Abstract: Methods, systems, and apparatus, including computer programs stored on a computer-readable storage medium, for video stabilization. In some implementations, a computer system obtains frames of a video captured by a recording device using an optical image stabilization (OIS) system. The computing system receives (i) OIS position data indicating positions of the OIS system during capture of the frames, and (ii) device position data indicating positions of the recording device during capture of the frames. The computing system determines a first transformation for a particular frame based on the OIS position data for the particular frame and device position data for the particular frame. The computing system determines a second transformation for the particular frame based on the first transformation and positions of the recording device occurring after capture of the particular frame. The computing system generates a stabilized version of the particular frame using the second transformation.

Abstract: An example method, apparatus, and computer-readable storage medium are provided to predict high-dynamic range (HDR) lighting from low-dynamic range (LDR) background images. In an example implementation, a method may include receiving low-dynamic range (LDR) background images of scenes, each LDR background image captured with appearance of one or more reference objects with different reflectance properties; and training a lighting estimation model based at least on the received LDR background images to predict high-dynamic range (HDR) lighting based at least on the trained model. In another example implementation, a method may include capturing a low-dynamic range (LDR) background image of a scene from an LDR video captured by a camera of the electronic computing device; predicting high-dynamic range (HDR) lighting for the image, the predicting, using a trained model, based at least on the LDR background image; and rendering a virtual object based at least on the predicted HDR lighting.

Abstract: The present disclosure provides systems and methods for authenticated control of content delivery. The method includes receiving a request for an item of content from a computing device, the request comprising a security token associated with the computing device and an identifier of a group of domains, identifying the group of domains from the identifier, and retrieving a security key associated with the group of domains. The method further includes decrypting a signature of the security token, identifying an authentication string, determining that the authentication string matches a server authentication string, and identifying characteristics of the security token. The characteristics of the security token include a confidence score. The method further includes comparing the confidence score of the security token to a threshold, determining that the confidence score does not exceed the threshold, and preventing transmission of content to the computing device.

Abstract: Implementations can identify a given assistant device from among a plurality of assistant devices in an ecosystem, obtain device-specific signal(s) that are generated by the given assistant device, process the device-specific signal(s) to generate candidate semantic label(s) for the given assistant device, select a given semantic label for the given semantic device from among the candidate semantic label(s), and assigning, in a device topology representation of the ecosystem, the given semantic label to the given assistant device. Implementations can optionally receive a spoken utterance that includes a query or command at the assistant device(s), determine a semantic property of the query or command matches the given semantic label to the given assistant device, and cause the given assistant device to satisfy the query or command.

Abstract: Techniques are disclosed that enable determining and/or utilizing a misrecognition of a spoken utterance, where the misrecognition is generated using an automatic speech recognition (ASR) model. Various implementations include determining a recognition based on the spoken utterance and a previous utterance spoken prior to the spoken utterance. Additionally or alternatively, implementations include personalizing an ASR engine for a user based on the spoken utterance and the previous utterance spoken prior to the spoken utterance (e.g., based on audio data capturing the previous utterance and a text representation of the spoken utterance).

Abstract: The various implementations described herein include a video camera assembly that includes: (1) a housing; (2) an image sensor encased in the housing and configured to capture activity of the smart home environment; (3) a wireless radio configured to transmit video frames captured by the image sensor to an electronic device via a remote server; (4) at least one infrared transmitter configured to selectively illuminate the smart home environment; (5) one or more circuit boards encased in the housing, the one or more circuit boards including at least one processor mounted thereon; and (6) a heating component coupled to the image sensor, the heating component configured to continuously maintain the image sensor at a temperature above a threshold temperature while the image sensor is capturing the activity of the smart home environment.

Abstract: Data log-base logical block devices are used to control parameter-based versioning at a block-device level. Memory in the block device stores data log indicating data and corresponding metadata, the metadata indicating a particular historical time. The block device may receive a request, such as a remote procedure call (RPC), for data from the logical block device, and identify data included in the data log having metadata that matches or satisfies a historical time parameter included in and extracted from the request.

Abstract: A method for an in-memory distributed cache includes receiving a write request from a client device to write a block of client data in random access memory (RAM) of a memory host and determining whether to allow the write request by determining whether the client device has permission to write the block of client data at the memory host, determining whether the block of client data is currently saved at the memory host, and determining whether a free block of RAM is available. When the client device has permission to write the block of client data at the memory host, the block of client data is not currently saved at the memory host, and a free block of RAM is available, the write request is allowed and the client is allowed to write the block of client data to the free block of RAM.

Abstract: The present disclosure provides systems and methods for distributed training of machine learning models. In one example, a computer-implemented method is provided for training machine-learned models. The method includes obtaining, by one or more computing devices, a plurality of regions based at least in part on temporal availability of user devices; selecting a plurality of available user devices within a region; and providing a current version of a machine-learned model associated with the region to the plurality of selected user devices within the region. The method includes obtaining, from the plurality of selected user devices, updated machine-learned model data generated by the plurality of selected user devices through training of the current version of the machine-learned model associated with the region using data local to each of the plurality of selected user devices and generating an updated machine-learned model associated with the region based on the updated machine-learned model data.

Abstract: A method includes receiving audio data corresponding to an utterance spoken by a user and processing, using a first recognition model, the audio data to generate a non-contextual candidate hypothesis as output from the first recognition model. The non-contextual candidate hypothesis has a corresponding likelihood score assigned by the first recognition model. The method also includes generating, using a second recognition model configured to receive personal context information, a contextual candidate hypothesis that includes a personal named entity. The method also includes scoring, based on the personal context information and the corresponding likelihood score assigned to the non-contextual candidate hypothesis, the contextual candidate hypothesis relative to the non-contextual candidate hypotheses.

