Abstract: Encoding and decoding is accomplished herein using a multi-reference picture buffer. Decoding includes receiving an encoded video stream, determining an expected error value representing possible frame errors, decoding a plurality of reference frames from the video stream, adding each of the plurality of reference frames to the reference picture buffers up to a maximum number of reference picture buffers, and determining an encoded video stream error value representing actual frame errors based on the decoded plurality of reference frames. If the encoded video stream error value is greater than the decoded expected error value, an error is signaled.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating a dynamic mesh for rendering with a graphical user interface. Graphical objects are rendered onto a layer having transparent pixels. The system infers what areas of the layer are drawn areas, and a mesh is generated based on the drawn areas.

Abstract: Methods, systems, and apparatus, for automatically generating search suggestions based on history data for multiple native application on a user device.

Abstract: This document describes techniques and apparatuses for implementing monochrome-color mapping using a monochromatic imager and a color map sensor. These techniques and apparatuses enable better resolution, depth of color, or low-light sensitivity than many conventional sensor arrays.

Abstract: Implementations relate to computer-assisted agendas for videoconferences. In some implementations, a computer-implemented method includes receiving, at a device, scheduling input associated with a videoconference, where the scheduling input includes a time for the videoconference and one or more identifiers of one or more participants for the videoconference. In response to receiving the scheduling input, the method generates a record of the videoconference in storage, and determines an agenda for the videoconference including a plurality of agenda items. Each agenda item is associated with a respective time period in the videoconference, and at least one of the time periods is automatically determined. The method detects that the videoconference is in progress, and in response, causes output during the videoconference on one or more participating user devices connected to the videoconference, where the output is based on at least one of the agenda items.

Abstract: A system includes an optical transceiver configured to transmit/receive at least one optical feed and a beam separator configured to separate the optical feed into a plurality of optical beams, and spatially combine the optical beams into the optical beam. The system also includes a dichroic mirror optically coupled to the beam separator and configured to reflect the optical beams, and allow beacon signals to pass therethrough. A position sensitive detector of the system optically couples to the dichroic mirror and is configured to sense an incidence position of each beacon signal allowed to pass through the dichroic mirror, and output a position error for each optical beam based on the sensed incidence positions. The system also includes a multi-axis optical steering system configured to direct each optical beam based on the corresponding position error outputted from the position sensitive detector and a corresponding transmit/receive target.

Abstract: A computer-executed technique for correcting previously received textual messages based on a later received textual message can include receiving and displaying a first textual message in a user interface. A second textual message is received and, based on this second textual message, the first textual message is analyzed to determine if it includes a correctable error. When the first textual message is determined to include an error, a language model is utilized to determine an incorrect word therein. A corrected word is programmatically determined and is used to automatically replace the incorrect word to generate a corrected first textual message. The user interface is modified such that the corrected first received textual message replaces the first received textual message.

Abstract: A computer-implemented method includes generating a metric for a first user that reflects preferences for image attributes, determining image attributes for a first set of images associated with a second user, selecting a subset of the first set of images for the first user based on the metric and the image attributes for the first set of images, and providing the subset of the first set of images to the first user.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for resolving location criteria using user location data. One of the methods includes receiving data identifying an action to be performed in response to received user requests and data identifying location criteria for performing the action; determining a cover sets for possible confidence radii; receiving a plurality of requests, each request being associated with a respective location and a confidence radius for the location; determining, for each of the plurality of requests, whether the location associated with the request is in the cover set corresponding to the confidence radius for the location associated with the request; and for each request for which the location associated with the request is in the cover set for the confidence radius corresponding to the location associated with the request, determining that the location criteria for performing the action are satisfied.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for neural network adaptive beamforming for multichannel speech recognition are disclosed. In one aspect, a method includes the actions of receiving a first channel of audio data corresponding to an utterance and a second channel of audio data corresponding to the utterance. The actions further include generating a first set of filter parameters for a first filter based on the first channel of audio data and the second channel of audio data and a second set of filter parameters for a second filter based on the first channel of audio data and the second channel of audio data. The actions further include generating a single combined channel of audio data. The actions further include inputting the audio data to a neural network. The actions further include providing a transcription for the utterance.

