Abstract: Dialog visualizations are created to enable analysis of interactions between a user and a speech recognition system used to implement user commands. Spoken commands from the user may be classified, along with system responses to the spoken commands, to enable aggregation of communication exchanges that form dialog. This data may then be used to create a dialog visualization. The dialog visualization may enable an analyst to visually explore different branches of the interactions represented in the dialog visualization. The dialog visualization may show a trajectory of the dialog, which may be explored in an interactive manner by the analyst.

Abstract: A system and method of deploying operating environments in an enterprise computing environment comprised of managed virtual or hardware servers is disclosed. A library of operating environments, each environment including at least one package including an operating system image and an application, is provided. A user controlling a cluster of servers may request creation of a test environment using an operating environment from the library, and test the environment with applications to ensure the user's needs are met. The user may request all servers within the user's cluster be provisioned with the operating environment through a deployment manager.

Abstract: Methods and systems for implementing an abstraction layer for streaming data sources are disclosed. A request to perform an operation based on one or more keys is received using a key-value interface. A streaming data source is selected based on the request. The operation is performed using the streaming data source, wherein the operation comprises storing or retrieving one or more values based on the one or more keys.

Abstract: Systems and methods are described to predict spikes in requests for content on a computing network based on referrer field values of prior requests associated with spikes. Specifically, a traffic spike prediction service is disclosed that can analyze information regarding past requests on the computing network to detect a spike in requests to a content item, where a significant number of request within the spike include a common referrer field value. The traffic spike prediction service can then detect a request to a second content also including the common referrer field value, and predict that a spike is expected to occur with respect to the second content. The traffic spike prediction service can manage the expected spike by increasing an amount of computing resources available to service requests to the second content.

Abstract: Systems and methods are described herein for generating adaptive feedback in response to a user request. Input indicative of a user request may be received and utilized to identify an item in an electronic catalog. A title for the item may be retrieved and provided, as input, to a machine-learning model. The machine-learning model may be trained to identify one or more segments of an input title. A shortened title may be generated from these identified segments and provided as output at the user device (e.g., via audible output provided at a speaker of the user device). In some embodiments, the length and content of the shortened title may vary based at least in part on the contextual intent of the user's request.

Abstract: A media storage service may maintain an asset including information to indicate a specific media content and a specific number of replicas to be stored by the media storage service for the specific media content. The media storage service may receive content recording-related requests from a recording manager to define or update the specific number for the specific media content. The media storage service may receive one or more segments specific to the media content from an encoder. The media storage service may create and store the specific number of replicas, as indicated in the asset, for the media content.

Abstract: A head-mounted wearable device provides many functions to a user. The functions may be provided based in part on whether the device is being worn (donned) or not worn (doffed). Force sensing resistors in the device provide output indicative of forces associated with wearing the device. A time series of values indicative of the magnitude of a force may be determined. An increase or decrease in force values measured at different times may indicate that the device has been donned or doffed, respectively. If a subsequent force value does not differ substantially from a previous force value, this may indicate that the device has remained donned or doffed.

Abstract: Generally described, one or more aspects of the present application relate to a public snapshot service for associating a block-level snapshot with certain snapshot properties that indicate a block load order, and creating a block storage volume from the snapshot by loading the blocks in the snapshot onto the block storage volume in the block load order indicated by the snapshot properties associated with the snapshot.

Abstract: Techniques for Scalable parallel elimination of approximately subsumed sets are described. A method for Scalable parallel elimination of approximately subsumed sets includes identifying a first plurality of blocks based at least on a plurality of records stored in a storage service of a provider network, determining a plurality of subsumption relationships between blocks from the first plurality of blocks, retaining a first subset of the first plurality of blocks and demoting a second subset of the first plurality of blocks based at least on the plurality of subsumption relationships, and iteratively performing dynamic blocking based at least on the first subset of the plurality of matching blocks and the second subset of the plurality of matching blocks to generate a subsequent pluralities of blocks.

