Abstract: Disclosed are system architectures and techniques for multiple access points for a data container. Control plane and data plane APIs are disclosed for generating access points and associating the access points with data containers, as well as for generating access point policies that specify permissions for the access points. More than one access point may be associated with a single data container. A storage service receives requests directed to the access points, and grants or denies the requests based on the permissions specified in the corresponding policies. Various types of use cases for using access policies are contemplated such as default, regional, or network-based (e.g., VPC-based) use-cases or the like. A system may implement layers of policies such as identity and access management policies, access point policies and data container policies.

Abstract: Described are apparatus, systems, and methods that are operable to conserve battery power of a wearable device by monitoring a state of the user wearing the wearable device and adjusting power state of one or more components of the device based on the determined user state and/or altering a wireless transmission of sensor data collected by the wearable device from the wearable device to a portable device associated with the user. For example, if it is determined that the user is in a sleep state, the wireless connection between the wearable device and the portable device may be terminated, the wireless transmitter and the processor transitioned to a low power state, and sensor data stored in a buffer memory of the wearable device until a defined period of time before the user exits the sleep state and/or upon satisfaction of a buffer usage threshold.

Abstract: This disclosure describes a system for fulfilling items at a materials handling facility. In some instances, a predicted items list that identifies items that are likely to be picked by a user are determined and, when the user arrives at the materials handling facility, those predicted items are presented to the user for selection. For example, predicted items may be determined and an inventory holder that holds one or more of those predicted items may be retrieved by a mobile drive unit (such as a Kiva mobile drive unit) and presented to the user at a retrieval area. The user may pick the items they desire from the presented inventory holder.

Abstract: Techniques are described for enabling software applications to obtain temporary security credentials used to interact with a cloud provider network and, upon the revocation of an active set of temporary security credentials used by an application (e.g., due to concerns about the temporary credential's potential exposure to one or more unauthorized third parties), to readily obtain new temporary security credentials that the application can use to continue operation with minimal interruption. The temporary security credentials can be used, for example, to enable the cloud provider network to authenticate requests sent by software applications or users to various services or other components of the cloud provider network. An operator of a cloud provider network may provide a software development kit (SDK) that application developers can use to incorporate functionality related to the management of temporary security credentials.

Abstract: Replicated instances in a distributed computing environment provide for automatic failover and recovery. A component monitors the status of event processors in a set or bucket and handles the failure of an event processor. For a large number of instances, the data environment can be partitioned such that each monitoring component is assigned a partition of the workload. At intervals, each event processor sends a “heartbeat” message to the event processors in the bucket covering the same workload partition, to inform the other event processors of the status of the event processor sending the heartbeat. If it is determined that a heartbeat is received from each event processor in the bucket, a current process can continue. In the event of monitoring component failure, the instances can be repartitioned, and the remaining monitoring components can be assigned to the new partitions to substantially evenly distribute the workload.

Abstract: An industrial monitoring system comprises monitoring devices that are enabled for debugging by a component comprising a first microcontroller unit and a second microcontroller unit. The monitoring device receives a debugging command based on a subscription, on a publish-subscribe communications channel, to events indicative of requests to perform debugging operation. The second microcontroller unit causes the first microcontroller unit to perform the command. Results of performing the command are returned by publishing an event to the publish-subscribe communications channel.

Abstract: Techniques are described for providing logical networking functionality for managed computer networks, such as for virtual computer networks provided on behalf of users or other entities. In some situations, a user may configure or otherwise specify a network topology for a virtual computer network, such as a logical network topology that separates multiple computing nodes of the virtual computer network into multiple logical sub-networks and/or that specifies one or more logical networking devices for the virtual computer network. After a network topology is specified for a virtual computer network, logical networking functionality corresponding to the network topology may be provided in various manners, such as without physically implementing the network topology for the virtual computer network. In some situations, the computing nodes may include virtual machine nodes hosted on one or more physical computing machines or systems, such as by or on behalf of one or more users.

Abstract: Autonomous ground vehicles capture images during operation, and process the images to recognize ground surfaces or features within their vicinity, such as by providing the images to a segmentation network trained to recognize the ground surfaces or features. Semantic maps of the ground surfaces or features are generated from the processed images. A point on a semantic map is selected, and the autonomous ground vehicle is instructed to travel to a location corresponding to the selected point. The point is selected in accordance with one or more goals, such as to maintain the autonomous ground vehicle at a selected distance from a roadway or other hazardous surface, or along a centerline of a sidewalk.

