Abstract: A clock selection circuit allows seamless switching between different clock signals in a clock distribution network. The clock selection circuit can be an Integrated Circuit (IC). The clock signals can be analyzed by a processor in communication with the IC to ensure the clock signals are validated. Analysis can include comparing time stamps between received pulses of the clock signals to determine if the clock signals are occurring at regular intervals. The processor can then assign a priority order to the clock signals and select one of the clock signals to use. An identifier associated with the selected clock signal can be programmed into the IC. The IC can then redistribute the selected clock signal to multiple other ICs in a hierarchical clock distribution network. Ultimately, the distributed clock signal can be received by server computers to ensure instances being executed have accurate and synchronized timing.

Abstract: A federated permission management service provides clients with customized access to a data set using customized authorization metadata. The federated permission management service may define and apply permissions that are defined at a data lake that provides access to many different data sets from many different sources, as well as those permissions that may be defined at the source of the data set, which may be provided when performing a data sharing request. By allowing for permissions to be specified at the data lake in addition to permissions specified at a source of a data set, the permission management service can provide a fine-grained access control to specific objects of the data set, such as specific columns, specific rows, or specific cells of a database to be shared, even for those data sets in the data lake having different sources.

Abstract: Described are systems and methods for active weathervaning of a hybrid flight aerial vehicle, such as an unmanned aerial vehicle (UAV). Active weathervaning of the hybrid flight aerial vehicle during can be provided during vertical takeoff and landing (VTOL)/hover flight without the assistance of any low-speed wind sensors and during transitions between VTOL/hover flight and fixed-wing, wing-borne, horizontal flight. Additionally, active weathervaning can be provided during propulsion mechanism failure conditions where the aerial vehicle may be experiencing failure conditions associated with one or more propulsion mechanisms.

Abstract: Technology is described for receiving an event notification produced by a source service. A negative cache may be searched for a cache entry for the event notification. The negative cache may be determined to not include the cache entry for the event notification. The event notification may be determined to satisfy a filter rule included in a filter rules table, which may indicate that the event notification is useful information for a destination service. The event notification may be transmitted to the destination service.

Abstract: A voice controlled assistant has a housing to hold one or more microphones, one or more speakers, and various computing components. The housing has an elongated cylindrical body extending along a center axis between a base end and a top end. The microphone(s) are mounted in the top end and the speaker(s) are mounted proximal to the base end. The microphone(s) and speaker(s) are coaxially aligned along the center axis. The speaker(s) are oriented to output sound directionally toward the base end and opposite to the microphone(s) in the top end. The sound may then be redirected in a radial outward direction from the center axis at the base end so that the sound is output symmetric to, and equidistance from, the microphone(s).

Abstract: Systems and methods are described herein for routing data by transferring a physical storage device for at least part of a route between source and destination locations. In one example, a computing resource service provider, may receive a request to transfer data from a customer center to a data center. The service provider may determine a route, which includes one or more of a physical path or a network path, for the data loaded onto a physical storage device to reach the data center from the customer center. Determining the route may include associating respective cost values to individual physical and network paths between physical stations between the customer and end data centers, and selecting one or more of the paths to reduce a total cost of the route. Route information may then be associated with the physical storage device based on the route.

Abstract: Systems and methods develop and apply one or more extractive summarization models for locating contact center conversation details in a transcript, extracting pertinent verbiage, and, in a transformation of the communication details, automatically generating summaries at one or more levels of abstraction, the summaries in full sentences, in a manner that a contact center agent understands. The models are trained using machine learning algorithms.

Abstract: Technologies directed to detecting and controlling different types of ZigBee devices on different ZigBee networks reliably and efficiently. In one method of operating a first device, the method includes detecting first and second Zigbee networks. The method sends a first request to rejoin the first Zigbee network, the first request being encoded with the first network key. While rejoined as part of the first Zigbee network, the method detects a first set of devices part of the first Zigbee network. The method repeats this for a second set of devices part of the second Zigbee network. The method determines, using the device data, a subset of controllable devices located in proximity, and sends a message with the command to each in response to receiving a command. Each message is encoded with a respective network key and a respective link key specified in the device data.

Abstract: Systems and methods are provided for use of use of fuzzy-match-based translation suggestions to augment machine translation of input sentences or other texts. A machine translation system may use a model trained to translate a source language input to a target language output based on pseudo-randomly selected translation suggestions in the target language, while at inference time the machine translation system may use translation selections associated with source language samples that have a high degree of similarity to the source language input to be translated. To efficiently use the translation suggestions, they may be encoded in context with the source language input to be translated, and the machine translation system may use the encoded translation suggestions with to generate a translation in the target language.

Abstract: Described is a system for counting stacked items using image analysis. In one implementation, an image of an inventory location with stacked items is obtained and processed to determine the number of items stacked at the inventory location. In some instances, the item closest to the camera that obtains the image may be the only item viewable in the image. Using image analysis, such as depth mapping or Histogram of Oriented Gradients (HOG) algorithms, the distance of the item from the camera and the shelf of the inventory location can be determined. Using this information, and known dimension information for the item, a count of the number of items stacked at an inventory location may be determined.

