In general, the subject matter described in this specification can be embodied in methods, systems, and program products for receiving user input that defines a search query, and providing the search query to a server system. Information that a search engine system determined was responsive to the search query is received at a computing device. The computing device is identified as in a first state, and a first output mode for audibly outputting at least a portion of the information is selected. The first output mode is selected from a collection of the first output mode and a second output mode. The second output mode is selected in response to the computing device being in a second state and is for visually outputting at least the portion of the information and not audibly outputting the at least portion of the information. At least the portion of information is audibly output.
G06F 40/58 - Utilisation de traduction automatisée, p. ex. pour recherches multilingues, pour fournir aux dispositifs clients une traduction effectuée par le serveur ou pour la traduction en temps réel
G10L 13/00 - Synthèse de la paroleSystèmes de synthèse de la parole à partir de texte
G10L 15/22 - Procédures utilisées pendant le processus de reconnaissance de la parole, p. ex. dialogue homme-machine
G10L 15/30 - Reconnaissance distribuée, p. ex. dans les systèmes client-serveur, pour les applications en téléphonie mobile ou réseaux
H04M 1/60 - Équipement de sous-station, p. ex. pour utilisation par l'abonné comprenant des amplificateurs de parole
H04M 1/72454 - Interfaces utilisateur spécialement adaptées aux téléphones sans fil ou mobiles avec des moyens permettant d’adapter la fonctionnalité du dispositif dans des circonstances spécifiques en tenant compte des contraintes imposées par le contexte ou par l’environnement
H04R 29/00 - Dispositifs de contrôleDispositifs de tests
2.
PRIVACY-SENSITIVE TRAINING OF USER INTERACTION PREDICTION MODELS
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for collaboratively training an interaction prediction machine learning model using a plurality of user devices in a manner that respects user privacy. In one aspect, the machine learning model is configured to process an input comprising: (i) a search query, and (ii) a data element, to generate an output which characterizes a likelihood that a given user would interact with the data element if the data element were presented to the given user on a webpage identified by a search result responsive to the search query.
G06F 18/214 - Génération de motifs d'entraînementProcédés de Bootstrapping, p. ex. ”bagging” ou ”boosting”
G06F 17/18 - Opérations mathématiques complexes pour l'évaluation de données statistiques
G06F 18/2415 - Techniques de classification relatives au modèle de classification, p. ex. approches paramétriques ou non paramétriques basées sur des modèles paramétriques ou probabilistes, p. ex. basées sur un rapport de vraisemblance ou un taux de faux positifs par rapport à un taux de faux négatifs
G06V 10/46 - Descripteurs pour la forme, descripteurs liés au contour ou aux points, p. ex. transformation de caractéristiques visuelles invariante à l’échelle [SIFT] ou sacs de mots [BoW]Caractéristiques régionales saillantes
3.
Container Device And Delivery Systems For Using The Same
Container devices and delivery systems for using the same are provided. In accordance with some embodiments of the disclosed subject matter, a method for delivering packages is provided that includes: receiving, at a delivery hub, a first package to be delivered to a recipient; causing the first package to be placed in a container to be delivered to the recipient; associating an identifier of the first package and an identifier of the container with the recipient; determining, at a first time point, whether the container is ready to be delivered to the recipient; in response to determining that the container is not ready to be delivered to the recipient, waiting for a second package to be delivered to the recipient; receiving the second package to be delivered to the recipient; causing the second package to be placed in the container; associating an identifier of the second package with the identifier of the container; determining, at a second time point, whether the container is ready to be delivered to the recipient; and, in response to determining that the container is ready to be delivered to the recipient, causing the container to be loaded onto a delivery vehicle.
G06Q 10/0832 - Marchandises spéciales ou procédures de manutention spéciales, p. ex. manutention de marchandises dangereuses ou fragiles
A47G 29/14 - Récipients pour déposer des aliments, p. ex. petit déjeuner, laitRécipients similaires pour colis avec accessoires pour éviter que les articles déposés ne soient indûment retirés
G06F 16/901 - IndexationStructures de données à cet effetStructures de stockage
G06Q 10/0631 - Planification, affectation, distribution ou ordonnancement de ressources d’entreprises ou d’organisations
G06Q 10/087 - Gestion d’inventaires ou de stocks, p. ex. exécution des commandes, approvisionnement ou régularisation par rapport aux commandes
G07C 9/00 - Enregistrement de l’entrée ou de la sortie d'une entité isolée
G07C 9/22 - Enregistrement de l’entrée ou de la sortie d'une entité isolée comportant l’utilisation d’un laissez-passer combiné à une vérification d’identité du titulaire du laissez-passer
G07C 9/27 - Enregistrement de l’entrée ou de la sortie d'une entité isolée comportant l’utilisation d’un laissez-passer une station centrale gérant l’enregistrement
This document describes techniques and devices for a radar recognition-aided search. Through use of a radar-based recognition system, gestures made by, and physiological information about, persons can be determined. In the case of physiological information, the techniques can use this information to refine a search. For example, if a person requests a search for a coffee shop, the techniques may refine the search to coffee shops in the direction that the person is walking. In the case of a gesture, the techniques may refine or base a search solely on the gesture. Thus, a search for information about a store, car, or tree can be made responsive to a gesture pointing at the store, car, or tree with or without explicit entry of a search query.
A hinge for a computing device includes at least one antenna extending from a body portion of the hinge into a body portion of the computing device. When incorporated into a head mounted wearable device, the hinge couples an arm portion to a front frame portion of the device, with the at least one antenna extending into an installation area defined by the front frame portion. The at least one antenna does not occupy installation space in the arm portions, providing additional space for other electrical components of the device, and provides for modularity in coupling a variety of arm portions housing different arrangements and/or combinations of electrical components to a variety of different front frames.
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for distributing digital component while securing user data are described. In one aspect, a method includes receiving, by a multi-platform server and from a client device, a request for a digital component for presentation by the client device. The request for the digital component includes (i) request data that is opaque to the multi-platform server, and (ii) sensitive user data that is managed by the client device. In response to receiving the request for the digital component, the multi-platform server transmits, to a first content platform, a contextual request that includes the request data from the client device and that does not include the sensitive user data. After transmitting the contextual request, the multi-platform server receives, from the first content platform, a contextual response that includes a selection data unit for a first repository of digital components.
A method for a statistics collection framework includes receiving a schema defining a relational database for storing a plurality of statistics corresponding to a query, the relational database including a plurality of data tables relationally connected according to the schema, each data table of the plurality of data tables corresponding to a respective statistic. The method includes receiving a query corresponding to data at a data store. The method also includes executing the query. During execution of the query, the method includes collecting, from a query execution database, the plurality of statistics related to the query, each statistic of the plurality of statistics corresponding to a respective data table of the plurality of data tables of the relational database and, for each statistic of the plurality of statistics, storing the respective statistic at the respective data table according to the schema.
A watch system may include a watchband including a flexible member configured to be mounted onto a wrist of a user, and a puck including watch functionality. The watchband may have an end that has a concave curved shape. The puck may have a connection interface that has a convex curved shape. The connection interface may be to be removably coupled to the end of the watchband. The watchband and the puck may have corresponding locking features that are configured to rotationally and translationally fix the puck to the watchband. The corresponding locking features may be configured to be engaged when the watchband is translated relative to the puck and rotated relative to the puck by a predetermined rotation angle.
Methods, systems, and apparatus, for a high throughput scan architecture. The scan architecture can include a clock controller, a decompressor, a scan chain, and a compressor. In some implementations, a set of values that represents a particular data pattern is received. A first data signal is generated using at least a portion of the values in the set of values, where the first data signal has a first frequency. A first series of latches and a second series of latches are used to extract alternating values of the at least portion of values from the first data signal, where the first series and second series of latches extract the alternating values at a second frequency that is a fraction of the first frequency. Outputs of the first and second series of latches are combined to generate a second data signal, where the second data signal has the first frequency.
One or more nodes of a radio access network (RAN) perform a method of configuring or reconfiguring a user equipment (UE). The method includes communicating (802) data with the UE while the UE is in an inactive state and configured for small data transmission (SDT) operation, and determining (804) to configure or reconfigure one or more radio resources for the UE while the UE is in the inactive state and configured for SDT operation. The method also includes, transmitting (810), in response to the determining and while the UE is in the inactive state and configured for SDT operation, a message to the UE to configure or reconfigure the one or more radio resources for the UE.
A radio access network (RAN); a core network (CN) or operations, administration, and management (OAM) node; and a user equipment (UE) can implement a method for managing Quality of Experience (QoE) reporting from the UE. The method includes: facilitating QoE reporting to a QE node for the UE; determining to perform a handover for the UE from a source node of the RAN to a target node of the RAN; and pausing the QoE reporting. The method may further providing information to identify a QoE configuration of multiple configurations. The information may include QoE configurations, reference identifiers, QoE configuration identifiers, and information about associations between the reference identifiers and the QoE configuration identifiers.
