Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium. In one aspect, a method includes receiving a text prompt describing a scene; processing the text prompt using a text encoder neural network to generate a contextual embedding of the text prompt; and processing the contextual embedding using a sequence of generative neural networks to generate a final video depicting the scene.

Abstract: A method for conserving resources in a distributed system includes receiving an event-criteria list from a resource controller. The event-criteria list includes one or more events watched by the resource controller and the resource controller controls at least one target resource and is configured to respond to events from the event-criteria list that occur. The method also includes determining whether the resource controller is idle. When the resource controller is idle, the method includes terminating the resource controller, determining whether any event from the event-criteria list occurs after terminating the resource controller, and, when at least one event from the event-criteria list occurs after terminating the resource controller, recreating the resource controller.

Abstract: Implementations relate to generating multi-modal response(s) through utilization of large language model(s) (LLM(s)). Processor(s) of a system can: receive natural language (NL) based input, generate a multi-modal response that is responsive to the NL based output, and cause the multi-modal response to be rendered. In some implementations, and in generating the multi-modal response, the processor(s) can process, using a LLM, LLM input (e.g., that includes at least the NL based input) to generate LLM output, and determine, based on the LLM output, textual content for inclusion in the multi-modal response and multimedia content for inclusion in the multi-modal response. In some implementations, the multimedia content can be obtained based on a multimedia content tag that is included in the LLM output and that is indicative of the multimedia content. In various implementations, the multimedia content can be interleaved between segments of the textual content.

Abstract: A method, includes: (i) receiving information about an ambient light level; (ii) receiving image frame data for an active matrix display panel with an array of pixels each having a light emitting diode (LED) and a pixel circuit to control current supplied to the LED; (iii) selecting a selected current-resistance compensation (IRC) setting based on the information about the ambient light value; (iv) selecting a selected source voltage level based on the selected IRC setting that was selected by the computing system; (v) generating compensated image frame data for the image frame based on the received image frame data and the selected IRC setting; and (vi) displaying the image frame by supplying data signals based on the compensated image frame data to corresponding pixels from the array of pixels, while applying a source voltage corresponding to the selected source voltage level to all of the pixels.

Abstract: The present disclosure is directed to positioning animated images within a dynamic keyboard interface. In particular, the methods and systems of the present disclosure can: receive, from a user device on which an application is executed, data indicating a context of: the application, and/or a dynamic keyboard interface provided in association with the application; identify, based at least in part on the data indicating the context, a plurality of different animated images, including an animated image comprising an advertisement, for presentation by the dynamic keyboard interface; communicate, to the user device, data indicating the plurality of different animated images; receive, from the user device, data indicating a selection of the animated image comprising the advertisement; and determine, based at least in part on the data indicating the selection and the data indicating the context, a position within the dynamic keyboard interface for presenting the animated image comprising the advertisement.

Abstract: A method for training a non-autoregressive TTS model includes receiving training data that includes a reference audio signal and a corresponding input text sequence. The method also includes encoding the reference audio signal into a variational embedding that disentangles the style/prosody information from the reference audio signal and encoding the input text sequence into an encoded text sequence. The method also includes predicting a phoneme duration for each phoneme in the input text sequence and determining a phoneme duration loss based on the predicted phoneme durations and a reference phoneme duration. The method also includes generating one or more predicted mel-frequency spectrogram sequences for the input text sequence and determining a final spectrogram loss based on the predicted mel-frequency spectrogram sequences and a reference mel-frequency spectrogram sequence. The method also includes training the TTS model based on the final spectrogram loss and the corresponding phoneme duration loss.

Abstract: A method (400) for a website authentication is provided. The method includes receiving a reference uniform resource locator (URL) (130) at a browser (120) and determining whether the reference URL adheres to an authentication protocol (202) including a credential identifier (132c) and a web authentication host URL (132a). When the reference URL adheres to the authentication protocol, the method also includes extracting the credential identifier and the web authentication host URL from the reference URL. The method further includes obtaining credential data (160) from a remote service (150) using the credential identifier and sending web authentication instructions (170) to the browser. The web authentication instructions include the web authentication host URL and the credential data.

Abstract: Techniques for implementing an uplink control channel in Non-Orthogonal Multiple Access (NOMA) asynchronous transmissions are disclosed. The techniques described herein provide an additional control channel in NOMA transmissions for a user equipment (UE) (102) to indicate certain information to a base station (106) to enable the base station (106) to detect and decode the data transmitted by the UE (102). For example, uplink control information (UCI) is transmitted (406) via an uplink control channel in connection with a group user equipment identifier (group UE-ID) to enable the base station (106) to detect and decode the UCI. After decoding (506) the UCI based on the group UE-ID, the base station (106) obtains information usable to distinguish (508) the UE (102), such as a UE-specific ID, as well as to detect (510) and decode (512) a data transmission from the UE (102) on an uplink data channel.

Abstract: A method for proactive notifications in a voice interface device includes: receiving a first user voice request for an action with an future performance time; assigning the first user voice request to a voice assistant service for performance; subsequent to the receiving, receiving a second user voice request and in response to the second user voice request initiating a conversation with the user; and during the conversation: receiving a notification from the voice assistant service of performance of the action; triggering a first audible announcement to the user to indicate a transition from the conversation and interrupting the conversation; triggering a second audible announcement to the user to indicate performance of the action; and triggering a third audible announcement to the user to indicate a transition back to the conversation and rejoining the conversation.