Abstract: Techniques for generating a personalization value that measures how tailored certain system interactions are for a user are described. A dialog exchange between a user and a skill may be determined, with the dialog exchange including user input data and system output data. It may be determined that the system output data was generated without respect to at least one previous user input or system output of the dialog exchanges. Based on this, a personalization value may be generated and sent to the skill.

Abstract: Disclosed herein are techniques for performing multi-layer neural network processing for multiple contexts. In one embodiment, a computing engine is set in a first configuration to implement a second layer of a neural network and to process first data related to a first context to generate first context second layer output. The computing engine can be switched from the first configuration to a second configuration to implement a first layer of the neural network. The computing engine can be used to process second data related to a second context to generate second context first layer output. The computing engine can be set to a third configuration to implement a third layer of the neural network to process the first context second layer output and the second context first layer output to generate a first processing result of the first context and a second processing result of the second context.

Abstract: A system and method relating to multi-level data de-duplication operations relating to data records associated with multiple user systems. The system includes a first set of computing instances execute a first set of de-duplication operations to generate a set of locally de-duplicated files associated with a data stream comprising the data records associated with the user systems. The system includes a storage system to store the set of locally de-duplicated files. The system includes a second set of computing instances to receive, in accordance with a frequency type of multiple frequency types, the set of locally de-duplicated files from the storage system and execute a second set of de-duplication operations to generate a set of globally de-duplicated files associated with the data records.

Abstract: Devices and techniques are generally described for dynamic policy determination using machine learning. In some examples, first goal data may be received. A first machine learning model may generate a first computer-executable policy based at least in part on the first goal data. In some examples, a first search query may be received. In various examples, the first search query may be modified by the first computer-executable policy to generate a modified search query. In some examples, first feedback data related to user interaction with search results of the modified search query may be determined. In various examples, at least one parameter of the first machine learning model may be updated based at least in part on the first feedback data to generate a second machine learning model.

Abstract: Techniques for speech processing using a deep neural network (DNN) based acoustic model front-end are described. A new modeling approach directly models multi-channel audio data received from a microphone array using a first model (e.g., multi-channel DNN) that takes in raw signals and produces a first feature vector that may be used similarly to beamformed features generated by an acoustic beamformer. A second model (e.g., feature extraction DNN) processes the first feature vector and transforms it to a second feature vector having a lower dimensional representation. A third model (e.g., classification DNN) processes the second feature vector to perform acoustic unit classification and generate text data. These three models may be jointly optimized for speech processing (as opposed to individually optimized for signal enhancement), enabling improved performance despite a reduction in microphones and a reduction in bandwidth consumption during real-time processing.

Abstract: As part of a compute instance migration, a compute instance which was executing at a first server begins execution at a second server before at least some state information of the compute instance has reached the second server. In response to a determination that a particular page of state information is not present at the second server, a migration manager running at one or more offload cards of the second server causes the particular page to be transferred to the second server via a network channel set up between the offload cards of both servers, and stores the page into main memory of the second server.

Abstract: Systems and methods to encode sounds in association with containers, as well as systems and methods to recognize encoded sounds associated with containers, are described. The containers may include various structures to encode sounds. Responsive to opening of containers, the various structures may emit the encoded sounds. An audio capture device may receive the encoded sounds, and a computing device may process the encoded sounds to identify audio signatures within the encoded sounds. The audio signatures may be matched with known audio signatures to identify the containers and/or information associated with the containers.

Abstract: An image may be evaluated by a computer vision system to determine whether it is fit for analysis. The computer vision system may generate an embedding of the image. An embedding quality score (EQS) of the image may be determined based on the image's embedding and a reference embedding associated with a cluster of reference noisy images. The quality of the image may be evaluated based on the EQS of the image to determine whether the quality meets filter criteria. The image may be further processed when the quality is sufficient, or otherwise the image may be removed.

Abstract: Technologies are provided for supporting storage device write barriers. A host computer can be configured to transmit a write barrier command to a storage device to indicate that one or more data access commands should be processed before one or more other data access commands are processed. For example, a host computer can transmit one or more data access commands to a storage device. The host computer can then transmit a write barrier command to the storage device. The storage device can be configured to receive the write barrier command and to associate a write barrier with the one or more data access commands. The host computer can continue to transmit additional data access commands to the storage device. However, the storage device will not process the additional data access commands until after the one or more data access commands associated with the write barrier have been processed.