A method includes obtaining a plurality of training samples that each include a respective speech utterance and a respective textual utterance representing a transcription of the respective speech utterance. The method also includes fine-tuning, using quantization and sparsity aware training with native integer operations, a pre-trained automatic speech recognition (ASR) model on the plurality of training samples. Here, the pre-trained ASR model includes a plurality of weights and the fine-tuning includes pruning one or more weights of the plurality of weights using a sparsity mask and quantizing each weight of the plurality of weights based on an integer with a fixed-bit width. The method also includes providing the fine-tuned ASR model to a user device.
G10L 15/06 - Création de gabarits de référenceEntraînement des systèmes de reconnaissance de la parole, p. ex. adaptation aux caractéristiques de la voix du locuteur
13.
DECENTRALIZED LEARNING OF LARGE MACHINE LEARNING (ML) MODEL(S)
Implementations described herein are directed to a framework for decentralized learning of large global machine learning (ML) model(s). In various implementations, remote processor(s) of a remote system can identify a global ML model, select client devices to participate in a given round of decentralized learning of the global ML model, and transmit, to each of the client devices, a processed version of the global ML model that is of a reduced transferrable size. Further, client device processor(s) of a client device can receive the processed version of the global ML model, obtain corresponding client data, perform partial model training, based on processing the corresponding client data, for the processed version of the global ML model to generate a corresponding update, and transmit the corresponding update back to the remote system. Moreover, the remote processor(s) can update, based on at least the corresponding update, the global ML model.
G10L 15/06 - Création de gabarits de référenceEntraînement des systèmes de reconnaissance de la parole, p. ex. adaptation aux caractéristiques de la voix du locuteur
G06N 3/098 - Apprentissage distribué, p. ex. apprentissage fédéré
G10L 15/183 - Classement ou recherche de la parole utilisant une modélisation du langage naturel selon les contextes, p. ex. modèles de langage
G10L 15/30 - Reconnaissance distribuée, p. ex. dans les systèmes client-serveur, pour les applications en téléphonie mobile ou réseaux
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium for training and using distilled machine learning models. In one aspect, a method includes obtaining a first input that includes training example sets that each include one or more feature values and, for each item, an outcome label that represents whether the item had a positive outcome. A first machine learning model is trained using the first input and is configured to generate a set of scores that represents whether the item will have a positive outcome when presented in the context of the training example set and with each other item in the example set. A distilled machine learning model is trained using the set of scores for each example set. The distilled machine learning model is configured to generate a distilled score.
Systems and methods for generating a 3D effect for a video stream are provided. A first video stream from a first client device of a first participant of a virtual meeting and a second video stream from a second client device of a second participant of the virtual meeting is identified. A background and a foreground layer of the first video stream is determined. A first and a second eye position of the second participant of the second video stream are determined. A presentation position of the background layer relative to the foreground layer is determined based on movement between the first and the second eye position of the second participant of the second video stream. A UI presenting the first video stream reflecting the determined presentation position of the background layer relative to the foreground layer is provided for display on the second client device.
G11B 27/02 - Montage, p. ex. variation de l'ordre des signaux d'information enregistrés sur, ou reproduits à partir des supports d'enregistrement ou d'information
16.
CONSTRUCTING AND PROGRAMMING QUANTUM HARDWARE FOR ROBUST QUANTUM ANNEALING PROCESSES
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.
G06N 10/40 - Réalisations ou architectures physiques de processeurs ou de composants quantiques pour la manipulation de qubits, p. ex. couplage ou commande de qubit
G06F 15/82 - Architectures de calculateurs universels à programmes enregistrés commandés par des données ou à la demande
G06F 17/11 - Opérations mathématiques complexes pour la résolution d'équations
G06N 7/01 - Modèles graphiques probabilistes, p. ex. réseaux probabilistes
G06N 10/20 - Modèles d’informatique quantique, p. ex. circuits quantiques ou ordinateurs quantiques universels
G06N 10/60 - Algorithmes quantiques, p. ex. fondés sur l'optimisation quantique ou les transformées quantiques de Fourier ou de Hadamard
This document describes a vapor chamber within an electronic device. In aspects, an electronic device includes a middle frame that provides mechanical support for the electronic device, a middle plate affixed to the middle frame to define an inner layer of a chassis, and a vapor chamber disposed inside the middle plate. The vapor chamber includes a first region proximate to a heat source and a second region opposite the first region. A coolant is evaporated in a first mode at the first region by heat absorbed from the heat source and is condensed in a second mode in the second region. This vapor chamber permits cooling of elements within the electronic device at lower cost and/or smaller size than many conventional cooling systems.
Various arrangements are presented for performing a hierarchical application launch of an application. A requests to register one or more smart home devices can be received in association with a user account. The smart home devices can be mapped to the user account based on receiving the requests. An application that is mapped to the user account may be launched; the application can analyze the one or more smart home devices registered to the user account and a user interface hierarchy. Based on analyzing the one or more smart home devices registered to the user account and the user interface hierarchy, an initial launch interface can be selected and output for presentation.
H04M 1/72415 - Interfaces utilisateur spécialement adaptées aux téléphones sans fil ou mobiles avec des moyens de soutien local des applications accroissant la fonctionnalité par interfaçage avec des accessoires externes pour la télécommande d’appareils
G05B 15/02 - Systèmes commandés par un calculateur électriques
19.
EFFICIENT, FLEXIBLE, AND SECURE DYNAMIC DIGITAL CONTENT CREATION
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating dynamic digital content in privacy preserving ways are described. In one aspect, a method includes receiving, by a trusted server and from multiple content platforms, digital component data for digital components. The server received, from each content platform, dynamic content selection logic for selecting discrete content elements for digital components of the content platform. The server selects, from digital components for which digital component data is stored in a digital component repository, candidate digital components based at least on user data included in a digital component request. For each candidate digital component, the server executes the dynamic content selection logic of the content platform that provided the digital component data for the candidate digital component, the executing resulting in selection of a particular layout and a particular subset of content elements for the digital component.
A computer-implemented method for determining a score indicative of mindfulness of a user is provided. The method includes obtaining heart rate variability data of the user during a guided breathing exercise in which the user inhales and exhales to mimic a respiration rate associated with mindfulness. The method includes filtering the heart rate variability data to generate filtered heart rate variability. The method includes determining a first standard deviation of interbeat intervals indicative of respiratory sinus arrythmia and included in a first segment of the filtered heart rate variability data that spans a discrete interval of time. The method includes determining a second standard deviation of all interbeat intervals included in the first segment of the filtered heart rate variability data. The method includes determining a score indicative of mindfulness of the user based on the first standard deviation and the second standard deviation.
Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for privacy preserving digital component provider. In some implementations, a method includes providing, by a user device and during a browsing session of content page at the user device, (1) a request for a digital component and (2) contextual data representing a context within which the content page is provided for display on the user device; obtaining an embedding vector that represents the contextual data as a set of features and the digital component; generating one or more adjusted embedding vectors for a first interest group, wherein the collection includes the embedding vector adjusted by one or more values; and providing the one or more adjusted embedding vectors to a server for generating a model for the first interest group.
G06F 16/9535 - Adaptation de la recherche basée sur les profils des utilisateurs et la personnalisation
G06F 16/955 - Recherche dans le Web utilisant des identifiants d’information, p. ex. des localisateurs uniformisés de ressources [uniform resource locators - URL]
G06F 21/62 - Protection de l’accès à des données via une plate-forme, p. ex. par clés ou règles de contrôle de l’accès
22.
GENERATING AUDIO WAVEFORMS USING ENCODER AND DECODER NEURAL NETWORKS
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for processing an input audio waveform using a generator neural network to generate an output audio waveform. In one aspect, a method comprises: receiving an input audio waveform; processing the input audio waveform using an encoder neural network to generate a set of feature vectors representing the input audio waveform; and processing the set of feature vectors representing the input audio waveform using a decoder neural network to generate an output audio waveform that comprises a respective output audio sample for each of a plurality of output time steps.
Methods, systems, and apparatus, including medium-encoded computer program products, for aggregating accounts using machine learning. User interaction data can be obtained for a user and can describe interactions by the user with a given account of multiple different accounts assigned to the user on one or more computer systems. An input that includes the user interaction data is processed using a machine learning model that is configured to produce a result that includes a first account embedding that differs from the user interaction data. From at least the first account embedding, an account group is determined that corresponds to the user interaction data. A first action is performed based on the account group, wherein the first action differs from a second action that would have been performed based on a different account group that is not the account group.
A method includes receiving a sequence of acoustic frames characterizing a spoken utterance in a particular native language. The method also includes generating a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames by a causal encoder that includes an initial stack of multi-head attention layers. The method also includes generating a second higher order feature representation for a corresponding first higher order feature representation by a non-causal encoder that includes a final stack of multi-head attention layers. The method also includes receiving, as input at each corresponding language-dependent adapter (LDA) module, a language ID vector identifying the particular native language to activate corresponding language-dependent weights specific to the particular native language. The method also includes generating a first probability distribution over possible speech recognition hypotheses by a decoder.
