Abstract: Inferences may be obtained to handle access requests at a non-relational database system. An access request may be received at a non-relational database system. The non-relational database system may determine that the access request uses a machine learning model to complete the access request. The non-relational database system may cause an inference to be generated using data items for the access request as input to the machine learning model. The access request may be completed using the generated inference.

Abstract: A first network device wirelessly transmits announcement messages in quick succession over multiple channels that represent normal operating channels of other network devices. The other network devices receive one of the announcement messages over their respective operating channels, and then send a message to one or more servers indicating that one of the announcement messages has been received. Various parameters and information related to the received announcement message, such as RSSI, SNR, and the contents of the announcement message, are included in the message sent to the one or more servers by each of the other network devices. The one or more servers then incorporate information in the messages sent from the other network devices into a network connectivity graph or table.

Abstract: A page is generated for a media program based on an image associated with the media program, such as an image of a creator of the media program. A predominant color within the image is identified, and a first color that is proximate or similar to the predominant color is selected from a color palette. A second color is selected as a counterpart to the first color. Code for a page including a section in the first color and a section in the second color, as well as the image, is generated and transmitted to a device for display by a user interface of an application for playing media. The page also includes interactive features enabling a listener to participate in the media program, and visual features identifying a participant speaking during the media program or one or more attributes of the speech of the participant.

Abstract: Implementations for detecting interactive content shared during media conferences are described. Initially, a frame to be presented to one or more participants during a media conference is received. Next, interactive content within the frame is identified. A bounding perimeter surrounding the interactive content is determined. At least one action based on the type of interactive content is determined. The bounding perimeter is transmitted for the interactive content within a plurality of frames to enable the at least one action associated with the interactive content.

Abstract: Systems and methods are disclosed for automated containers for packing of multi-item packages. In one embodiment, an example system may include an automated container having a first wireless charging coil, a first moveable sidewall, a second moveable sidewall, and a first actuator configured to move at least one of the first moveable sidewall or the second moveable sidewall. The system may include a controller configured to determine a first size of a first item, determine a second size of a second item, determine a first position for the first moveable sidewall based at least in part on the first size and the second size, determine a second position for the second moveable sidewall at least in part on the first size and the second size, cause the first moveable sidewall to move to the first position, and cause the second moveable sidewall to move to the second position.

Abstract: Some embodiments provide a method for training a machine-trained (MT) network to classify inputs into multiple categories. The method propagates a set of input training items through the MT network to generate a set of corresponding outputs. Each input training item belongs to a category and the output for each input training item includes, for each category, a computed probability of the input belonging to the category. The method computes a value for a loss function based on the generated outputs. The loss function includes a first term based on the computed probabilities of each input belonging to its category and not based on individual computed probabilities of the inputs belonging to other categories and a second term based on the individual computed probabilities of each input belonging to each of the categories. The method uses the computed value for the loss function to train the MT network.

Abstract: A virtualized computing service provides a computing instance capable of requesting attestation of the authenticity of the hypervisor implementing the computing instances. An attestation device included in a virtualization host maintains a log of hash values representing hypervisor versions that have been implemented at the virtualization host. Also, an independent auditor (e.g., attestation service) is provided software configurations that are known to be authentic. The independent auditor generates hash values for the authentic hypervisor versions. The computing instance receives a response from the local attestation device indicating hash values of hypervisor versions currently and/or previously deployed on the virtualization host, and the computing instance forwards the hash values to the independent auditor to authenticate that they match the hash values of the known authentic hypervisor versions.

Abstract: In various embodiments described in the present disclosure data objects are ingested and associated with a profile. For example, an object mapping indicates a set of keys included in a data object which are used to search an index to determine a profile that is associated with the data object. In various embodiments, data from the data object is then used to update the profile.

Abstract: A method for providing access to an aggregated profile can include receiving data from data sources. The data can include one or more identification indicators for target entities. Each identification indicator can be mapped to one or more entities of the target entities. Historical interactions that may be associated with the target entities can be identified based at least in part on the mapping. A particular target entity of the target entities can be identified based at least in part on the mapping and the one or more historical interactions. An aggregated profile can be generated for the particular target entity based at least in part on the mappings and the one or more historical interactions. Access can be provided to the aggregated profile.