Abstract: Techniques for generation of synthetic queries from customer data for training of document querying machine learning (ML) models as a service are described. A service may receive one or more documents from a user, generate a set of question and answer pairs from the one or more documents from the user using a machine learning model trained to predict a question from an answer, and store the set of question and answer pairs generated from the one or more documents from the user. The question and answer pairs may be used to train another machine learning model, for example, a document ranking model, a passage ranking model, a question/answer model, or a frequently asked question (FAQ) model.

Abstract: Techniques are disclosed for locally printing a print job generated by a remote web services platform. It may be that a local computer and an application of the remote web services platform that generates the print job lack a bidirectional communications link (like a VPN connection). A printer service on the web services platform presents itself as a printer to an application on the web services platform. The printer service receives a print job from the application and transmits it to a client-side print service. The client print service then sends the print job to a local printer.

Abstract: A distributed data warehouse system maintains data blocks on behalf of clients, and stores primary and secondary copies of data blocks on different disks or nodes in a cluster. The data warehouse system may back up data blocks in a key-value backup storage system. In response to a query targeting a data block previously stored in the cluster, the data warehouse system may determine whether a consistent, uncorrupted copy of the data block is available in the cluster (e.g., by applying a consistency check). If not (e.g., if a disk or node failed), the data warehouse system may automatically initiate an operation to restore the data block from the backup storage system, using a unique identifier of the data block to access a backup copy. The target data may be returned in a query response prior to restoring primary and secondary copies of the data block in the cluster.

Abstract: A shard polling system fairly distributes stored items from producers to consumer processes and includes polling threads that poll for items from respective portions of a storage source, place the items in respective queues, and increment a global permit counter, restricted to a configurable maximum, that tracks the quantity of messages across the respective queues. The polling threads are restricted by respective shard permit counters that limit a quantity of items that may be moved from a storage source to a respective queue. A global thread applies a fairness technique, such as round robin, to fairly obtain the items from each of the respective queues and places the items in a global queue, decrementing the global permit counter, and incrementing the respective shard permit counter for each item obtained. Consumer threads poll the global queue (e.g., a blocking output queue) for the messages.

Abstract: A host application installed on a computing device detects that the computing device is in proximity of a location associated with another application that can be installed onto the computing device. As a result of the computing device being in proximity of the location, the host application obtains, from an application vending server, application code for the application. The host application uses the application code to install the other application onto the computing device. If the host application detects that the computing device is no longer in proximity of the location, the host application uninstalls the other application from the computing device.

Abstract: This disclosure describes systems, devices, and methods for performing and facilitating tenant migration between multiple bare-metal servers. An example method includes receiving an indication of an impending reboot of a first bare-metal server. The first bare-metal server may be hosting a tenant. The method further includes identifying a second bare-metal server in a pre-initialized state. The method also includes causing the first bare-metal server to migrate data associated with the tenant to the second bare-metal server in advance of the reboot.

Abstract: The following description is directed to a configurable logic platform. In one example, a configurable logic platform includes host logic and a reconfigurable logic region. The reconfigurable logic region can include logic blocks that are configurable to implement application logic. The host logic can be used for encapsulating the reconfigurable logic region. The host logic can include a host interface for communicating with a processor. The host logic can include a management function accessible via the host interface. The management function can be adapted to cause the reconfigurable logic region to be configured with the application logic in response to an authorized request from the host interface. The host logic can include a data path function accessible via the host interface. The data path function can include a layer for formatting data transfers between the host interface and the application logic.

Abstract: Techniques for analyzing stored video upon a request are described.

Abstract: Systems and methods are described herein for generating enhanced search results utilizing third-party website content within a search engine provided by an electronic catalog of a service provider. This content may be collected in advance of query processing and analyzed to identify a category indicating some attribute of the content (e.g., terms mentioned, topics discussed, object depicted in images/videos/3D data of the content, etc.). Items may be matched to the website through analyzing the textual and/or visual representation data of the website to textual and/or visual representation data associated with an item offered within the electronic catalog. A query may be subsequently received and a third-party website may be identified as being relevant to the search query. In response to the query, the third-party website may be included in a search result list along with images and/or text identifying items pertaining to that website.