Abstract: Implementations relate to prioritized data transmission features over networks. A method can include deriving description data from content data. The description data describes content of the content data and includes one or more portions corresponding to one or more portions of the content data. The description data is transmitted to a destination over a network to be made accessible over the network at the destination to one or more users. The content data also is transmitted to the destination to be made accessible. At least one portion of the description data finishes transmitting and is made accessible before its corresponding portion of the content information finishes transmitting and is made accessible.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for aggregating requests for resources. In one aspect, a method includes receiving a set of requests for computing resources. Each request specifies one or more computing resources and, for each of the one or more computing resources, a quantity of the computing resource requested. A quantity-based provider condition that a request must meet for inclusion in an allocation process is identified. Each received request that fails to meet the quantity-based provider condition is included in a proper subset. Requests in the proper subset of requests are grouped into one or more groups based at least on the quantity of the one or more resources specified by each request of the proper subset and the quantity-based provider condition. A respective combined request is generated for each group of requests.

Abstract: Systems and methods of providing content for display on a computing device via a computer network using a location feature index are provided. A data processing system can receive a request for content from the computing device, and can determine a geographic location of the computing device associated with the request for content. The data processing system can identify a keyword indicating a non-geographic semantic feature of the determined geographic location. The identification can be based on the determined geographic location and from a location feature index that maps geographic areas to keywords that indicate non-geographic semantic features of the geographic areas. The data processing system can select, based on the keyword, a candidate content item for display on the computing device.

Abstract: A method for pull frame interpolation includes receiving an encoded bitstream including information representing a plurality of frames of video data, decoding the plurality of frames, including identifying a plurality of motion vectors indicating motion from a first frame of the plurality of video frames to a second frame of the plurality of video frames, identifying an interpolation point between the first frame and the second frame, identifying a plurality of candidate interpolation motion vectors indicating motion from the first frame to the interpolation point and from the second frame to the interpolation point based on the plurality of motion vectors, selecting an interpolation motion vector from the plurality of candidate interpolation motion vectors based on a metric, and generating an interpolated frame at the interpolation point based on the selected interpolation motion vector, which may include correcting an artifact in the interpolated frame by blending the interpolated frame.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for speech recognition. One of the methods includes receiving first audio data corresponding to an utterance; obtaining a first transcription of the first audio data; receiving data indicating (i) a selection of one or more terms of the first transcription and (ii) one or more of replacement terms; determining that one or more of the replacement terms are classified as a correction of one or more of the selected terms; in response to determining that the one or more of the replacement terms are classified as a correction of the one or more of the selected terms, obtaining a first portion of the first audio data that corresponds to one or more terms of the first transcription; and using the first portion of the first audio data that is associated with the one or more terms of the first transcription to train an acoustic model for recognizing the one or more of the replacement terms.

Abstract: Methods, systems, and apparatus include computer programs encoded on a computer-readable storage medium, including a method for providing content. An example content presentation opportunity is identified as being associated with a reservation system. The content presentation opportunity is characterized, including creating a signature defined by signals. Creating the signature includes defining a value for each signature element, wherein an element corresponds to one or more of signals. The value for a respective element is based on a value for a respective signal as defined by the content presentation opportunity. At least a portion of the signature is provided for presentation to a user. The user is queried for an indication of a relative importance of signature elements. The signature is evaluated in context of an exchange system including creating a proposed campaign including selection criteria based on the signature and the received indication. Information related to the campaign is provided.

Abstract: In embodiments of distributed channel sampling across a mesh network, a commissioning device propagates a scanning request, which includes a number of scanning parameters, to nodes in a mesh network, causing the nodes to perform energy detection (ED) scans using the scanning parameters. The commissioning device receives energy measurements in scanning reports from the nodes and analyzes the measurements to determine an operating channel for the mesh network. The commissioning device updates the operating channel in network configuration information that is sent to a leader device in the mesh network, for propagation to the mesh network.

Abstract: In some implementations, a process for communicating data over audio is performed. In one aspect, one or more ordered sequences of audio attribute values that are selected based on a musical relationship between the audio attribute values and associated with data values may be played by a first device and received by a second device. This technique may allow for sound-based communications to take place between devices that listeners may find pleasant.