G10L 15/02 - Extraction de caractéristiques pour la reconnaissance de la paroleSélection d'unités de reconnaissance
G10L 15/06 - Création de gabarits de référenceEntraînement des systèmes de reconnaissance de la parole, p. ex. adaptation aux caractéristiques de la voix du locuteur
G10L 15/30 - Reconnaissance distribuée, p. ex. dans les systèmes client-serveur, pour les applications en téléphonie mobile ou réseaux
Aspects of the disclosure are directed to merging data lake openness with scalable metadata for managed tables in a cloud database platform, allowing for atomicity, consistency, isolation, and durability (ACID) transactions, performant data manipulation language (DML), higher throughput stream ingestion, data consistency, schema evolution, time travel, clustering, fine-grained security, and/or automatic storage optimization. Table data is stored in various open-source file formats in cloud storage while physical metadata of the table data is stored in a scalable metadata storage system.
Aspects of the disclosure are directed to merging data lake openness with scalable metadata for managed tables in a cloud database platform, allowing for atomicity, consistency, isolation, and durability (ACID) transactions, performant data manipulation language (DML), higher throughput stream ingestion, data consistency, schema evolution, time travel, clustering, fine-grained security, and/or automatic storage optimization. Table data is stored in various open-source file formats in cloud storage while physical metadata of the table data is stored in a scalable metadata storage system.
A method includes receiving training data that includes unspoken text utterances and un-transcribed non-synthetic speech utterances. Each unspoken text utterance is not paired with any corresponding spoken utterance of non-synthetic speech. Each un-transcribed non-synthetic speech utterance is not paired with a corresponding transcription. The method also includes generating a corresponding synthetic speech representation for each unspoken textual utterance of the received training data using a text-to-speech model. The method also includes pre-training an audio encoder on the synthetic speech representations generated for the unspoken textual utterances and the un-transcribed non-synthetic speech utterances to teach the audio encoder to jointly learn shared speech and text representations.
G10L 13/047 - Architecture des synthétiseurs de parole
G10L 13/08 - Analyse de texte ou génération de paramètres pour la synthèse de la parole à partir de texte, p. ex. conversion graphème-phonème, génération de prosodie ou détermination de l'intonation ou de l'accent tonique
A method includes receiving audio data corresponding to an utterance spoken by a particular user and captured in streaming audio by a user device. The method also includes performing speaker identification on the audio data to identify an identity of the particular user that spoke the utterance. The method also includes obtaining a keyword detection model personalized for the particular user based on the identity of the particular user that spoke the utterance. The keyword detection model is conditioned on speaker characteristic information associated with the particular user to adapt the keyword detection model to detect a presence of a keyword in audio for the particular user. The method also includes determining that the utterance includes the keyword using the keyword detection model personalized for the particular user.
G10L 17/02 - Opérations de prétraitement, p. ex. sélection de segmentReprésentation ou modélisation de motifs, p. ex. fondée sur l’analyse linéaire discriminante [LDA] ou les composantes principalesSélection ou extraction des caractéristiques
G10L 17/04 - Entraînement, enrôlement ou construction de modèle
Methods, systems, and apparatus, including medium-encoded computer program products, for secure workflows that enhance data security are described. In one aspect, a digital component request is received. In response to receiving the digital component request, a multi-stage workflow for selecting a digital component is identified, and can include customizable stages. The execution of workflow stages includes: (A) identifying a given customizable stage; (B) for the stage: (i) identifying, a customization specific to the stage that generates an output for use in selecting the digital component; (ii) initiating an isolated execution environment for each customization; (iii) executing, within each isolated execution environment, the customization for which the isolated execution environment was initiated; and (iv) obtaining the output generated by the code of each isolated execution environment; and (C) executing a final stage to select a digital component based on the outputs. The selected digital component is sent to the client device.
G06F 21/53 - Contrôle des utilisateurs, des programmes ou des dispositifs de préservation de l’intégrité des plates-formes, p. ex. des processeurs, des micrologiciels ou des systèmes d’exploitation au stade de l’exécution du programme, p. ex. intégrité de la pile, débordement de tampon ou prévention d'effacement involontaire de données par exécution dans un environnement restreint, p. ex. "boîte à sable" ou machine virtuelle sécurisée
A method including determining a viewpoint, generating a first image using an image generator, the first image including an object in a first orientation based on the viewpoint, modifying the image generator based on a second orientation of the object, and generating a second image based on the first image using the modified image generator.
Aspects of the disclosure are directed to a transactional agent for user interactions. The agent can seamlessly respond to user requests in a conversational manner while maintaining the conversational state. The agent can include a multi-stage modular model architecture, including a semantic understander and a semantic matcher. The semantic understander can be configured to understand common conversation conventions and/or patterns to produce a structure representation of a user request. The semantic matcher can be configured to map items and modifiers to product entries for a particular domain.
Provided are computer systems which demonstrate improved cryptographic agility via inclusion of multiple cryptographic operating modes. In one example, one or more devices included within a computing system are designed to include multiple cryptographic operating modes from the outset (e.g., prior to deployment of the system, “by design”). Additionally or alternatively, one or more devices included within the computing system (e.g., a gateway computing device) can be updated to include the multiple cryptographic operating modes after deployment of the system (e.g., in an “ad hoc” fashion). A system can also include both device(s) that have multiple operating modes by design and device(s) that have multiple operating modes introduced in an ad hoc fashion. Inclusion of the multiple cryptographic operating modes can serve to enhance the security of at least the communications of the computing system that are performed by or follow the updated device(s).
H04L 9/14 - Dispositions pour les communications secrètes ou protégéesProtocoles réseaux de sécurité utilisant plusieurs clés ou algorithmes
H04L 9/32 - Dispositions pour les communications secrètes ou protégéesProtocoles réseaux de sécurité comprenant des moyens pour vérifier l'identité ou l'autorisation d'un utilisateur du système
33.
Golden Prompt Generation based on Authoritative Publications
A method for golden prompt generation based on authoritative publications includes receiving an initial authoritative publication associated with a specific topic. The method includes retrieving, using the initial authoritative publication additional authoritative publications associated with the specific topic. The method includes generating, using natural language processing, a set of golden prompts from the set of authoritative publications. Each golden prompt of the set of golden prompts includes text from the set of authoritative publications. The method includes fine-tuning a pre-trained model using the set of authoritative publications. The method includes generating, using the fine-tuned model and the set of golden prompts, a set of predictions. The method includes determining, using the set of predictions and the set of authoritative publications, an error rate of the fine-tuned model. The error rate indicates a similarity between the set of predictions and the set of authoritative publications.
A heat exchanger includes a first manifold having an inlet opening and a second manifold having an outlet opening. A group of conduits fluidly connect the first manifold and the second manifold to one another such that a flow path is established for liquid to flow from the inlet opening to the outlet opening. The flow path includes a select portion that extends through all conduits within the group of conduits. Valves are located in the first manifold and the second manifold. The valves are operable to change the select portion of the flow path from between a first state, wherein the conduits within group of conduits are fluidly connected in parallel with one another, and a second state, wherein the conduits within the group of conduits are fluidly connected in series with one another.
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for obtaining a first schedule, for a first hardware block of an integrated circuit device, where the first schedule identifies a first set of operations to be performed by the first hardware block. Obtaining a second schedule for a second hardware block of the integrated circuit device, where the second schedule identifies a second set of operations to be performed by the second hardware block and where operations of the second schedule are coordinated with operations of the first schedule such that the first schedule triggers the first hardware block to send data to the second block at a first pre-scheduled value of a counter, and the second schedule triggers the second hardware block to accept the data at an input at a second pre-scheduled value of the counter that is after the first pre-scheduled value. Performing, by the first hardware block, the first set of operations according to the first schedule, and performing, by the second hardware block, the second set of operations according to the second schedule.
G06F 9/48 - Lancement de programmes Commutation de programmes, p. ex. par interruption
G06F 9/30 - Dispositions pour exécuter des instructions machines, p. ex. décodage d'instructions
G06F 9/32 - Formation de l'adresse de l'instruction suivante, p. ex. par incrémentation du compteur ordinal
G06F 9/38 - Exécution simultanée d'instructions, p. ex. pipeline ou lecture en mémoire
G06F 15/78 - Architectures de calculateurs universels à programmes enregistrés comprenant une seule unité centrale
G06N 3/063 - Réalisation physique, c.-à-d. mise en œuvre matérielle de réseaux neuronaux, de neurones ou de parties de neurone utilisant des moyens électroniques
36.
SYSTEM AND METHOD FOR VIDEO ENCODING USING CONSTRUCTED REFERENCE FRAME
Video coding using constructed reference frames may include generating, by a processor in response to instructions stored on a non-transitory computer readable medium, a reconstructed video. Generating the reconstructed video may include receiving an encoded bitstream. Video coding using constructed reference frames may include generating a reconstructed non-showable reference frame. Generating the reconstructed non-showable reference frame may include decoding a first encoded frame from the encoded bitstream. Video coding using constructed reference frames may include generating a reconstructed frame. Generating the reconstructed frame may include decoding a second encoded frame from the encoded bitstream using the reconstructed non-showable reference frame as a reference frame. Video coding using constructed reference frames may include including the reconstructed frame in the reconstructed video and outputting the reconstructed video.