Abstract: System and methods are provided for rapid provisioning of data management services. A user creates a data management service through a user interface. The creation of the data management services is initiated by the user without any programming by the user in a scalable, rapidly provisioned environment. The data management service natively supports the CRUD model (create, read, update, and delete) for data entities, metadata, references, as well as other features, such as search capabilities and/or data lineage. Once the data management service has been provisioned, the service scales as usage of the service increases. The data management service is an abstraction layer and the organization that created the service can remain unaware of the underlying implementation details of the service.

Abstract: Systems and methods for providing cryptographic services. A cryptography service obtains a request to provision a computing device to perform cryptographic operations. The cryptography service generates executable code for a protected execution environment. The computing device obtains and executes the executable code. The computing device fulfills requests for cryptographic operations in the protected execution environment.

Abstract: Techniques for enhancing functionalities of a voice-based user device are described. In an example a computer system generates, from a first taxonomy, a second taxonomy for a user based on an item purchase history. The computer system also identifies items based on this history and generates a list that associates the items with categories of the second taxonomy. Based on a user utterance, the computer system generates text-to-speech content from the list. This content includes category information and item information. The category information identifies a first category from the categories. The item information identifies a first item from the items, where this item is associated with the first category. The category information and the item information are presented to the user. The computer system adds an identifier of the first item to an item checkout process based on the presentation of the item information.

Abstract: Aspects of the present disclosure relate to machine learning techniques for identifying collections of items, such as furniture items, that are visually complementary. These techniques can rely on computer vision and item imagery. For example, a first portion of a machine learning system can be trained to extract aesthetic item qualities or attributes from pixel values of images of the items. A second portion of the machine learning system can learn correlations between these extracted aesthetic qualities and the level of visual coordination between items. Thus, the disclosed techniques use computer vision machine learning to programmatically determine whether items visually coordinate with one another based on pixel values of images of those items.

Abstract: Features are disclosed for an interface for verifiable tracking of an item through a supply chain using a distributed electronic ledger. For example, when an item is added to a catalog system, the item information may be included in the creation element at the start of the ledger for the item. A certification authority may be included to verify that items received correspond to the item included in the catalog based on one or more certification rules.

Abstract: A system actuates a force mechanism to control a force applied to items on a shelf. An amount of force to be applied to items on a shelf can depend upon context data associated with a user. The user context data may be based on an analysis of sensor data, such as image data or proximity data. Once user context data has been determined, configuration data can be generated and transmitted to the force mechanism to apply the designated amount of force to the items on the shelf. One way to apply the designated amount of force to the items on the shelf is to lift one end of the shelf upwards to create an angle between the shelf and a horizontal surface, using a height adjustment mechanism and a hinge. The variation in the angle produces different levels of force on the items.

Abstract: Certain embodiments herein relate to implementing collaborative bidding in online auctions. A determination may be made whether bidding entities, or bids associated with bidding entities, share promotional interests, such as promoting the same product. Bids submitted by bidding entities that share the same promotional interests may be combined to create a single, combined bid that may compete against other bids in an online auction. Costs associated with promoting the product may be allocated among such bidding entities in various ways, including proportional allocation of cost based on a bid price submitted by the bidding entity.

Abstract: A hybrid Bacon-Shor surface code is implemented using a fault tolerant quantum computer comprising hybrid acoustic-electric qubits. A control circuit includes an asymmetrically threaded superconducting quantum interference devices (ATS) that excites phonons in a mechanical resonator by driving a storage mode of the mechanical resonator and dissipates phonons from the mechanical resonator via an open transmission line coupled to the control circuit. The hybrid Bacon-Shor surface code only couples four phononic modes per given ATS, reducing cross-talk as compared to other systems that couple more phononic modes per ATS. Also, measurements are performed such that three parity measurements are taken between a phononic readout mode and a transmon qubit in a given syndrome measurement cycle.

Abstract: Techniques implemented by a game-hosting service to place a load balancer between client devices engaged in a game session and servers that maintain state data for the game session. Rather than having a single server maintain state data and host the game session, the load balancer will broadcast action data received from the client devices to the server that is hosting the game session, or “primary server,” and also to a redundant server that is maintaining state data for the game session, or “secondary server.” Thus, if the primary server is unable to continue hosting the session, the secondary server may begin hosting the game session using the state data maintained for the game session. Further, by offloading the network input/output (I/O) tasks of communicating with the client devices from the server to the load balancer, the server can allocate larger amounts of resources to hosting game sessions.