Abstract: A computer-implemented method for k-anonymizing a dataset to provide privacy guarantees for all columns in the dataset can include obtaining, by a computing system including one or more computing devices, a dataset comprising data indicative of a plurality of entities and at least one data item respective to at least one of the plurality of entities. The computer-implemented method can include clustering, by the computing system, the plurality of entities into at least one entity cluster. The computer-implemented method can include determining, by the computing system, a majority condition for the at least one entity cluster, the majority condition indicating that the at least one data item is respective to at least a majority of the plurality of entities. The computer-implemented method can include assigning, by the computing system, the at least one data item to the plurality of entities in an anonymized dataset based at least in part on the majority condition.

Abstract: Among other things, an apparatus comprises quantum units; and couplers among the quantum units. Each coupler is configured to couple a pair of quantum units according to a quantum Hamiltonian characterizing the quantum units and the couplers. The quantum Hamiltonian includes quantum annealer Hamiltonian and a quantum governor Hamiltonian. The quantum annealer Hamiltonian includes information bearing degrees of freedom. The quantum governor Hamiltonian includes non-information bearing degrees of freedom that are engineered to steer the dissipative dynamics of information bearing degrees of freedom.

Abstract: Implementations relate to techniques for providing context-dependent search results. The techniques can include receiving a query and background audio. The techniques can also include identifying the background audio, establishing concepts related to the background audio and obtaining terms related to the concepts related to the background audio. The techniques can also include obtaining search results based on the query and on at least one of the terms. The techniques can also include providing the search results.

Abstract: Systems and methods for training a machine-learned model are provided. A method can include can include obtaining an unlabeled audio signal, sampling the unlabeled audio signal to select one or more sampled slices, inputting the one or more sampled slices into a machine-learned model, receiving, as an output of the machine-learned model, one or more determined characteristics associated with the audio signal, determining a loss function for the machine-learned model based at least in part on a difference between the one or more determined characteristics and one or more corresponding ground truth characteristics of the audio signal, and training the machine-learned model from end to end based at least in part on the loss function. The one or more determined characteristics can include one or more reconstructed portions of the audio signal temporally adjacent to the one or more sampled slices or an estimated distance between two sampled slices.

Abstract: The subject matter described herein provides systems and techniques for adding an identity (ID) header to IP packets associated with a VoIP call. This ID header may be used to authenticate the source provider/originator of a VoIP call, may be used to traceback to the source provider/originator of the VoIP call, and may be used to create a relationship between the source provider/originator and the destination provider/destination of the VoIP call. Such steps may be performed by a public proxy/platform. The ID header may include a certificate and/or a public encryption key, from a public certificate authority (CA) infrastructure, which assists in authenticating the source provider/originator of the VoIP call. The public proxy/platform may directly route authenticated VoIP calls through a VoIP network towards its destination, bypassing a public switched telephone network (PSTN).

Abstract: A head mounted display includes a combiner configured to receive display light from a micro-display. The world-facing surface of the combiner has a curvature that corresponds to a user's vision correction prescription. The head mounted display also includes a corrective layer having a second curvature that corresponds to the user's vision correction prescription. The corrective layer is disposed on the eye-facing surface of the combiner such that the focal point of the display light is adjusted for the specific user as the display light exits the combiner towards the user's eye.

Abstract: A flexible annotation framework normalizes auxiliary information from diverse sources, ranks the information for an individual search result, and provides a lightweight or full display of the auxiliary information in an annotation for the search result. An annotation thus displays information not typically part of the details included in the search result. An example method comprises, for at least one item in a search result page, identifying at least one annotation of a first annotation type in an annotation data store that references the item, identifying at least one annotation for a second annotation type in an annotation data store that references the item, ranking the annotation of the first annotation type and the annotation of the second annotation type and providing the highest ranked annotation as part of a search result for the item in the search result page.

Abstract: The present disclosure provides a system and method for delegating authority to cloud IoT devices, with such delegated authority enabling the cloud IoT devices to access second cloud services outside of a core network. The IoT device uses its IoT identity to obtain a token for accessing the second service within a predefined time window. The token may be used to access the second service without further authentication by the second service. Accordingly, the IoT device can take particular actions, such as downloading files, etc., during the predefined time window. After the predefined time window, the IoT device may no longer access the second service without obtaining another token.

Abstract: A method includes receiving, via a user interface of a client device, a request to populate one or more cells of a plurality of cells of a document having a tabular structure, wherein the one or more cells correspond to a first attribute pertaining to a first column header and a first object pertaining to a first row header; analyzing the request and one or more additional cells corresponding to one or more additional attributes and one or more additional objects of the document to obtain contextual information for the request; generating a query based at least in part on the contextual information; initiating an execution of the query to obtain a response using one or more data sources; and causing the user interface to be modified to populate the response in the one or more cells corresponding to the first attribute and the first object.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media for evaluating robot learning. In some implementations, a system receives classification examples from a plurality of remote devices over a communication network. The classification examples can include (i) a data representation generated by a remote device based on sensor data captured by the remote device and (ii) a classification corresponding to the data representation. The system assigns quality scores to the classification examples based on a level of similarity of the data representations with other data representations. The system selects a subset of the classification examples based on the quality scores assigned to the classification examples. The system trains a machine learning model using the selected subset of the classification examples.