H04N 19/80 - Détails des opérations de filtrage spécialement adaptées à la compression vidéo, p. ex. pour l'interpolation de pixels
H04N 19/105 - Sélection de l’unité de référence pour la prédiction dans un mode de codage ou de prédiction choisi, p. ex. choix adaptatif de la position et du nombre de pixels utilisés pour la prédiction
H04N 19/107 - Sélection du mode de codage ou du mode de prédiction entre codage prédictif spatial et temporel, p. ex. rafraîchissement d’image
H04N 19/117 - Filtres, p. ex. pour le pré-traitement ou le post-traitement
H04N 19/127 - Établissement des priorités des ressources en matériel ou en calcul
H04N 19/139 - Analyse des vecteurs de mouvement, p. ex. leur amplitude, leur direction, leur variance ou leur précision
H04N 19/172 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant le codage adaptatif caractérisés par l’unité de codage, c.-à-d. la partie structurelle ou sémantique du signal vidéo étant l’objet ou le sujet du codage adaptatif l’unité étant une zone de l'image, p. ex. un objet la zone étant une image, une trame ou un champ
H04N 19/176 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant le codage adaptatif caractérisés par l’unité de codage, c.-à-d. la partie structurelle ou sémantique du signal vidéo étant l’objet ou le sujet du codage adaptatif l’unité étant une zone de l'image, p. ex. un objet la zone étant un bloc, p. ex. un macrobloc
H04N 19/179 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant le codage adaptatif caractérisés par l’unité de codage, c.-à-d. la partie structurelle ou sémantique du signal vidéo étant l’objet ou le sujet du codage adaptatif l’unité étant une scène ou une prise de vues
H04N 19/23 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant le codage d'objets vidéo avec codage des zones présentes sur l’ensemble d’un segment vidéo, p. ex. plans-objets vidéo, image de fond ou mosaïque
H04N 19/527 - Estimation de vecteurs de mouvement globaux
H04N 19/61 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant un codage par transformée combiné avec un codage prédictif
Methods, systems, and apparatus, for sensing a touch on a display panel including an array of pixels. A method includes controlling first pixels to operate in an illumination state; controlling second pixels to repeatedly switch between operating in the illumination state and operating in a sensing state; generating sensing signals indicative of levels of light detected by the second pixels; detecting a touch input to the display panel based on the generated sensing signals; and in response to detecting the touch input to the display panel, changing at least one of (i) a frequency at which the second pixels switch between operating in the illumination state and in the sensing state, (ii) a duty cycle for operating the second pixels in the sensing state, or (iii) which of the pixels in the array of pixels are controlled to switch between operating in the illumination state and in the sensing state.
Provided is a generative sequence processing model that is specifically finetuned for cybersecurity applications. This specialized generative sequence processing model can be finetuned on a comprehensive range of cybersecurity data and associated finetuning tasks, providing the model with rich capabilities for analyzing, understanding, describing, and taking action with respect to the real-time cybersecurity data generated by one or more cybersecurity operations tools. Thus, the creation and use of the generative sequence processing model represents a solution to the technical challenge of usefully analyzing and acting upon a large volume of data generated by a number of disparate cybersecurity operations tools deployed by an organization.
G06F 21/57 - Certification ou préservation de plates-formes informatiques fiables, p. ex. démarrages ou arrêts sécurisés, suivis de version, contrôles de logiciel système, mises à jour sécurisées ou évaluation de vulnérabilité
G06N 3/008 - Vie artificielle, c.-à-d. agencements informatiques simulant la vie fondés sur des entités physiques commandées par une intelligence simulée de manière à reproduire des formes de vie intelligentes, p. ex. fondés sur des robots reproduisant les animaux ou les humains dans leur apparence ou leur comportement
39.
CAPTURING SPATIAL SOUND ON UNMODIFIED MOBILE DEVICES WITH THE AID OF AN INERTIAL MEASUREMENT UNIT
The present disclosure provides computer-implemented methods, systems, and devices for capturing spatial sound for an environment. A computing system captures, using two or more microphones, audio data from an environment around a mobile device. The computing system analyzes the audio data to identify a plurality of sound sources in the environment around the mobile device based on the audio data. The computing system determines, based on characteristics of the audio data and data produced by one or more movement sensors, an estimated location for each respective sound source in the plurality of sound sources. The computing system generates a spatial sound recording of the audio data based, at least in part, on the estimated location of each respective sound source in the plurality of sound sources.
H04R 1/40 - Dispositions pour obtenir la fréquence désirée ou les caractéristiques directionnelles pour obtenir la caractéristique directionnelle désirée uniquement en combinant plusieurs transducteurs identiques
G01S 3/808 - Systèmes pour déterminer une direction ou une déviation par rapport à une direction prédéterminée utilisant des transducteurs espacés et mesurant la différence de phase ou de temps entre les signaux provenant de ces transducteurs, c.-à-d. systèmes à différence de parcours
Features described herein pertain to privacy-preserving tracking. Depth information derived from depth data generated by a sensor that includes photodiodes having a field-of-view within a portion of an environment surrounding an electronic device can be used to recognize a gesture performed by at least one person within the field-of-view and the electronic device can be controlled based on the gesture. The environment surrounding the electronic device can be illuminated with infrared light in which a portion of the infrared light reflected from an object in a field-of-view of the sensor can be captured by the sensor. A center of gravity of the object can be tracked captured light and a notification can be generated based on the tracked center of gravity.
This disclosure provides systems, devices, apparatus, and methods, including computer programs encoded on storage media, for PDU set dependency awareness. A first wireless communication device transmits (810), to a second wireless communication device, an indication of a dependency between first protocol data unit, PDU, Set and a second PDU Set within a PDU Set group or a dependency between a first Data Burst and a second Data Burst within a Data Burst group. The first device transmits (830), to the second device, the first PDU Set or the first Data Burst.
A method includes receiving, by a server hosting a foundation model and via an application programming interface (API), a prompt from a robot that indicates a current set of affordances associated with the robot. The method also includes receiving, by the server and via the API, a query from the robot for instructions to complete a task. The method also includes generating, by the server and using the foundation model, at least one command in response to the query. The at least one command is executable by the robot to complete the task, and the at least one command is generated based on the current set of affordances associated with the robot. The method also includes sending, by the server and via the API, the at least one command to the robot.
The technology described herein is directed to artificial intelligence (Al) powered tools that can generate, enhance, and evaluate digital imagery. For example, the Al-powered tools can be used to generate personalized digital illustrations based on user profile information. In some examples, the tools can modify the personalized digital illustrations, such as by modifying a facial expression of a person depicted in the personalized digital illustration to correspond to a mood or tone of the illustration.
A computing device obtains a query and determines a one or more alternative query terms based on the query. The computing device identifies application modules that match the alternative query terms. The computing device identifies content by at least performing a queryless query, where the queryless query is based on the one or more alternative query terms. The computing device outputs at least one of a query preview that includes an indication of at least one application module from the one or more application modules, or at least one query result that include an indication of the content.
A user equipment (UE) employs a historical profile of previous handovers to respond to a measurement report associated with a cell handover. The historical profile indicates a preferred cell, among a plurality of available cells, wherein the preferred cell is associated with a threshold number of previously successful handovers. Accordingly, by employing the historical profile to respond to the measurement report request, the UE is able to influence the network to select the preferred cell.
The technology provides a flexible schema approach where an app or other service does not need to know anything about language or machine learning in order to interact with natural language associated with a user of the app. This enables tailoring for a particular application or a given task in an application. One method includes receiving a natural language input of a user (702) and then generating a dynamic schema for use with an application running on the computing system (704). The computing system applies, using a trained language model, the dynamic schema to the natural language input to generate schema data having one or more corresponding field and value pairings (706), and provides the schema data to the application (708). The application can generate an output based upon at least one of the field and value pairings (710) and present the output to the user (712).
Systems, devices, methods, and non-transitory, machine-readable media for glass-break sound detection are provided. A glass-break sound detector may be developed. A glass-break sound waveform may be decomposed into sound components that may be recombined in combinations to derive a glass-break sound dataset, which may be developed based on room reverberations and noise. A model may be trained to facilitate detection of glass-break sounds using the glass-break sound dataset. The model may be configured to receive second audio data and predict whether a sound event corresponds to a glass break. A glass-break detection device may monitor sounds in a home and feed an audio stream corresponding to the sounds to a detector. The detector may be configured with the model and may detect a sound event in the audio stream. The detector may predict whether the sound event corresponds to a glass break.
G01H 17/00 - Mesure des vibrations mécaniques ou des ondes ultrasonores, sonores ou infrasonores non prévue dans les autres groupes de la présente sous-classe
G08B 13/04 - Déclenchement mécanique par bris de glace
G10L 25/30 - Techniques d'analyse de la parole ou de la voix qui ne se limitent pas à un seul des groupes caractérisées par la technique d’analyse utilisant des réseaux neuronaux
G10L 25/51 - Techniques d'analyse de la parole ou de la voix qui ne se limitent pas à un seul des groupes spécialement adaptées pour un usage particulier pour comparaison ou différentiation
48.