Abstract: Techniques for partitioning data flow operations between execution on a compute instance and an attached accelerator instance are described. A set of operations supported by the accelerator is obtained. A set of operations associated with the data flow is obtained. An operation in the set of operations associated with the data flow is identified based on the set of operations supported by the accelerator. The accelerator executes the first operation.

Abstract: Disclosed are various embodiments for the extraction of isolated nodes during source code refactoring. A graph model representative of a computing application is generated having nodes and bridges that connect some nodes to other nodes. An application component corresponding to a selected one of the nodes may be extracted from the computing application. An independently deployable component of the computing application may be generated and deployed on a network service such that the independently deployable component is accessible through a network-based call.

Abstract: A system in which players participate as characters in a virtual game world via game clients executing on client devices. An observer mode is provided that enables users to navigate in the game world without requiring a game client so that a user can navigate the game world while not actively playing the game. When a user accesses the game in observer mode, a view of the game world is rendered from the viewpoint of an observer's character; the observer's character may not correspond to a character in the game and the observer's character may not be visible to the players in the game. The rendered view of the game world is streamed to a viewing application on a personal device. Navigation inputs allow the observer's character to move freely within the game world.

Abstract: Techniques are described for enabling a software modernization assessment service of a cloud provider network to maintain a modernization knowledge base and to use the knowledge base to generate modernization recommendations for users' software applications. A modernization knowledge base comprises one or more modernization ontologies, where a modernization ontology defines concepts and relationships used to describe modernization tool capabilities and limitations, modernization strategies, etc. The modernization assessment service uses the modernization knowledge base to automate various software modernization processes including, for example, providing modernization recommendations for software applications (e.g., applications identified by users as candidates for modernization) and generating modernization assessment reports.

Abstract: Systems and methods are provided to perform multiply-accumulate operations of normalized numbers in a systolic array to enable greater computational density, reduce the size of systolic arrays required to perform multiply-accumulate operations of normalized numbers, and/or enable higher throughput operation. The systolic array can be provided normalized numbers by a column of normalizers and can lack support for denormal numbers. Each normalizer can normalize the inputs to each processing element in the systolic array. The systolic array can include a multiplier and an adder. The multiplier can have multiple data paths that correspond to the data type of the input. The multiplier and adder can employ expanded exponent range to operate on normalized floating-point numbers and can lack support for denormal numbers.

Abstract: A first instance is caused to execute software code to perform a first portion of a workflow in response to receipt of a workflow request, and performance of the first portion results in submission of an operation request to an entity. A resume workflow request is received from the entity, where the resume workflow request includes a handle to a snapshot that corresponds to a state of execution of the software code and a response to the operation request to the entity. Using the handle to the snapshot and the response to the operation request, a second instance is caused to execute the software code from the first state to perform a second portion of the workflow. A workflow result is received from an instance that executes a last portion of the workflow, and the workflow is provided result in response to the workflow request.

Abstract: A data storage system includes multiple head nodes and multiple data storage sleds mounted in a rack. For a particular volume or volume partition one of the head nodes is designated as a primary head node for the volume or volume partition. The primary head node is configured to store data for the volume in a data storage of the primary head node and cause the data to be replicated to a secondary head node. The primary head node is also configured to cause the data for the volume to be stored in a plurality of respective mass storage devices each in different ones of the plurality of data storage sleds of the data storage system.

Abstract: A service manages a plurality of virtual machine instances for low latency execution of user codes. The service can provide the capability to execute user code in response to events triggered on an auxillary service to provide implicit and automatic rate matching and scaling between events being triggered on the auxiliary service and the corresponding execution of user code on various virtual machine instances. An auxiliary service may be configured as an event triggering service to detect events and generate event messages for execution of the user codes. The service can request, receive, or poll for event messages directly from the auxiliary service or via an intermediary message service. Event messages can be rapidly converted to requests to execute user code on the service. The time from processing the event message to initiating a request to begin code execution is less than a predetermined duration, for example, 100 ms.

Abstract: Techniques for anomaly detection within a group of hosts are described.

Abstract: Availability of future capacity can be determined for various categories of resources. In some embodiments, logistic regression models are used to obtain binary predictions of availability. In this way, a customer or other entity can obtain a determination as to the availability of a specific number and category or resources, or resource instances, at a future period of time. In some embodiments alternatives may be provided, which may provide improved characteristics or may have a higher likelihood of availability. Such approaches can also be used to monitor current and future capacity demands and enable a provider to better plan appropriate adjustments to the physical resources that provide that capacity.