Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for determining neural network architectures. One of the methods includes generating, using a controller neural network having controller parameters and in accordance with current values of the controller parameters, a batch of output sequences. The method includes, for each output sequence in the batch: generating an instance of a child convolutional neural network (CNN) that includes multiple instances of a first convolutional cell having an architecture defined by the output sequence; training the instance of the child CNN to perform an image processing task; and evaluating a performance of the trained instance of the child CNN on the task to determine a performance metric for the trained instance of the child CNN; and using the performance metrics for the trained instances of the child CNN to adjust current values of the controller parameters of the controller neural network.

Abstract: This document describes systems and techniques that may be used to aggregate information about open parking spots from various different parking providers or organizations.

Abstract: A method for accessing customer data includes receiving an access request requesting access to customer data stored on a storage abstraction. The access request includes a justification that specifies a purpose/reason for requesting access to the customer data. The method also includes validating the justification, and after validating the justification, transmitting the justification to an external key management service associated with a customer of the customer data. The external key management service is configured to grant or deny access to the customer data based on the justification. The method also includes receiving an approved access token from the external key management service when the external key management service grants access to the customer data and accessing the customer data stored on the storage abstraction using the approved access token received from the external key management service.

Abstract: A method and workstation for generating a transcript of a conversation between a patient and a healthcare practitioner is disclosed. A workstation is provided with a tool for rendering of an audio recording of the conversation and generating a display of a transcript of the audio recording using a speech-to-text engine, thereby enabling inspection of the accuracy of conversion of speech to text. A tool is provided for scrolling through the transcript and rendering the portion of the audio according to the position of the scrolling. There is a highlighting in the transcript of words or phrases spoken by the patient relating to symptoms, medications or other medically relevant concepts. Additionally, there is provided a set of transcript supplement tools enabling editing of specific portions of the transcript based on the content of the corresponding portion of audio recording.

Abstract: Systems, methods performed by data processing apparatus and computer storage media encoded with computer programs for receiving an utterance from a user in a multi-user environment, each user having an associated set of available resources, determining that the received utterance includes at least one predetermined word, comparing speaker identification features of the uttered predetermined word with speaker identification features of each of a plurality of previous utterances of the predetermined word, the plurality of previous predetermined word utterances corresponding to different known users in the multi-user environment, attempting to identify the user associated with the uttered predetermined word as matching one of the known users in the multi-user environment, and based on a result of the attempt to identify, selectively providing the user with access to one or more resources associated with a corresponding known user.

Abstract: A compact electronic device has a touch sensor and/or a microphone that are concealed within a housing at least partially wrapped by an acoustically porous cover. In some implementations, the touch sensor includes a sensing portion and a contact portion extending from the sensing portion. While the sensing portion is placed in proximity to an interior surface of the housing to detect a touch on the housing, the contact portion is bent to electrically couple the sensing portion to a circuit board via two distinct electrical paths. In some implementations, an exterior surface of the housing includes a sealing area surrounding an aperture on the housing, and the acoustically porous cover is affixed to the sealing area via an adhesive. The adhesive covers the sealing area and permeates a thickness of the acoustically porous cover, thereby enabling formation of a controlled sound path to access the microphone via the aperture.

Abstract: An environmental frequency sensing device, includes logic that performs signal strength (SS) level separation on a received band of frequencies to produce SS level separated frequencies. The logic performs frequency grouping on the SS level separated frequencies for each signal strength level to produce magnitude information for each grouping. The logic generates peak data by detecting peaks of the produced magnitude information. The logic generates an edge event indicating a signal edge based on arrival or departure of a given peak and compares, on a frequency basis, generated edges to stored fingerprint data of a signal of interest. Based on the comparison, the logic provides detected signal data indicating current use of a range of frequencies in an environment.

Abstract: Active noise cancellation systems and methods include a feedforward path configured to receive a reference signal comprising ambient noise and adaptively generate an anti-noise signal to cancel the ambient noise. The adaptive filter is tuned in accordance with at least one parameter, which is set by a logic device configured to determine an ambient noise condition based the reference signal by estimating a fullband power of the reference signal, estimating a low-frequency power of the reference signal, comparing the fullband power and low-frequency power to one or more thresholds, and/or setting one or more ambient noise flags. The ambient noise condition may include a quiet background, a wideband noise condition, and/or a low-frequency dominant noise condition.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for hotword detection on multiple devices are disclosed. In one aspect, a method includes the actions of receiving, by a first computing device, audio data that corresponds to an utterance. The actions further include determining a first value corresponding to a likelihood that the utterance includes a hotword. The actions further include receiving a second value corresponding to a likelihood that the utterance includes the hotword, the second value being determined by a second computing device. The actions further include comparing the first value and the second value. The actions further include based on comparing the first value to the second value, initiating speech recognition processing on the audio data.

Abstract: A method includes displaying, via a graphical user interface (GUI), a first graphical representation of a dataset on a display screen in communication with the data processing hardware, the first graphical representation representing a statistical summary of the dataset. The method includes receiving, via the GUI, an indication of a user interaction. In response to the received indication of the user interaction, the method includes displaying, via the GUI, a second graphical representation of the dataset on the display screen, the second graphical representation representing a plurality of statistical summaries of the dataset, each respective statistical summary of the plurality of statistical summaries computed for a respective time period of a plurality of time periods.