NON-CODEBOOK BASED TRANSMISSION WITH DISCRETE FOURIER TRANSFORM SPREAD ORTHOGONAL FREQUENCY DIVISION MULTIPLEXING (DFT-S-OFDM) WAVEFORM
This disclosure provides systems, devices, apparatus, and methods, including computer programs encoded on storage media, for a non-codebook based transmission with a DFT-s-OFDM waveform. A UE receives (710), from a network entity (104), a configuration indicating an SRS resource set. The UE (120) transmits (730), to the network entity (104), a plurality of SRSs via the SRS resource set. The UE (102) receives (740), from the network entity (104), an uplink grant scheduling a PUSCH based on a transform precoding enabled OFDM waveform, the uplink grant indicating an SRS resource indicator (SRI) associated with the SRS resource set. The UE (102) transmits (750), to the network entity (104) based on a first precoder indicated by the SRI, the PUSCH using the transform precoding enabled OFDM waveform, the first precoder comprising a rank of at least two layers.
H04B 7/0404 - Systèmes de diversitéSystèmes à plusieurs antennes, c.-à-d. émission ou réception utilisant plusieurs antennes utilisant plusieurs antennes indépendantes espacées la station mobile comprenant plusieurs antennes, p. ex. pour mettre en œuvre une diversité en voie ascendante
A fabrication process uses casting to form portions of a waveguide having ultra-flat surfaces. A casting resin is coated between a prism mold and a top flat mold via inkjet, slot die, spray coating, etc. The top flat mold is lowered to conform with the casting resin and the casting resin is then cured to form a bottom prism array. After curing, the bottom prism array is demolded from the prism mold and the top flat mold is used as a carrier wafer to support the bottom prism array. The bottom prism array is selectively coated with a reflective coating and a second casting process is performed by coating the bottom prism array with casting resin to form a reflective waveguide.
The present disclosure provides techniques to reduce or eliminate contaminants in a waveguide resulting from the waveguide fabrication process. A blocking layer is deposited on a grating layer to protect the grating layer from contamination, e.g., via diffusion, from a hardmask layer that is used to pattern the grating layer. Accordingly, the optical gratings resulting from the patterning of the grating layer have little or no contamination from the hardmask layer, thereby increasing the optical performance of the final waveguide product.
G02B 6/12 - Guides de lumièreDétails de structure de dispositions comprenant des guides de lumière et d'autres éléments optiques, p. ex. des moyens de couplage du type guide d'ondes optiques du genre à circuit intégré
G02B 6/34 - Moyens de couplage optique utilisant des prismes ou des réseaux
A method for deleting obsolete files from a file system is provided. The method includes receiving a request to delete a reference to a first target file of a plurality of target files stored in a file system, the first target file having a first target file name. A first reference file whose file name includes the first target file name is identified. The first reference file is deleted from the file system. The method further includes determining whether the file system includes at least one reference file, distinct from the first reference file, whose file name includes the first target file name. In accordance with a determination that the file system does not include the at least one reference file, the first target file is deleted from the file system.
G06F 16/215 - Amélioration de la qualité des donnéesNettoyage des données, p. ex. déduplication, suppression des entrées non valides ou correction des erreurs typographiques
52.
METHODS, SYSTEMS, AND MEDIA FOR IDENTIFYING A PLURALITY OF SETS OF COORDINATES FOR A PLURALITY OF DEVICES
Methods, systems, and media for identifying a plurality of sets of coordinates for a plurality of devices are provided. In some embodiments, the method comprises: identifying each device in a plurality of devices associated with a user account; instructing the plurality of devices to perform an audio sequence; receiving a plurality of transit times from the plurality of devices; determining a plurality of distances based on the plurality of transit times; determining a plurality of sets of coordinates based on the plurality of distances; associating to each of the plurality of devices a corresponding unique one of the plurality of sets of coordinates; and causing at least one of the plurality of devices to play spatial audio determined from the plurality of sets of coordinates.
The technology described herein is directed to artificial intelligence (AI) powered tools that can generate, enhance, and evaluate digital imagery. For example, the AI-powered tools can be used to generate personalized digital illustrations based on user profile information. In some examples, the tools can modify the personalized digital illustrations, such as by modifying a facial expression of a person depicted in the personalized digital illustration to correspond to a mood or tone of the illustration.
The present disclosure provides computer-implemented methods, systems, and devices for capturing spatial sound for an environment. A computing system captures, using two or more microphones, audio data from an environment around a mobile device. The computing system analyzes the audio data to identify a plurality of sound sources in the environment around the mobile device based on the audio data. The computing system determines, based on characteristics of the audio data and data produced by one or more movement sensors, an estimated location for each respective sound source in the plurality of sound sources. The computing system generates a spatial sound recording of the audio data based, at least in part, on the estimated location of each respective sound source in the plurality of sound sources.
H04R 1/40 - Dispositions pour obtenir la fréquence désirée ou les caractéristiques directionnelles pour obtenir la caractéristique directionnelle désirée uniquement en combinant plusieurs transducteurs identiques
55.
System And Method For Identifying Places Using Contextual Information
The present disclosure provides a computing device and method for providing personal specific information based on semantic queries. The semantic queries may be input in a natural language form, and may include user specific context, such as by referring to prior or future events related to a place the user is searching for. With the user's authorization, data associated with prior or planned activities of the user may be accessed and information from the accessed data may be identified, wherein the information is correlated with the user specific context. One or more query results are determined based on the identified information and provided for output to the user.
A central node (CN) and a radio access network (RAN) can implement a method for managing paging for multicast and broadcast services (MBS). The method includes: receiving, from an MBS network, an identifier for an MBS session; transmitting, to a RAN, a message including the identifier for the MBS session; transmitting, to the RAN, one or more parameters associated with the MBS session; and transmitting, to the RAN, one or more MBS data packets to be broadcast to a user equipment (UE) in accordance with the one or more parameters.
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for privacy-sensitive training of a neural network. In one aspect, a system comprises a central memory configured to store current values of a set of neural network parameters and one or more computers that are configured to implement a plurality of worker computing units, where each worker computing unit is configured to repeatedly perform operations comprising obtaining current values of the set of neural network parameters from the central memory, sampling a batch of network inputs from a set of training data, determining a respective gradient corresponding to each network input, determining an aggregated gradient based on the gradients, identifying a subset of a set of gradient values as target values, generating a noisy gradient by combining random noise with the target gradient values, and updating the current values of the set of neural network parameters.
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for determining an architecture for a task neural network that is configured to perform a particular machine learning task on a target set of hardware resources. When deployed on a target set of hardware, such as a collection of datacenter accelerators, the task neural network may be capable of performing the particular machine learning task with enhanced accuracy and speed.
FINE-TUNING GENERATIVE MODEL UTILIZING INSTANCES AUTOMATICALLY GENERATED FROM LESS COMPUTATIONALLY EFFICIENT DECODING AND SUBSEQUENT UTILIZATION THEREOF WITH MORE COMPUTATIONALLY EFFICIENT DECODING
Implementations disclose utilizing a less computationally efficient decoding method in automatically generating corresponding single generative content predictions for training instances and fine-tuning a student generative model based on those automatically generated training instances. Those implementations are further directed to then utilizing, in an inference time environment, the fine-tuned student generative model and a more computationally efficient decoding method in generating generative predictions—and without any utilization of the less computationally efficient decoding method in generating the generative predictions.
The present disclosure is directed to encoding images. In particular, one or more computing devices can receive data representing one or more machine learning (ML) models configured, at least in part, to encode images comprising objects of a particular type. The computing device(s) can receive data representing an image comprising one or more objects of the particular type. The computing device(s) can generate, based at least in part on the data representing the image and the data representing the ML model(s), data representing an encoded version of the image that alters at least a portion of the image comprising the object(s) such that when the encoded version of the image is decoded, the object(s) are unrecognizable as being of the particular type by one or more object-recognition ML models based at least in part upon which the ML model(s) configured to encode the images were trained.
G06V 10/94 - Architectures logicielles ou matérielles spécialement adaptées à la compréhension d’images ou de vidéos
G06V 10/774 - Génération d'ensembles de motifs de formationTraitement des caractéristiques d’images ou de vidéos dans les espaces de caractéristiquesDispositions pour la reconnaissance ou la compréhension d’images ou de vidéos utilisant la reconnaissance de formes ou l’apprentissage automatique utilisant l’intégration et la réduction de données, p. ex. analyse en composantes principales [PCA] ou analyse en composantes indépendantes [ ICA] ou cartes auto-organisatrices [SOM]Séparation aveugle de source méthodes de Bootstrap, p. ex. "bagging” ou “boosting”
G06V 10/82 - Dispositions pour la reconnaissance ou la compréhension d’images ou de vidéos utilisant la reconnaissance de formes ou l’apprentissage automatique utilisant les réseaux neuronaux
61.