Abstract: Techniques for tracking and maintaining queues used for executing pending quantum objects using respective quantum processing units (QPUs) are disclosed. An amount of time to execute a given quantum object depends on many factors, and a non-deterministic nature of quantum computing resources is such that, while knowing an expected wait time in a queue for access to a given QPU is useful, it is difficult to reliably determine. A quantum computing service that manages submission and execution of quantum objects to respective QPUs may apply QPU-specific machine learning models in order to predict expected wait times and provide that information to customers. By generating labeled datasets using ground truth wait times pertaining to already-executed quantum objects, respective machine learning models may be trained using a supervised learning technique, which may be a self-contained and re-occurring process.

Abstract: An attestation service is configured to receive a request to enable attestation for a compute instance according to an attestation policy indicating one or more baseline health measurement values for validating compute instances. The attestation service provides a network endpoint for the compute instance to request attestation. The attestation service receives, via the network endpoint from a compute instance, one or more health measurement values of the compute instance. The attestation service validates the compute instance based at least on a comparison of the one or more current health measurement values and the one or more baseline health measurement values. The attestation service, in response to validating the compute instance, generates an attestation token indicating that the compute instance is authorized to access a secured resource of the provider network.

Abstract: Changes made to a database table are accumulated, in durable storage, and snapshots of partitions of the table are obtained. For successive snapshots of a partition, the system accesses a previous snapshot, applies changes from the accumulated changes, and stores the updated snapshot to a durable data store. The accumulated changes and the successive partition snapshots are made available to restore the database to any point in time across a continuum between successive snapshots. Although each partition of the table may have a backup snapshot that was generated at a time different from when other partition snapshots were generated, changes from respective change logs may be selectively log-applied to distinct partitions of a table to generate an on-demand backup of the entire table at common point-in-time across partitions. Point-in-time restores of a table may rely upon a similar process to coalesce partition snapshots that are not aligned in time.

Abstract: Approaches presented herein can allocate resources in such a way that sufficient capacity will be provided to perform a job or task, while minimizing any excess capacity included with those allocated resources. A number of jobs can be performed with differently sized resource instances in some embodiments, to determine an instance size, from a set of available sizes, that is appropriate for each of those jobs. Various parameters for those jobs can be determined, and those values associated with the determined instance sizes. When a new job is received that is to be performed, the parameter values for that job can be compared against the corresponding values for these testing jobs, and an instance size can be selected where a testing job that was successfully performed with that instance size had the same or larger values for these parameters.

Abstract: Techniques for performing voice-based user recognition are described. When a device receives audio data corresponding to a spoken user input, the device may generate spoken user input embedding data representing speech characteristics of the spoken user input. The device may also identify user embedding data representing speech characteristics of a user known to the device. A machine learning (ML) model may process the spoken user input embedding data to generate reduced spoken user input embedding data including a reduced number of dimensions, where the reduced number of dimensions are functions of the higher number of dimensions of the spoken user input embedding data.

Abstract: Systems and methods for automated actions for application policy violations are disclosed. For example, policy violation evaluation components may monitor requests and/or responses from one or more applications to identify content policy violations. When a violation is identified, an automated decision engine utilizes data representing the policy violation along with, in example, contextual information about the policy violation to identify a rule from a rules database that is associated with the policy violation. An action is determined from the selected rule, and a command is generated to perform the action in response to the policy violation.

Abstract: A system configured to detect custom acoustic events, where the system generates an acoustic event profile for the custom acoustic event based on a natural language description and without a sample of the sound. The system may generate a profile for a new custom sound based on a natural language description provided by the user; for example, a “microwave beep.” If the system does not have an existing event profile for a microwave beep, the system may ask the user questions to determine whether any existing event profiles are close (e.g., is the sound similar to a “fan,” “alarm,” “appliance beep,” etc.). The system may detect an event that may be a possible match for the custom sound and ask the user to verify whether the detected event corresponds to the custom sound. The system may update the event profile based on the user's response.

Abstract: The present disclosure generally relates to systems and methods for configuring Security Assertion Markup Language (SAML) access network-based services. An identity provider can authenticate the customer and provide SAML authentication information to a SAML configuration service. Based on the customer's authentication information, the SAML configuration service can discover SAML-enabled services and analyze the customer's usage pattern of each discovered service. The SAML configuration service may prioritize the discovered services based on the analysis and transmit a list of the discovered services to the customer based on a sequence of the prioritization. The SAML configuration service also can configure SAML configuration associated with each SAML-enabled service by identifying one or more parameters and applying parsing rules.

Abstract: Systems and methods for intelligent device grouping are disclosed. An environment, such as a home, may have a number of voice-enabled devices and accessory devices that may be controlled by the voice-enabled devices. One or more models, such as linguistics model(s) and/or device affinity models may be utilized to determine which accessory devices are candidates for inclusion in a device group, and a recommendation for grouping the devices may be provided. Device-group naming recommendations may also be generated and may be sent to users.