Abstract: A method for accessing one or more service processes of service includes executing at least one service enclave and executing an enclave sandbox that wraps the at least one service enclave. The at least one service enclave provides an interface to the one or more service processes. The enclave sandbox is configured to establish an encrypted communication tunnel to the at least one service enclave interfacing with the one or more service processes, and communicate program calls to/from the one or more service processes as encrypted communications through the encrypted communication tunnel.

Abstract: Systems and methods for cancelling a query are disclosed. A method includes establishing a connection over a network with a user computing device associated with a user, receiving, from the user computing device via the connection, a query request requesting query results from a database and, based on the query request, instructing, via the network, a server to begin executing a database query against the database. The method also includes determining that the connection is disconnected and, in response to determining that the connection is disconnected, instructing, via the network, the server to stop executing the database query.

Abstract: A process for recommending pre-constructed queries in data analytics includes writing different records to a correlation data structure correlating different data classifications of data to different queries and, subsequent to the writing, establishing a communicative connection by a data analytics application to an underlying database. Thereafter, a data model for data in the database may be constructed in the data analytics application and at least one of the different queries may be selected in the correlation data structure that correlates to the classification of the data in the data model. Finally, the selected one of the different queries may be displayed in the data analytics application to an end user so as to provide an intelligent recommendation for the addition of the selected one of the different queries without requiring the end user to alone and without assistance discover the suitability of the selected one of the different queries.

Abstract: A system for collaboratively generating and/or managing a media channel is provided. The system includes an identification component, a generation component and a permissions component. The identification component identifies a first set of media content associated with a first social group and a second set of media content associated with a second social group. The generation component generates a media channel based on the first and second sets of media content identified as being respectively associated with the first and second social groups. The permissions component grants access to the media channel based at least on permissions generated for the first and second sets of media content in connection with the respective first and second social groups.

Abstract: A doorbell chime bypass circuit includes a first node, a second node, and a bi-directional FET switch in series with the first node and the second current node. The bi-directional FET switch includes a first FET and a second FET in series, and is configured to cease conducting current between the first and second nodes when gate voltages of the first and second FETs are below a cut-off threshold. The bypass circuit further includes a sensing circuit configured to determine a level of current flowing through the bi-directional FET switch, and a switch controller configured to set the gate voltages of the first and second FETs to a level below the cut-off threshold when the sensing circuit senses that the level of current meets a doorbell press current threshold, causing the bi-directional FET switch to cease conducting current between the first and second nodes.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-readable storage media, for speech recognition using attention-based sequence-to-sequence models. In some implementations, audio data indicating acoustic characteristics of an utterance is received. A sequence of feature vectors indicative of the acoustic characteristics of the utterance is generated. The sequence of feature vectors is processed using a speech recognition model that has been trained using a loss function that uses N-best lists of decoded hypotheses, the speech recognition model including an encoder, an attention module, and a decoder. The encoder and decoder each include one or more recurrent neural network layers. A sequence of output vectors representing distributions over a predetermined set of linguistic units is obtained. A transcription for the utterance is obtained based on the sequence of output vectors. Data indicating the transcription of the utterance is provided.

Abstract: An indication of a first change is received at a first coordinate location of a client model of the collaborative electronic document. The first coordinate location is within a coordinate structure of the collaborative electronic document. The first change is applied at a second coordinate location of a server model of the collaborative electronic document. The first change is associated with a first revision identifier. An attribution model is generated that includes a coordinate structure that corresponds to at least a portion of the coordinate structure of the collaborative electronic document and includes a third coordinate location that corresponds to the first coordinate location and the second coordinate location. The third coordinate location of the attribution model is updated with the first revision identifier. The attribution information for the first change is stored in association with the first revision identifier.

Abstract: Technology for automatic cross-device meeting authentication. An example method involving initializing, by a first computing device, a real-time meeting communication session, wherein the first computing device is connected to a media system; receiving information indicative of a mobile computing device of a user being in a physical presence of the first computing device, wherein the user is associated with the mobile device and a second computing device; granting control of the real-time meeting communication session from the mobile computing device of the user to the first computing device; transmitting an instruction that causes the second computing device to display a user interface element associated with the real-time meeting communication session, wherein the instruction is sent after granting the control; and granting control of the real-time meeting communication session to the second computing device of the user in response to a selection of the user interface element.

Abstract: A method for estimating an embedding capacity includes receiving, at a deterministic reference encoder, a reference audio signal, and determining a reference embedding corresponding to the reference audio signal, the reference embedding having a corresponding embedding dimensionality. The method also includes measuring a first reconstruction loss as a function of the corresponding embedding dimensionality of the reference embedding and obtaining a variational embedding from a variational posterior. The variational embedding has a corresponding embedding dimensionality and a specified capacity. The method also includes measuring a second reconstruction loss as a function of the corresponding embedding dimensionality of the variational embedding and estimating a capacity of the reference embedding by comparing the first measured reconstruction loss for the reference embedding relative to the second measured reconstruction loss for the variational embedding having the specified capacity.

Abstract: A video playlist associated with a set of videos is distinguished in a graphical user interface using a dynamic thumbnail to represent the playlist. The dynamic thumbnail comprises a static portion comprising a first image associated with the set of videos and a dynamic portion comprising one or more second images associated with the set of videos. An image provided in the dynamic portion is configured to change while the first image remains the same in response to a shift in the graphical user interface that results in a change in position of the thumbnail about the user graphical user interface.