MULTI-CHANNEL VIDEO RATE CONTROL FOR EXTENDED REALITY STREAMING
A cloud-based extended reality (XR) system includes a server configured to encode a set of frames each associated with an XR scene to be displayed. To encode the set of frames, the server estimates a total number of encoded output bits for the set of frames based on a set of quantization parameters (QPs). The set of QPs includes a corresponding QP for each frame of the set of frames and one or more predetermined relationships between the corresponding QPs. The server then compares the estimated total number of encoded output bits to a target frame size threshold. Based on the estimated total number of encoded bits being outside the target frame size threshold, the server updates the set of QPs so as to maintain the predetermined relationships between the QPs.
This document describes methods and systems of adaptive fingerprint-enrollment to finger characteristics using local under-display fingerprint sensors, UDFPS, in an electronic device. The electronic device includes an adaptive-enrollment module that determines characteristics of a fingerprint based on information corresponding to a touch input detected by a touch-display device, including size and shape of an area of the touch input. Based on the fingerprint characteristics, a number and location of enrollment touches used for completing enrollment of the fingerprint are adjusted to minimize the number of enrollment touches required to complete the enrollment, minimize the amount of time needed to complete the enrollment, and maximize coverage of the fingerprint. The adaptive-enrollment module also provides visual guidance to guide the user to touch the adjusted locations of the enrollment touches and, if needed, feedback to instruct the user to adjust the location of their finger to align with the visual guidance.
A method includes receiving training data that includes a plurality of sets of training utterances each associated with a respective language. Each training utterance includes a corresponding reference speech representation paired with a corresponding input text sequence. For each training utterance, the method includes generating a corresponding encoded textual representation for the corresponding input text sequence, generating a corresponding speech encoding for the corresponding reference speech representation, generating a shared encoder output, and determining a text-to-speech (TTS) loss based on the corresponding encoded textual representation, the corresponding speech encoding, and the shared encoder output. The method also includes training a TTS model based on the TTS losses determined for the training utterances in each set of the training utterances to teach the TTS model to learn how to synthesize speech in each of the respective languages.
G10L 13/02 - Procédés d'élaboration de parole synthétiqueSynthétiseurs de parole
G10L 25/30 - Techniques d'analyse de la parole ou de la voix qui ne se limitent pas à un seul des groupes caractérisées par la technique d’analyse utilisant des réseaux neuronaux
64.
Systems and Methods for Receiving Incoming Messages Via Satellite
A computer-implemented method is provided. The method includes determining, by a user device, that the user device is within range of a particular satellite. The method also includes sending, by the user device, a connection request to the particular satellite in response to determining that the user device is within range of the particular satellite. The connection request is forwarded to a network device, via the particular satellite, to indicate to the network device that the user device is available to receive messages via the particular satellite. The method also includes receiving, by the user device and via the particular satellite, a message from the network device in response to sending the connection request.
This document describes systems for and techniques of alternating-current (AC) power harmonic-based circuit state detection. In various aspects, a system includes a component, a bypass circuit for the component, and a controller with an AC power harmonic-based circuit state detector that can determine a state of the bypass circuit. The AC power harmonic-based circuit state detector may convert an AC voltage of the AC power to a direct current (DC) voltage, filter the DC voltage to obtain a voltage of a harmonic of the AC power, and compare the voltage of the harmonic to a threshold to determine that the bypass circuit is in a fault state (blown fuse). By so doing, the controller of the system can notify a user that the bypass circuit needs to be reset or replaced to reenable operation of the system and avoid poor user experience typically associated with a non- or mis-functioning system.
H02J 3/01 - Dispositions pour réduire les harmoniques ou les ondulations
H02J 1/00 - Circuits pour réseaux principaux ou de distribution, à courant continu
H02J 3/00 - Circuits pour réseaux principaux ou de distribution, à courant alternatif
H02J 3/14 - Circuits pour réseaux principaux ou de distribution, à courant alternatif pour règler la tension dans des réseaux à courant alternatif par changement d'une caractéristique de la charge du réseau par interruption, ou mise en circuit, des charges du réseau, p. ex. charge équilibrée progressivement
H04N 7/18 - Systèmes de télévision en circuit fermé [CCTV], c.-à-d. systèmes dans lesquels le signal vidéo n'est pas diffusé
66.
SECURE MULTI-PARTY COMPUTATION WITH ATTESTATION USING A TRUSTED EXECUTION ENVIRONMENT
Disclosed herein are systems, methods, and computer-readable media for enabling more secure multi-party computations (MPCs) using a trusted execution environment (TEE). In one aspect, a method includes executing, by a first MPC computer, a secure MPC protocol in a first TEE of the first MPC computer. The first MPC computer generates a request to a second MPC computer executing the secure MPC protocol in a second TEE of the second MPC computer. The first TEE determines that one or more attestation conditions are met by the first MPC computer executing the secure MPC protocol in the first TEE. In response to determining that the one or more attestation conditions are met, the first TEE generates an attestation token including one or more digital signatures for the secure MPC protocol executing in the first TEE. The first MPC computer sends the attestation token with the request to the second MPC computer.
H04L 9/32 - Dispositions pour les communications secrètes ou protégéesProtocoles réseaux de sécurité comprenant des moyens pour vérifier l'identité ou l'autorisation d'un utilisateur du système
G06F 21/53 - Contrôle des utilisateurs, des programmes ou des dispositifs de préservation de l’intégrité des plates-formes, p. ex. des processeurs, des micrologiciels ou des systèmes d’exploitation au stade de l’exécution du programme, p. ex. intégrité de la pile, débordement de tampon ou prévention d'effacement involontaire de données par exécution dans un environnement restreint, p. ex. "boîte à sable" ou machine virtuelle sécurisée
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for evicting cache data using machine learning. One of the methods includes determining that particular data is not stored in a cache that is full; determining, using information for the particular data, a predicted eviction accuracy of a machine learning system; determining whether the predicted eviction accuracy of the machine learning system satisfies a threshold eviction accuracy; and in response to determining that the predicted eviction accuracy of the machine learning system satisfies the threshold eviction accuracy: sending, to the machine learning system, a request for an identifier for data stored in the cache; receiving, from the machine learning system, an identifier for data stored in the cache; evicting the data referenced by identifier from a location in the cache; and storing the particular data at the location in the cache.
G06F 12/121 - Commande de remplacement utilisant des algorithmes de remplacement
G06F 12/127 - Commande de remplacement utilisant des algorithmes de remplacement avec maniement spécial des données, p. ex. priorité des données ou des instructions, erreurs de maniement ou repérage utilisant des algorithmes de remplacement supplémentaires
G06N 3/044 - Réseaux récurrents, p. ex. réseaux de Hopfield
A method includes training, using an un-supervised learning technique, an auxiliary ASR model based on a first set of un-transcribed source task speech utterances to determine a first task vector, training, using the un-supervised learning technique, the auxiliary ASR model based on a second set of un-transcribed speech utterances to determine a second task vector, and training, using the un-supervised learning technique, the auxiliary ASR model based on un-transcribed target task speech utterances to determine a target task vector. The method also includes determining a first correlation between the first and target task vectors, determining a second correlation between the second and target task vectors, and adapting parameters of a trained primary ASR model based on the first and second source task vectors and the first and second correlations to teach the primary ASR model to learn how to recognize speech associated with the target task.
G10L 15/06 - Création de gabarits de référenceEntraînement des systèmes de reconnaissance de la parole, p. ex. adaptation aux caractéristiques de la voix du locuteur
An example folding device includes a hinge assembly that is coplanar with the continuous display of the device in order to decrease the thickness of the device. The hinge assembly includes torque members that increase the amount of force needed to rotate the assemblies. In this way, the torque members may provide the device with a more rigid feel. Also in this way, the torque members may enable the device to hold intermediate positions between fully open and fully closed.
Aspects of the technology provide a symbiotic graphical display on a client device such as a smartwatch. The system includes at least one emissive display element and at least one non-emissive display element. The display elements are arrayed in layers or other configurations such that content or other information is concurrently aligned across the respective display surfaces of the different elements. A first set of content is rendered using the non-emissive display element while a second set of content is rendered using the emissive display element. Depending on characteristics or features of a given content item, that item may be rendered by one or both of the display elements. Certain content may be transitioned from the emissive display element to the non-emissive display element according to a time threshold or other criteria.
G04C 17/00 - Indication optique du temps par des moyens électriques
G04C 3/00 - Horloges ou montres électromécaniques indépendantes d'autres garde-temps et dans lesquelles le mouvement est entretenu par des moyens électriques
G04G 9/00 - Moyens visuels d'indication de l'heure ou de la date
This document describes techniques and apparatuses for computational photography under low-light conditions for an image-capture device on a mobile computing device. In aspects, described are techniques and apparatuses for an image-capture device to utilize sensor data in determining whether to enable flash photography or capture multiple images of the scene without use of a flash under low-light conditions. In other aspects, an image-capture device may utilize device data in determining whether to enable flash photography or capture multiple images of the scene without use of a flash under low-light conditions. The disclosed techniques and apparatuses may provide improved computational photography under low-light conditions for an image-capture device on a mobile computing device.
