Abstract: Encoding and decoding using warped reference list includes generating a reconstructed frame from an encoded bitstream by, for decoding a current block for the reconstructed frame, obtaining a dynamic reference list, obtaining a warped reference list, decoding a warped reference list index value, obtaining optimal predicted warped model parameters from the warped reference list in accordance with the index value, decoding differential warped model parameters, obtaining, as optimal warped model parameters, a result of adding the optimal predicted warped model parameters and the differential warped model parameters, obtaining predicted block data in accordance with the optimal warped model parameters, decoding residual block data, and obtaining, as decoded block data for the current block, a result of adding the residual block data and the predicted block data.

Abstract: Techniques are described herein for providing people suggestions in collaborative online text editors. A method includes: receiving user interface input that corresponds to a document in a document editing application; automatically parsing the received user interface input to identify a name included in the user interface input; in response to identifying the name included in the user interface input, providing an option to create a link in the document between the name and a corresponding contact in a contact store; receiving additional user interface input that indicates acceptance of the option to create the link in the document; and in response to receiving the additional user interface input, automatically creating the link in the document between the name and the corresponding contact in the contact store.

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: Implementations described herein generally relate to generating a modification selectable element that may be provided for presentation to a user in a smart dictation session with an automated assistant. The modification selectable element may, when selected, cause a transcription, that includes textual data generated based on processing audio data that captures a spoken utterance and that is automatically arranged, to be modified. The transcription may be automatically arranged to include spacing, punctuation, capitalization, indentations, paragraph breaks, and/or other arrangement operations that are not specified by the user in providing the spoken utterance. Accordingly, a subsequent selection of the modification selectable element may cause these automatic arrangement operation(s), and/or the textual data locationally proximate to these automatic arrangement operation(s), to be modified.

Abstract: An optical engine of a display system includes first and second sets of laser light sources, which may be independently controllable, with laser light from each set of laser light sources being respectively combined to form first and second elliptical laser light beams. The first and second elliptical laser light beams may propagate along parallel or angularly separated optical paths prior to incidence at a scan mirror of an optical scanner of the system. In some embodiments, the first and second elliptical laser light beams are incident on separate regions of the reflective surface of the scan mirror. In some embodiments, the first and second elliptical laser light beams are incident on partially overlapping regions of the reflective surface of the scan mirror.

Abstract: An alert peripheral device that provides sensory notification to a user of the device includes: a power subsystem; a communication mechanism by which notification signals is received from a first user equipment (UE) that generates and transmits the notification signals in response to detection of specific events at the first UE; and a response notification mechanism that provides a sensory response of the peripheral device following receipt of a notification of a detected event (NDE) signal. The device further includes an embedded controller coupled to each of the other components and which includes firmware that when executed on the embedded controller configures the embedded controller to: establish a communication link between the communication mechanism and the first UE; and in response to detecting a receipt of the NDE signal from the first UE, trigger the response notification mechanism to exhibit the sensory response.

Abstract: Methods, apparatus, and computer readable media related to interaction between a user and an automated assistant during a dialog between the user and the automated assistant. Some implementations are directed to adapting a graphical and/or audible presentation of search results provided by the automated assistant for presentation to the user. The adaptation may be in response to attribute(s), of one or more of the search results, referenced in spoken and/or typed textual input provided by the user during the dialog. Some of those implementations may enable a user to provide textual input to navigate the search results within the dialog and within resource and/or interface constraints associated with the dialog. Some of those implementations may additionally and/or alternatively enable adapting, based on textual input provided by a user to the automated assistant, when and/or whether search results having certain attributes are provided to the user by the automated assistant.

Abstract: A head-mounted display (HMD) system including a lens element supported by a support structure, the lens element having a waveguide with an incoupler configured to receive light from an optical scanner of the HMD. The incoupler is configured with multiple features varying in at least one of height, spacing, angle, or density. The features may be separated into discrete zones along the incoupler such that at least one of height, spacing, angle, or density of the plurality of features is varied over the incoupler and constant within a given zone or the features may be varied continuously across the incoupler.

Abstract: Systems and methods for automatically generating remarketing lists based on client sessions are provided. The system can include a visitor management module to identify client sessions established at a website. The visitor management module can determine node pairs including destination nodes and referrer nodes. The system can include a graph construction module to construct a graph including the node pairs and edges between the node pairs representing click-through rates. The system can include a graph clustering module to assign information resources to a first group based on one of i) a number of edges between the nodes and a conversion node or ii) click-through rates of node pairs including. The system includes a remarketing list generation module to assign the first group of information resources to a remarketing list. The remarketing list generation module can assign client identifiers that accessed the first group of information resources to the remarketing list.

Abstract: Systems and methods for providing dialog data, from an initially invoked automated assistant to a subsequently invoked automated assistant. A first automated assistant may be invoked by a user utterance, followed by a dialog with the user that is processed by the first automated assistant. During the dialog, a request to transfer dialog data to a second automated assistant is received. The request may originate with the user, by the first automated assistant, and/or by the second automated assistant. Once authorized, the first automated assistant provides the previous dialog data to the second automated assistant. The second automated assistant performs one or more actions based on the dialog data.

