Abstract: Systems and methods are provided for to estimate the pose of a human subject's head from a sequence of images received from a single depth camera by processing the images to generate a continuous estimate of the head pose in a 3-dimensional (3D) space, and to generate a 3D head model for display and further use. The subject is provided instructions to rotate their head in a first direction until a threshold angle of rotation is reached and then are provided instructions to rotate their head in a second direction. The depth camera provides a sequence of captured images which are processed to extract head meshes. After capture is complete the head meshes are merged to generate a 3D model of the subject's head.

Abstract: Systems and methods are disclosed for anticipating a video switch to accommodate a new speaker in a video conference comprising a real time video stream captured by a camera local to a first videoconference endpoint is analyzed according to at least one speaker anticipation model. The speaker anticipation model predicts that a new speaker is about to speak. Video of the anticipated new speaker is sent to the conferencing server in response to a request for the video on the anticipated new speaker from the conferencing server. Video of the anticipated new speaker is distributed to at least a second videoconference endpoint.

Abstract: This disclosure relates to solutions for eliminating undesired audio artifacts, such as background noises, on an audio channel. A process for implementing the technology can include receiving a set of audio segments, analyzing the segments using a first ML model to identify a first probability of unwanted background noises in the segments, and if the first probability exceeds a threshold, analyzing the segments using a second ML model to determine a second probability that the one or more background features exist in the segments. In some aspects, the process can include attenuating audio artifacts in the segments, if the second probability exceeds a second threshold. In some implementations, dynamic time stretching and shrinking can be applied to the noise attenuation. Systems and machine-readable media are also provided.

Abstract: A turbine component including a shrouded airfoil with a flow conditioner configured to direct leakage flow and coolant to be aligned with main hot gas flow is provided. The flow conditioner is positioned on a shroud base radially adjacent to the tip of the airfoil and includes a ramped radially outer surface positioned further radially inward than a radially outer surface of the shroud base. The ramped radially outer surface extends from a first edge to a second edge in a direction generally from the suction side to the pressure side of the airfoil, such that the first edge is positioned further radially inward than the second edge. Multiple coolant ejection holes are positioned on the ramped radially outer surface. The coolant ejection holes are connected fluidically to an interior of the airfoil.

Abstract: Systems and methods are disclosed for anticipating a video switch to accommodate a new speaker in a video conference comprising a real time video stream captured by a camera local to a first videoconference endpoint is analyzed according to at least one speaker anticipation model. The speaker anticipation model predicts that a new speaker is about to speak. Video of the anticipated new speaker is sent to the conferencing server in response to a request for the video on the anticipated new speaker from the conferencing server. Video of the anticipated new speaker is distributed to at least a second videoconference endpoint.

Abstract: Systems, methods, and devices are disclosed for training a model. Media data is separated into one or more clusters, each cluster based on a feature from a first model. The media data of each cluster is sampled and, based on an analysis of the sampled media data, an accuracy of the media data of each cluster is determined. The accuracy is associated with the feature from the first model. Based on a subset dataset of the media data being outside a threshold accuracy, the subset dataset is automatically forwarded to a crowd source service. Verification of the subset dataset is received from the crowd source service, and the verified subset dataset is added to the first model.

Abstract: This disclosure relates to solutions for eliminating undesired audio artifacts, such as background noises, on an audio channel. A process for implementing the technology can include receiving a set of audio segments, analyzing the segments using a first ML model to identify a first probability of unwanted background noises in the segments, and if the first probability exceeds a threshold, analyzing the segments using a second ML model to determine a second probability that the one or more background features exist in the segments. In some aspects, the process can include attenuating audio artifacts in the segments, if the second probability exceeds a second threshold. In some implementations, dynamic time stretching and shrinking can be applied to the noise attenuation. Systems and machine-readable media are also provided.

Abstract: Systems, methods, and computer-readable media are disclosed for detecting and rendering endnotes in digital content. A user selection of a link rendered at a first location in digital content may be detected and various filtering criteria may be evaluated to determine whether the link references a second location in the digital content that is associated with endnote content. If the link is determined to reference endnote content, the endnote content may be extracted and rendered in association with a portion of the digital content corresponding to the first location.