Abstract: Techniques for a service provider network to identifying impacts affecting transmission of a live event to multiple client devices are discussed herein. A system can gather data from a variety of sources associated with the service provider network (e.g., a video service, a client device, a social media service, etc.) and identify patterns in network activity and/or user behavior that are indicative of a potential problem to deliver the live event. In some examples, the system can initiate a query for information associated with the potential problem, and output the query results to a model and/or a user interface for review. The system can, in various examples, determine an action to remedy the potential problem.

Abstract: A method includes receiving first data associated with an acceleration of a first device, determining that the acceleration is greater than a first threshold, receiving second data generated by a camera, determining, based at least in part on the second data, a difference between first pixels within a first frame and second pixels with a second frame, determining that the difference is greater than a second threshold, and sending third data to a second device.

Abstract: Improved crash resilience is implemented in shared memory processing systems. A processing application including a plurality processing groups may be configured, and a processing group including a plurality of processes may allocate memory from a global shared memory to a group-private shared memory arena. Then, an abnormal execution condition for a failing process of the processing group may be detected during execution of a processing job. Responsive to the detection, a recovery of execution of the processing group may be performed while continuing execution of processing groups of the application. The recovery may include restoring the processing group to a normal execution state and returning the allocated memory of the group-private memory arena to the global shared memory.

Abstract: Described is a system and method for developing and dynamically updating user paths through a materials handling facility. A user path through a materials handling facility may be developed that will route a user past one or more items (e.g., pick list items, recommended items, wish list items, promotional items) planned for the user. If the user does not follow the user path, the user path may be dynamically updated to account for the current location of the user and/or to consider other actions performed by the user.

Abstract: Metadata indicating that a virtual traffic hub enabling connectivity between a plurality of isolated networks has been established is stored. A determination is made that a first entry of a first isolated network attached to the hub is to be represented in a second routing table of a second isolated network attached to the hub, e.g., to enable network packets originating at resources of the second isolated network to be transmitted via the hub to the first isolated network. A new entry corresponding to the first entry is included in the second routing table.

Abstract: Systems and methods for providing and using a liner including one or more rigid over-molds for a chassis of a computing device cooled with a liquid fluid are disclosed. The liner may include a flexible membrane configured to be positioned in a chassis to form a liquid barrier that retains a liquid cooling fluid and one or more rigid over-molds configured to align components in fixed positions relative to other respective components or the chassis. The rigid over-molds may be located on the interior and/or the exterior of the liner to fixedly couple to components and/or the chassis. A set of over-molds may also be configured to enable insertion of connectors exterior to the liner, wherein a first rigid over-mold on the interior of the liner may be aligned with a second rigid over-mold on the exterior of the liner.

Abstract: Systems and methods directed to a service that analyzes media and outputs a confidence score of the media based on inconsistencies within media and differences between source media and the media, are described herein. In one aspect, a media file may be obtained at a computing resource service provider. The media file may be analyzed using at least one of a plurality of functions to generate media results that indicate inconsistencies within the media file, and/or differences between the media file and corresponding media registered with the computing resource service provider or contextual media that contains similar content to the media file, where individual functions of the plurality of functions are executed in parallel by at least one compute unit. A confidence score, which indicates a significance of the inconsistencies within the media file and/or differences from the registered or contextual media, may be generated based on the media results.

Abstract: A data set management of a provider network may allow a user to create new data set instances. When a data set instance is created, data set lineage metadata is also generated to describe the new data set instance, including the transformation that was applied to data in order to create the data set instance. When modifications are made to source data (e.g., a data bucket), then the modifications are propagated via transformations to the parent data set instance and to any child data set instances according to the data set lineage metadata in order to update the data set instances. When modifications are made to a parent data set instance to create an updated parent data set instance, then the modifications are propagated via transformations to any child data set instances according to the data set lineage metadata. Transformations and transformation patterns may also be defined and scheduled.

Abstract: A consumer database of a service provider network may perform writes to a producer database. To do so, the consumer database obtains metadata and a transaction context from the producer database and uses that information to perform a write from the consumer database to the producer database. A consumer database may access (read and/or write) a producer database that uses a different topology to manage data. To do so, the consumer database obtains topology metadata from the producer database and uses the topology metadata to perform a query (read or write) from the consumer database to the producer database. A consumer database may access a producer database based on a mapping of local user IDs to consumer databases at the producer database. The access may be granted to the consumer database based on permissions assigned to the local user IDs.