Abstract: Techniques are described for selectively adapting and/or selectively utilizing a noise reduction technique in detection of one or more features of a stream of audio data frames. For example, various techniques are directed to selectively adapting and/or utilizing a noise reduction technique in detection of an invocation phrase in a stream of audio data frames, detection of voice characteristics in a stream of audio data frames (e.g., for speaker identification), etc. Utilization of described techniques can result in more robust and/or more accurate detections of features of a stream of audio data frames in various situations, such as in environments with strong background noise. In various implementations, described techniques are implemented in combination with an automated assistant, and feature(s) detected utilizing techniques described herein are utilized to adapt the functionality of the automated assistant.

Abstract: Implementations set forth herein relate to selectively relying on additional suggestion model(s) when generating selectable suggestions, while also maintaining access to a default suggestion model. The selectable suggestions can be generated using one or more additional multi-domain machine learning (ML) models, which can be optionally available to the client application, regardless of whether a default suggestion model remains useful for generating suitable suggestions. In some implementations, as the client application employs various additional multi-domain ML models, a particular model can be identified as improving suggestions for the client application, at least based on user feedback and/or other data. The particular model can then be selected to replace and/or supplement the default suggestion model, in order to provide more accurate suggestions that, when selected, initialize actions that can preserve time and computational resources.

Abstract: A UE activates a timer for applying access control to transmissions associated with a certain access category during an access control period, in response to a first system information message received via a radio interface and a mobile-originated access request (1702). While the timer is running, the UE receives a second message that indicates a potential transition of the UE (i) from a current state associated with a protocol for controlling radio resources to another state associated with the protocol, or (ii) from a current cell to a new cell (1704). In response to the second message, the UE continues to apply the access control to the transmissions for the access category, for a remainder of the access control period (1706).

Abstract: Provided is a knowledge distillation technique for training a student language model that, relative to a larger teacher language model, has a significantly smaller vocabulary, lower embedding dimensions, and/or hidden state dimensions. Specifically, aspects of the present disclosure are directed to a dual-training mechanism that trains the teacher and student language models simultaneously to obtain optimal word embeddings for the student vocabulary. In some implementations, this approach can be combined with learning shared projection matrices that transfer layer-wise knowledge from the teacher language model to the student language model. Example experimental results have also demonstrated higher compression efficiency and accuracy when compared with other state-of-the-art compression techniques, including the ability to compress the BERTBASE model by more than 60×, with only a minor drop in downstream task metrics, resulting in a language model with a footprint of under 7 MB.

Abstract: Implementations relate to mitigating client device latency in rendering of remotely generated automated assistant content. Some of those implementations mitigate client device latency between rendering of multiple instances of output that are each based on content that is responsive to a corresponding automated assistant action of a multiple action request. For example, those implementations can reduce latency between rendering of first output that is based on first content responsive to a first automated assistant action of a multiple action request, and second output that is based on second content responsive to a second automated assistant action of the multiple action request.

Abstract: Systems and methods for calibrating and initializing a system comprising a first qubit, a second qubit, and a coupler are provided. A method of initializing the system can include calibrating the system; providing a first idle resonance frequency of the first qubit and a second idle resonance frequency of the second qubit; providing the first bias for the first idle resonance frequency and the second bias for the second idle resonance frequency; determining a value of the third resonance frequency at which a difference between the first degree of coupling and the second degree of coupling is minimized, when the first resonance frequency is the first idle resonance frequency and the second resonance frequency is the second idle resonance frequency; and providing the third bias according to the determined value of the third resonance frequency.

Abstract: A portion Y of a degraded frame is restored using a projection operation that uses a first projection parameter ?, a second projection parameter ?, and at least two guide portions. Restoring the portion Y of the degraded frame includes generating, using first restoration parameters, a first guide portion Y1 for the portion Y; generating, using second restoration parameters, a second guide portion Y2 for the portion Y; and generating a reconstructed portion YR, wherein the projection operation is based on ?(Y1?Y)+?(Y2?Y).

Abstract: Systems and methods for augmented-reality tutoring can utilize optical character recognition, natural language processing, and/or augmented-reality rendering for providing real-time notifications for completing a determined task. The systems and methods can include utilizing one or more machine-learned models trained for quantitative reasoning and can include providing a plurality of different user interface elements at different times.

Abstract: Implementations described herein are directed to enabling collaborative ranking of interpretations of spoken utterances based on data that is available to an automated assistant and third-party agent(s), respectively. The automated assistant can determine first-party interpretation(s) of a spoken utterance provided by a user, and can cause the third-party agent(s) to determine third-party interpretation(s) of the spoken utterance provided by the user. In some implementations, the automated assistant can select a given interpretation, from the first-party interpretation(s) and the third-party interpretation(s), of the spoken utterance, and can cause a given third-party agent to satisfy the spoken utterance based on the given interpretation.