure uses a single script within an online publisher website to automatically insert affiliate links within the publisher’s content. Upon execution, these script sends a request an online advertising provider. Page content is analyzed using machine learning techniques to identify entities that refer to products. The identified products are matched with products available for sale from an online vendor and corresponding links are provided to the …
Selective Provision Of Notifications To Devices That Are Offline, Victor Cărbune, Sandro Feuz
Selective Provision Of Notifications To Devices That Are Offline, Victor Cărbune, Sandro Feuz
Defensive Publications Series
Client mobile devices that are offline are unable to receive notifications via the internet. This disclosure describes techniques to selectively transmit important notifications via alternate channels such as phone calls, SMS, etc. to client mobile devices that are offline. A machine learning model is trained to predict importance scores for incoming notifications, based on training data that includes human-provided ground truth scores. The trained machine learning model ranks the importance of various notifications. The techniques ensure that important notifications are delivered even when a client mobile device is disconnected from the internet.
Volumetric Rotational Control, Michael Alger
Volumetric Rotational Control, Michael Alger
Defensive Publications Series
Techniques are described that utilize a volumetric rotational control that can be activated and controlled in an intuitive manner using natural movements. To use the volumetric rotational control, the user operates a virtual reality (VR) controller to virtually touch a VR object that represents a selectable control that includes a collider. A collider is a physics object that can be added to a virtual object to give the virtual object “physical” properties and cause the virtual object to be subject to the “physics” of the VR environment. The user activates the volumetric rotational control by causing the VR controller to …