To provide navigation directions in response to a request for viewing a particular type of geographical feature on a route to a destination, a computing device receives a request for navigation directions from a starting location to a destination that specifies a particular type of geographical feature for viewing on a route to the destination. The computing device identifies at least one candidate route for navigating to the destination location that includes a road segment from which the particular type of geographical feature can be viewed, and selects a route from the set of candidate routes based at least in part on an extent to which the particular type of geographical feature can be viewed from each candidate route. Then the computing device provides a set of navigation directions for presentation on a client device for navigating to the destination location via the selected route.
Methods, systems, and apparatuses, including computer programs encoded on computer storage media, for configuring a set of language model neural networks, e.g., a first large language model and a second smaller-sized language model, and performing a machine learning task on new inputs using the set of language model neural networks. Configuring the language model neural networks and performing a machine learning task can include leveraging the ability of a first large language model to follow prompt-engineered instructions and perform chain-of-thought reasoning, while also fine-tuning a second, smaller language model neural network to optimize the machine learning task performance.
An example computing device includes a flexible display coupled to a housing that includes a support plate having a first joint coupled to a first end of the support plate and a second joint coupled to a second end of the support plate. A slide module has a slot that guides a slide movement of the second joint along a path of movement within the slot as the support plate pivots about the first joint, where the support plate moves according to the first joint and the second joint to support at least the portion of the flexible display when the flexible display is unfolded and moves according to the first joint and the second joint to create a gap between at least a portion of the support plate and at least the portion of the flexible display when the flexible display is folded.
Implementations relate to generating and using multimodal embeddings. In various implementations, first modality data may be obtained and encoded into first modality embedding(s) using a trained first modality encoder that is stored in memory of edge-based client device(s). Second modality data may be obtained and encoded into second modality embedding(s) using a trained second modality encoder that is also stored in the memory of the edge-based client device(s). The first and second modality embeddings may be processed using an edge-based multimodal LLM that is also stored locally in memory of the edge-based client device(s) to generate a multimodal contextual embedding, which may be provided to a remote server that hosts a central LLM, e.g., in conjunction with a natural language input provided by the user. Information generated using the central LLM, responsive to the natural language input, may be received from the remote server.
G06V 10/80 - Fusion, c.-à-d. combinaison des données de diverses sources au niveau du capteur, du prétraitement, de l’extraction des caractéristiques ou de la classification
G06V 40/16 - Visages humains, p. ex. parties du visage, croquis ou expressions
G10L 15/183 - Classement ou recherche de la parole utilisant une modélisation du langage naturel selon les contextes, p. ex. modèles de langage
G10L 15/30 - Reconnaissance distribuée, p. ex. dans les systèmes client-serveur, pour les applications en téléphonie mobile ou réseaux
The present disclosure is generally directed to methods, apparatus, and computer-readable media (transitory and non-transitory) for learning to automatically navigate interactive web documents and/or websites. More particularly, various approaches are presented for training various deep Q network (DQN) agents to perform various tasks associated with reinforcement learning, including hierarchical reinforcement learning, in challenging web navigation environments with sparse rewards and large state and action spaces. These agents include a web navigation agent that can use learned value function(s) to automatically navigate through interactive web documents, as well as a training agent, referred to herein as a “meta-trainer,” that can be trained to generate synthetic training examples. Some approaches described herein may be implemented when expert demonstrations are available. Other approaches described herein may be implemented when expert demonstrations are not available. In either case, dense, potential-based rewards may be used to augment the training.
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for enabling artificial intelligence to generate new images based on contextual data and to generate digital components based on the images. In one aspect, a method includes receiving one or more queries from a client device of a user. A digital component is selected based on the one or more queries. A customized digital component is generated by obtaining an image of an object corresponding to the selected digital component and generating, using a language model, an image editing prompt for editing the image based on digital component data related to the digital component and query data including the one or more queries and contextual data. The image and the image editing prompt are provided to an image editing model. An edited image is received and used to generate the customized digital component.
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating speech from text. One of the systems includes one or more computers and one or more storage devices storing instructions that when executed by one or more computers cause the one or more computers to implement: a sequence-to-sequence recurrent neural network configured to: receive a sequence of characters in a particular natural language, and process the sequence of characters to generate a spectrogram of a verbal utterance of the sequence of characters in the particular natural language; and a subsystem configured to: receive the sequence of characters in the particular natural language, and provide the sequence of characters as input to the sequence-to-sequence recurrent neural network to obtain as output the spectrogram of the verbal utterance of the sequence of characters in the particular natural language.
G10L 13/08 - Analyse de texte ou génération de paramètres pour la synthèse de la parole à partir de texte, p. ex. conversion graphème-phonème, génération de prosodie ou détermination de l'intonation ou de l'accent tonique
G06N 3/084 - Rétropropagation, p. ex. suivant l’algorithme du gradient
G10L 13/04 - Détails des systèmes de synthèse de la parole, p. ex. structure du synthétiseur ou gestion de la mémoire
G10L 15/16 - Classement ou recherche de la parole utilisant des réseaux neuronaux artificiels
G10L 25/18 - Techniques d'analyse de la parole ou de la voix qui ne se limitent pas à un seul des groupes caractérisées par le type de paramètres extraits les paramètres extraits étant l’information spectrale de chaque sous-bande
G10L 25/30 - Techniques d'analyse de la parole ou de la voix qui ne se limitent pas à un seul des groupes caractérisées par la technique d’analyse utilisant des réseaux neuronaux
79.
Two-Level Text-To-Speech Systems Using Synthetic Training Data
A method includes obtaining training data including a plurality of training audio signals and corresponding transcripts. Each training audio signal is spoken by a target speaker in a first accent/dialect. For each training audio signal of the training data, the method includes generating a training synthesized speech representation spoken by the target speaker in a second accent/dialect different than the first accent/dialect and training a text-to-speech (TTS) system based on the corresponding transcript and the training synthesized speech representation. The method also includes receiving an input text utterance to be synthesized into speech in the second accent/dialect. The method also includes obtaining conditioning inputs that include a speaker embedding and an accent/dialect identifier that identifies the second accent/dialect. The method also includes generating an output audio waveform corresponding to a synthesized speech representation of the input text sequence that clones the voice of the target speaker in the second accent/dialect.
G10L 13/08 - Analyse de texte ou génération de paramètres pour la synthèse de la parole à partir de texte, p. ex. conversion graphème-phonème, génération de prosodie ou détermination de l'intonation ou de l'accent tonique
G10L 13/047 - Architecture des synthétiseurs de parole
80.
Systems and Methods for Anonymizing Large Scale Datasets
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.
A non-planar lightguide directs a display light from an incoupler surface towards an eye of a user via a reduced number of internal reflective interactions with a world-facing surface of the non-planar lightguide and an eye-facing lens surface of the non-planar lightguide.
A method (500) includes receiving audio data (120) corresponding to an utterance (106) spoken by a particular user (10) and captured in streaming audio (118) by a user device (102). The method also includes performing speaker identification on the audio data to identify an identity (205) of the particular user that spoke the utterance. The method also includes obtaining a keyword detection model (420) personalized for the particular user based on the identity of the particular user that spoke the utterance. The keyword detection model is conditioned on speaker characteristic information (250) associated with the particular user to adapt the keyword detection model to detect a presence of a keyword in audio for the particular user. The method also includes determining that the utterance includes the keyword using the keyword detection model personalized for the particular user.
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a visualization of a story. In one aspect, a method comprises receiving an input from a user of a user device that comprises text representing a story to be visualized through a plurality of images; in response to receiving the input: obtaining a plurality of sets of words from the text; for each set of words in the plurality of sets of words: generating an input prompt for the set of words; obtaining an image that represents the input prompt by providing the input prompt to a first machine learning model; and providing, for presentation on the user device, the plurality of images generated by the first machine learning model that visualize the story.
G06F 3/04845 - Techniques d’interaction fondées sur les interfaces utilisateur graphiques [GUI] pour la commande de fonctions ou d’opérations spécifiques, p. ex. sélection ou transformation d’un objet, d’une image ou d’un élément de texte affiché, détermination d’une valeur de paramètre ou sélection d’une plage de valeurs pour la transformation d’images, p. ex. glissement, rotation, agrandissement ou changement de couleur
Multi-speaker-selective protection is described for smart audio amplifier systems having multiple speakers all concurrently playing substantially the same audio. Current-voltage (IV) sense signals from all of the speakers are received by a smart selector. The smart selector identifies one of the IV sense signals as a reference and computes IV deltas between the identified reference and the others of the IV sense signals. A controlling sense signal is selected as the one of the IV sense signals manifesting the largest IV delta. Speaker protection can be computed based on the controlling sense signal and applied to all the speakers based on the single set of speaker protection computations. In some embodiments, speakers are grouped, a respective controlling sense signal is selected for each group, and speaker protection is computed for and applied to each group based on its respective controlling sense signal.