Abstract: Embodiments herein describe a controller for a machine that wirelessly transmits a destination address to a robot which is tasked with transporting an item to the destination corresponding to the received address. Using the address, the robot can select a predefined path stored in its memory for that address. Because multiple robots can move items in the machine at the same time, the robots may collide if their predefined paths intersect. The machine can use a variety of different techniques to prevent collisions between the robots as the robots deliver items to different locations along their predefined paths.

Abstract: A technology is described for modifying a message frequency of a device. In one example, a network connection is identified between a device and a service provider environment. A message frequency is set for which to send reachability messages between the device and the service provider environment to maintain the network connection. The reachability messages may be sent between the device and the service provider environment at the message frequency. After testing whether the networking connection is maintained at the message frequency, the message frequency may be modified to a modified message frequency.

Abstract: Techniques for accommodating irregular timing information in streamed media data are described. According to some embodiments, at least a portion of a media stream that includes a video component is received, the video component including a plurality of video frames and video timing data, and wherein the media stream includes an indication of a video frame rate. A first duration of a first video frame of the plurality of video frames is determined to be different than an expected video frame duration, wherein the expected video frame duration is based at least in part on the indication of the video frame rate. A timestamp of at least one video frame of the plurality of video frames is adjusted to change the first duration.

Abstract: Techniques for dataset schema and metadata management are described. A data model management service provides a logically centralized repository that allows discovery of datasets via exploration of dataset metadata. The data model management service can ensure the validity of the datasets and the associated metadata by enforcing dataset validation rules and/or requiring dataset metadata to be provided and limiting visibility into non-compliant datasets.

Abstract: Methods and apparatus for token-based admission control for replicated writes are disclosed. Data objects are divided into partitions, and corresponding to each partition, at least a master replica and a slave replica are stored. A determination as to whether to accept a write request directed to the partition is made based at least in part on one or more of (a) available throughput capacity at the master replica, and (b) an indication, obtained using a token-based protocol, of available throughput capacity at the slave replica. If the write request is accepted, one or more data modification operations are initiated.

Abstract: A system for testing the streaming capacity of a media streaming system, or of one or more origin servers thereof, includes a control system that controls the provisioning of computing resources, such as virtual machine instances, configured as load generators. The load generators establish a plurality of concurrent streams of content from the media streaming system, thus representing a plurality of connected user devices. The streams impart a load on the media streaming system; additional load generators can be added to a scalable group to increase the load. The load generators can produce monitoring data describing errors in network activity (e.g., dropped packets) and in the stream data itself (e.g., synchronization errors). A metric analysis system calculates performance metrics based on the monitoring data, and signals the control system to add or remove load generators until the load is as high as possible without degrading service, revealing the streaming capacity.

Abstract: A system, method, and computer readable storage medium for inter-process rendering pipeline for shared process remote web content rendering. The method includes obtaining an image rendering request through an inter-process communication channel established by an application in which the image rendering request including a content area. The method further includes selecting a content rendering engine based on one or more image resources needed to generate web content as specified in the image rendering request, determining the one or more image resources corresponding to the image rendering request by using the content rendering engine, and generating the web content on the content area based at least in part on the one or more image resources.

Abstract: A program to be executed to perform a packet processing operation on a packet associated with a resource group, as well as security settings of the resource group, are received. The program is transmitted to a set of fast path nodes which were assigned to the resource group based on the group's metadata. With respect to a particular packet, security operations based on the settings are performed and the program is executed at a fast path node. Based at least partly on the results of the program, a packet routing action corresponding to the received packet is performed.

Abstract: In response to a determination that a first logical table is to be created in a data sheet, a respective row identifier is generated and stored for individual rows of the first logical table. To indicate a relationship between a particular cell of another logical table and a particular row of the first logical table, the row identifier of the particular row is stored. After a modification of a value stored in the particular row, the row identifier is used to determine an updated value to be displayed in the particular cell of the other logical table.