Abstract: A system includes a user interface presented to a developer. The developer selects a first function to supplement functionality of a first application with external functionality available from third party applications. A code generation module provides a software object to the developer for incorporation into a first state of the first application. The first state includes a user interface element associated with an entity. User selection of the user interface element initiates preparation of a query wrapper including a combination of the entity's name and a predefined text string corresponding to the first function. The query wrapper is transmitted to a search system and a result set is received and displayed. A first item of the result set includes an access mechanism for a specified state of a target application. User selection of the first item causes the access mechanism to open the target application to the specified state.

Abstract: Systems, methods, and devices are disclosed for detecting an active speaker in a two-way conference. Real time audio in one or more sub band domains are analyzed according to an echo cancellor model. Based on the analyzed real time audio, one or more audio metrics are determined from output from an acoustic echo cancellation linear filter. The one or more audio metrics are weighted based on a priority, and a speaker status is determined based on the weighted one or more audio metrics being analyzed according to an active speaker detection model. For an active speaker status, one or more residual echo or noise is removed from the real time audio based on the one or more audio metrics.

Abstract: Disclosed is a system, method and computer readable medium enabling collaboration service providers to more accurately predict packet loss, jitter and delay based on current session, historical session and user location parameters. The prediction can be used to forecast the occurrence of poor media quality at the current location and potential future locations.

Abstract: Machine-learned models are selectively distributed to a plurality of computer servers according to conditions associated with the computer servers. A server receives travel information from a travel coordination system. The travel information describes a plurality of conditions. The server identifies a hierarchy of one or more parent-child relationships based on the plurality of conditions. The server trains machine-learned models using the plurality of conditions described by the travel information. The server selects machine-learned models for the plurality of conditions responsive to the identified hierarchy. The server distributes machine-learned models to the plurality of computer servers responsive to the identified hierarchy.

Abstract: Disclosed is a system, method and computer readable medium enabling collaboration service providers to more accurately predict packet loss, jitter and delay based on current session, historical session and user location parameters. The prediction can be used to forecast the occurrence of poor media quality at the current location and potential future locations.

Abstract: A process for producing a shoe sole, having a hybrid material of a polyurethane foam as a matrix material and an inlay component of expanded particles of a thermoplastic polyurethane is provided. The process includes preparing an inlay form by joining the expanded particles of a thermoplastic polyurethane in an amount and size of the desired inlay; preparing the shoe sole by placing the prepared inlay form in a shoe sole mold such that an edge of the inlay component is 0.2 cm or more from an edge of the shoe sole mold; embedding the inlay form within a reaction mixture in the shoe sole mold; and reacting the reaction mixture to form the matrix in the shoe sole mold.

Abstract: Systems and methods are disclosed for anticipating a video switch to accommodate a new speaker in a video conference comprising a real time video stream captured by a camera local to a first videoconference endpoint is analyzed according to at least one speaker anticipation model. The speaker anticipation model predicts that a new speaker is about to speak. Video of the anticipated new speaker is sent to the conferencing server in response to a request for the video on the anticipated new speaker from the conferencing server. Video of the anticipated new speaker is distributed to at least a second videoconference endpoint.

Abstract: Systems and methods are provided for estimating head pose data from a sequence of depth images of a human subject, and processing the data to generate a continuous estimate of the head pose in a 3-dimensional (3D) space, and to generate a 3D head model for display and further use. The subject is provided instructions to rotate their head in a first direction until a threshold angle of rotation is reached and then are provided instructions to rotate their head in a second direction. The depth camera provides a sequence of captured images which are processed to extract head meshes. After capture is complete the head meshes are merged to generate a 3D model of the subject's head.

Abstract: A method includes receiving a request from a user device executing a native application. The request indicates an accessed state of the native application to which the native application is set. The method includes identifying one or more categories of the accessed state of the native application and recommended application resource identifiers based on features of the accessed state. Each of the recommended application resource identifiers references a different state of a native application. When a native application is selected by the user, the method includes setting the native application to the different state. For each of the application resource identifiers, the method includes identifying one or more recommended categories of the state referenced by the recommended application resource identifier and determining whether to recommend the different state referenced by the recommended application resource identifier based on features of the accessed state and features of the state of the application.