According to an aspect, a method includes identifying frames of video data, identifying background portions in a subset of the frames, and determining an composite background based on the background portions for the subset. The method further includes determining at least one noise parameter for the video data based on the composite background and a model. The method also includes updating the video data based on the at least one noise parameter and the model.
Aspects of the disclosure are directed to merging data lake openness with scalable metadata for managed tables in a cloud database platform, allowing for atomicity, consistency, isolation, and durability (ACID) transactions, performant data manipulation language (DML), higher throughput stream ingestion, data consistency, schema evolution, time travel, clustering, fine-grained security, and/or automatic storage optimization. Table data is stored in various open-source file formats in cloud storage while physical metadata of the table data is stored in a scalable metadata storage system.
A user equipment (UE) device (102) in a mobile cellular network (100) implements one or more techniques to perform early retransmissions of session initiation protocol (SIP) non-INVITE messages (401). For example, the UE device receives a first indication (405) that a current radio access technology (RAT) (108) at the UE device has changed from a first RAT to a second RAT. Responsive to receiving the first indication, the UE device retransmits a SIP non-INVITE message (401) that is pending a response from an Internet Protocol Multimedia Subsystem (IMS) network (126) without waiting for a specification defined retransmission interval to expire.
H04L 65/1104 - Protocole d'initiation de session [SIP]
H04L 65/80 - Dispositions, protocoles ou services dans les réseaux de communication de paquets de données pour prendre en charge les applications en temps réel en répondant à la qualité des services [QoS]
H04L 69/40 - Dispositions, protocoles ou services de réseau indépendants de la charge utile de l'application et non couverts dans un des autres groupes de la présente sous-classe pour se remettre d'une défaillance d'une instance de protocole ou d'une entité, p. ex. protocoles de redondance de service, état de redondance de protocole ou redirection de service de protocole
H04W 36/14 - Resélection d'un réseau ou d'une interface hertzienne
An application may receive a prompt from a user related to an input for a text field of digital content displayed on a user device. An application may generate context data about the digital content. An application may provide the prompt and the context data to a generative language model. An application may receive a response generated by the generative language model and provide the response as a suggestion for the input for the text field.
Two components can transfer data across a communication fabric (110), but the data transfer consumes power and time. To save power or time, a compression technique avoids repeatedly making data transfers across the fabric. In example implementations, for an initiator side, a transmission storage unit (204-1) stores one or more values (304-1…304-N) that were previously transmitted to a target (114). Compression logic circuitry (202-1) compares a first value (302) of a request (306) with the stored values. Responsive to the first value matching a stored value, the transmission logic circuitry prevents the first value from being transmitted to the target and instead transmits an indication (122) that the stored value is to be received by the target. For a target side, a reception storage unit (204-2) stores the values (304-1…304-N) that were previously received from an initiator (112). Decompression logic circuitry (202-2) receives the indication and provides the stored value (304) to the target from the reception storage unit using the indication.
An intermediate representation can be obtained of audio data comprising audio of a particular type, wherein audio of the particular type comprises a plurality of audio content attributes. The intermediate representation can be processed with a machine-learned disentanglement model to obtain two or more token streams for two or more respective audio content attributes of the plurality of audio content attributes. The two or more token streams include a first token stream for a first audio content attribute of the two or more audio content attributes, wherein the first token stream is tokenized at a first time scale, and a second token stream for a second audio content attribute of the two or more audio content attributes, wherein the second token stream is tokenized at a second time scale different than the first time scale.
G10L 25/30 - Techniques d'analyse de la parole ou de la voix qui ne se limitent pas à un seul des groupes caractérisées par la technique d’analyse utilisant des réseaux neuronaux
G10L 19/20 - Vocodeurs utilisant des modes multiples utilisant un codage spécifique de la catégorie de son, des encodeurs hybrides ou un codage basé objet
G10L 25/51 - Techniques d'analyse de la parole ou de la voix qui ne se limitent pas à un seul des groupes spécialement adaptées pour un usage particulier pour comparaison ou différentiation
G10L 21/003 - Changement de la qualité de la voix, p. ex. de la hauteur tonale ou des formants
G10L 25/63 - Techniques d'analyse de la parole ou de la voix qui ne se limitent pas à un seul des groupes spécialement adaptées pour un usage particulier pour comparaison ou différentiation pour estimer un état émotionnel
91.
MACHINE-LEARNED OUTPUT USING INTERLEAVED UNIMODAL MODELS
Methods and systems for generating an output from a machine-learned model are disclosed herein. The method includes receiving a multimodal input and providing a first input portion from the multimodal input to a first unimodal model. The method also includes providing a second input portion from the multimodal input to a second unimodal model and providing a first representation from the first unimodal model to the second unimodal model. The method further includes providing a second representation from the second unimodal model to the first unimodal model, and processing the second input portion using the second unimodal model based in part on the first representation. The method also includes processing the first input portion using the first unimodal model based in part on the second representation and generating an output from the machine-learned model based in part on the processing by the first and second unimodal models.
G06N 3/0442 - Réseaux récurrents, p. ex. réseaux de Hopfield caractérisés par la présence de mémoire ou de portes, p. ex. mémoire longue à court terme [LSTM] ou unités récurrentes à porte [GRU]
G06N 3/063 - Réalisation physique, c.-à-d. mise en œuvre matérielle de réseaux neuronaux, de neurones ou de parties de neurone utilisant des moyens électroniques
G06N 3/084 - Rétropropagation, p. ex. suivant l’algorithme du gradient
An example mobile computing device includes a housing; a battery; one or more electronic components; and a chassis assembly removably connected to the housing, wherein the battery is connected to the chassis assembly and the chassis assembly provides a heat path from the one or more electronic components to the battery.
A method includes obtaining a three-dimensional (3D) object comprising a first plurality of polygons that form a mesh and generating a first set of lighting values associated with the 3D object. The lighting values of the first set of lighting values are associated with one or more simulated lighting conditions used to illuminate the 3D object. One or more polygons to be removed are identified from the mesh. A compressed version of the 3D object is generated by removing the identified one or more polygons from the mesh. The compressed version comprises a second plurality of polygons. A second set of lighting values associated with the initial compressed version of the 3D object is generated. A difference between the first set of lighting values and the second set of lighting values is determined and an output compressed version of the 3D object is generated based upon the determined difference.
Projected motion field hole filling includes determining a motion field for the current frame using a proper subset of blocks of the current frame, the motion field comprising a respective motion vector projecting onto a reference frame for each block of the proper subset of blocks. For a block of the current frame without a motion vector in the motion field, the respective motion vector for a spatially neighboring block of the proper subset of blocks is reused as a projected motion vector for the block within the motion field. A motion vector candidate list may be determined using the motion field. A reference motion vector for the current block may be selected from the motion vector candidate list. The reference motion vector may be used to encode the current block into an encoded bitstream or decode the current block from an encoded bitstream.
H04N 19/52 - Traitement de vecteurs de mouvement par encodage par encodage prédictif
H04N 19/105 - Sélection de l’unité de référence pour la prédiction dans un mode de codage ou de prédiction choisi, p. ex. choix adaptatif de la position et du nombre de pixels utilisés pour la prédiction
H04N 19/176 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques utilisant le codage adaptatif caractérisés par l’unité de codage, c.-à-d. la partie structurelle ou sémantique du signal vidéo étant l’objet ou le sujet du codage adaptatif l’unité étant une zone de l'image, p. ex. un objet la zone étant un bloc, p. ex. un macrobloc
H04N 19/70 - Procédés ou dispositions pour le codage, le décodage, la compression ou la décompression de signaux vidéo numériques caractérisés par des aspects de syntaxe liés au codage vidéo, p. ex. liés aux standards de compression
A system and method for protocol independent receive side scaling (RSS) includes storing a plurality of RSS hash M-tuple definitions, each definition corresponding to one of a set of possible protocol header combinations for routing an incoming packet, the set of possible protocol header combinations being modifiable to include later-developed protocols. Based on initial bytes of the incoming packet, a pattern of protocol headers is detected, and used to select one of the plurality of RSS hash M-tuple definitions. The selected RSS hash M-tuple definition is applied as a protocol-independent arbitrary set of bits to the headers of the incoming packet to form a RSS hash M-tuple vector, which is used to compute a RSS hash. Based on the RSS hash, a particular queue is selected from a set of destination queues identified for the packet, and the packet is delivered to the selected particular queue.
The present disclosure provides for a heatshield that can be actively cooled during a rework process. The heatshield may include a backer plate and a metal plate. A plurality of vents may extend from air inlet ducts to a top surface of the backer plate such that the plurality of vents directs cooling gas forced into the heatshield towards the metal plate and a first ball grid array (BGA) package. The cooling gas may maintain the solder joint temperature of the first BGA package below a reflow temperature and below a solidus temperature to prevent reflow-related solder joint defects from occurring in the first BGA package during rework of a second BGA package.
H01L 23/467 - Dispositions pour le refroidissement, le chauffage, la ventilation ou la compensation de la température impliquant le transfert de chaleur par des fluides en circulation par une circulation de gaz, p. ex. d'air
H01L 23/498 - Connexions électriques sur des substrats isolants