Abstract: Implementations relate to generating standardized metrics from device specific metrics that are generated during an interaction between a user and an automated assistant. The metrics indicate events that occurred while processing an interaction of a user with the automated assistant and are specific to the particular configuration of the device with which the user is interacting. Conversion mappings are determined based on device characteristics that can be utilized to convert the device metrics into standardized metrics. Analysis metrics are generated based on the standardized metrics that are incapable of being generated from the device metrics. Some implementations include visually rendering the analysis metrics such that one or more of the analysis metrics are rendered more prominently than other metrics.

Abstract: Methods, systems, and apparatus for a matrix multiply unit implemented as a systolic array of cells are disclosed. Each cell of the matrix multiply includes: a weight matrix register configured to receive a weight input from either a transposed or a non-transposed weight shift register; a transposed weight shift register configured to receive a weight input from a horizontal direction to be stored in the weight matrix register; a non-transposed weight shift register configured to receive a weight input from a vertical direction to be stored in the weight matrix register; and a multiply unit that is coupled to the weight matrix register and configured to multiply the weight input of the weight matrix register with a vector data input in order to obtain a multiplication result.

Abstract: An optical links diagnostic system (LDS) and its operation within an optical circuit switch (OCS) for measurement and diagnosis of fiber-optic network fiber performance and quality is disclosed. The LDS can contain two photodetectors, a laser source, and be coupled to an OCS. Optical circulators can further be linked to the OCS. The LDS can be used both as an optical time domain reflectometer (OTDR) or as an optical return loss (ORL) meter and can automate the diagnosis of the fiber optical network fiber insertion loss and return loss.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for providing suggestions within a document. In one aspect, a method includes obtaining textual input provided to a document editing application by a user device, the textual input being provided to the document editing application for inclusion in a document; identifying performance measures associated with the current editing session for the document, each performance measure being based on session data obtained from the user device during a document editing session, the session data being for the textual input and prior text that was included in the document prior to the textual input; providing the performance measures as input to a suggestion model that was trained using historical performance measures identified in performance logs for historical document editing sessions of users; and throttling textual suggestions during the current editing session based on the output of the suggestion model.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for brightness control for under-display fingerprint sensor are disclosed. A method includes receiving, at a computing device, an indication to activate an under-display sensor that is located underneath a display of the computing device; activating a collection of LEDs of the display to provide illumination for the under-display sensor, including activating an LED in the collection of LEDs by: establishing, by drive circuitry of the LED, a first overdriven voltage across an LED-driving transistor that is arranged to energize the LED; establishing a second overdriven voltage across the LED-driving transistor; and establishing a steady state voltage across the LED-driving transistor; and activating the under-display sensor by reading a signal from the under-display sensor once the collection of LEDs has activated, including once the steady state voltage has been programmed across the LED-driving transistor.

Abstract: Systems and methods for publishing content are provided. One method includes generating a content item template for a content item configured for publication within one or more resources. The method further includes storing the content item template within a memory for later publication. The method further includes receiving input data from the content provider and detecting a trigger signal associated with the content item. The method further includes entering the content item into an auction in response to detecting the trigger signal. The method further includes adding the input data provided by the content provider via the interface to the content item template to generate the content item prior to publication of the content item within the one or more resources.

Abstract: Example aspects of the present disclosure are directed to systems and methods that learn a compressed representation of a machine-learned model (e.g., neural network) via representation of the model parameters within a reparameterization space during training of the model. In particular, the present disclosure describes an end-to-end model weight compression approach that employs a latent-variable data compression method. The model parameters (e.g., weights and biases) are represented in a “latent” or “reparameterization” space, amounting to a reparameterization. In some implementations, this space can be equipped with a learned probability model, which is used first to impose an entropy penalty on the parameter representation during training, and second to compress the representation using arithmetic coding after training. The proposed approach can thus maximize accuracy and model compressibility jointly, in an end-to-end fashion, with the rate-error trade-off specified by a hyperparameter.

Abstract: Data transfer in a secure processing environment is provided. A digital assistant can receive audio input detected by a microphone of a computing device. The digital assistant can determine, based on the audio input, to invoke a third-party application associated with the computing device. The digital assistant can generate, responsive to the determination to invoke the third-party application, a packaged data object. The digital assistant can forward, to the third-party application invoked by the digital assistant component to execute in a secure processing environment on the computing device, the packaged data object. The third-party application can transmit, responsive to a digital component request triggered in the third-party application, the packaged data object to a digital component selector to execute a real-time selection process based on the packaged data object.

Abstract: A simulation test is run on a first machine learning model trained using first training data historically collected over a time period. The first training data includes a set of training inputs and a set of target outputs. In response to a determination that a result of the simulation test run on the first machine learning model satisfies one or more criteria, a size of the set of target outputs of the first training data is determined. Second training data for training a second machine learning model is obtained. A size of a set of target outputs of the second training data meets or exceeds the size of the target outputs of the first training data. The second machine learning model is trained using the second training data.

Abstract: An example method includes determining, by a controller of an image capture system, a plurality of sets of exposure parameter values for one or more exposure parameters. The plurality of sets of exposure parameter values are determined at an exposure determination rate. The method further includes capturing, by an image capture device of the image capture system, a plurality of images. Each image of the plurality of images is captured according to a set of exposure parameter values of the plurality of sets of exposure parameter values. The method also includes sending, by the controller of the image capture system to an image processing unit, a subset of the plurality of images. Each subset of images is sent at a sampling rate, and the sampling rate is less than the exposure determination rate.

Abstract: Replicas are selected in a large distributed network, and the roles for these replicas are identified. In one example, a leader is selected from among candidate computing clusters. To make this selection, an activity monitor predicts or monitors the workload of one or more clients. Different activities of the workload are given corresponding weights. The delay in performing requested activities, modified by these weights is found, and the candidate leader with the lowest weighted delay is selected as the leader.