Abstract: Methods and systems for improved quality of a streaming session using multiple simultaneous streams. For the multiple simultaneous streams an audio/video device (A/V device) records and generates a high-resolution stream and a low-resolution stream for simultaneous transmission to a server. The server selects one of the two streams for retransmission to a destination client device. The server also monitors the streaming session and estimates a total available bandwidth between the server and the A/V device and assigns a confidence value to the bandwidth estimation. The server periodically transmits the bandwidth estimate and confidence value to the A/V device to improve the efficiency of the streams being generated by the A/V device. The A/V device can use the received bandwidth estimate and confidence value to adapt the resolution of each of the streams to efficiently use the total available bandwidth between the A/V device and the server.

Abstract: Data is encoded to be incrementally authenticable. A plaintext is used to generate a ciphertext that comprises a plurality of authentication tags. Proper subsets of the authentication tags are usable to authenticate respective portions of plaintexts obtained from the ciphertext. Portions of the plaintext can be obtained and authenticated without decrypting the complete ciphertext.

Abstract: Systems and methods are described controlling ingestion of data items within a data stream by executions of a serverless function on a serverless compute system. A poller device can act as an intermediary between the data stream and the serverless function, iteratively retrieving data items from the data stream and passing them in invocations of the serverless function. To allow for fine-grained control of ingestion without requiring implementation of complex logic at the poller device, the poller device can enable the serverless function to pass instructions controlling subsequent operation of the poller device. Each execution of the serverless function may determine whether subsequent operation of the poller device should be altered, and if so, instruct the poller device accordingly. The poller device can then modify its operation pursuant to the instructions, enabling highly granular control of streaming data ingestion without inhibiting existing benefits of serverless computing.

Abstract: The present disclosure generally relates to a system of a delivery device for combining sensor data from various types of sensors to generate a map that enables the delivery device to navigate from a first location to a second location to deliver an item to the second location. The system obtains data from RGB, LIDAR, and depth sensors and combines this sensor data according to various algorithms to detect objects in an environment of the delivery device, generate point cloud and pose information associated with the detected objects, and generates object boundary data for the detected objects. The system further identifies object states for the detected object and generates the map for the environment based on the detected object, the generated object proposal data, the labeled point cloud data, and the object states. The generated map may be provided to other systems to navigate the delivery device.

Abstract: An optical transceiver includes a built-in optical switch to switch between diverse fiber paths in switches in a datacenter or in switches between two datacenters. The built-in optical switch can be used to switch between racks in a datacenter to increase capacity for any rack that requests it. A controller, which receives a signal from a server computer in one of the racks, can be external to the optical transceiver or within the optical transceiver. In either case, the server computer can be provided with additional bandwidth when needed. For connections between datacenters, the built-in optical switch allows for optical line protection, but without the need for a splitter circuit, which incurs a significant power loss and requires a more expensive transceiver. Consequently, the built-in optical switch within an optical transceiver can be used in a variety of contexts to increase efficiency and reduce overall costs for network devices.

Abstract: Described are systems and methods for monitoring, detecting, and/or protecting various systems of an aerial vehicle, such as an unmanned aerial vehicle (UAV). Embodiments of the present disclosure can provide a multi-tiered system to provide monitoring, detection, and/or initiation of protection protocols in response to detected faults in connection with the electronics associated with UAV systems, such as the motor drive and/or control systems that may drive the propulsion systems of the UAV.

Abstract: A clock disciplining scheme uses a pulse per second (PPS) signal that is distributed throughout a network to coordinate timing. In determining the time, jitter can occur due to latency between detection of the PPS signal and a software interrupt generated there from. This jitter affects the accuracy of the clock disciplining process. To eliminate the jitter, extra hardware is used to capture when the PPS signal occurred relative to a hardware clock counter associated with the clock disciplining software. In one embodiment, the extra hardware can be a sampling logic, which captures a state of a hardware clock counter upon PPS detection. In another embodiment, the extra hardware can initiate a counter that calculates a delay by the clock disciplining software in reading the hardware clock counter. The disciplining software can then subtract the calculated delay from a hardware clock counter to obtain the original PPS signal.

Abstract: A latch mechanism for a sidewalk delivery robot or container includes a plunger that extends to engage a latch body on a lid to lock the lid closed, a cable for pulling the plunger out of engagement with the lid to free the lid for opening, and a temporary catch mechanism for holding the plunger in the retracted position during an initial phase of the opening process. The catch, and therefore the plunger, is released when a forward-extending release arm of the catch is engaged by the latch body during the opening process to pivot the catch. A slot hinge mechanism includes a spring piston to guard against finger pinching and a slider that is attached to the piston of the latch mechanism by the cable. Actuation of the hinge mechanism retracts the piston to its retracted position.

Abstract: Techniques for performing speech processing using multi-modal widget information are described. A system may receive input data corresponding to a user input. The system may also receive widget context data corresponding to one or more multi-modal widgets active at a device. The system may use the widget context data to perform natural language understanding (NLU) processing with respect to the user input, and for selecting a skill component for responding to the user input. The system may send a widget identifier to the skill component when invoking the skill to respond to the user input.

Abstract: Disclosed are various embodiments for inferring brand similarities using graph neural networks and selection prediction. In one embodiment, a brand-to-brand graph is generated indicating similarities between a set of brands according to at least one of: click-through data or conversion data. Using a first graph neural network (GNN) tower, the brand-to-brand graph is analyzed to determine brand similarities among a first brand identified from a search query and a first set of other brands. Using a second GNN tower, the brand-to-brand graph is analyzed to determine brand similarities among a second brand and a second set of other brands. A level of similarity between the first brand and the second brand is determined based at least in part on an output of the first GNN tower and an output of the second GNN tower.

Abstract: Aspects of the present disclosure relate to the sharing of synchronization information corresponding to content among devices in a communication network. An originating device collects synchronization information related to the presentation of content to a user via the originating device. The synchronization information is defined according to reference points related to the presentation of the content to the user via the originating device, including a current state of the presentation of the content and/or additional information obtained during the presentation of the content. Based on detection of a synchronization event, the originating device transmits the collected synchronization information, resulting in the receipt of the synchronization information by one or more receiving devices. Thereafter, the presentation of the content to the user on one of the receiving devices can incorporate the synchronization information collected and transmitted by the originating device.

Abstract: Systems and methods are described herein for implementing a hybrid codec to compress and decompress image data using both lossy and lossless compression. In one example encoding process, it may be determined whether a first block of pixels of a frame of image data contains an edge. A type of compression by which to encode the first block may be selected based on that determination. The first block may be compressed using the selected type of compression. At least one second value associated with the first block of pixels may be set to indicate at least oof the compressed value or the type of compression used to compress the first block.

Abstract: Techniques for performing machine learning inference calls in database query processing are described. A method for performing machine learning inference calls in database query processing may include generating a query plan to optimize a query for batch processing of data stored in a database service, the query plan including a batch mode operator to execute a function reference and an execution context associated with the batch mode operator, executing the query plan to invoke a function associated with the function reference, wherein the function sends a batch of requests, generated using the execution context, to a remote service and obtains a plurality of responses from the remote service, and generating a query response based on the plurality of responses.

Abstract: A data service implements a configurable data compressor/decompressor using a recipe generated for a particular data set type and using compression operators of a common registry (e.g., pantry) that are referenced by the recipe, wherein the recipe indicates at which nodes of a compression graph respective ones of the compression operators of the registry are to be implemented. The configurable data compressor/decompressor provides a customizable framework for compressing data sets of different types (e.g., belonging to different data domains) using a common compressor/decompressor implemented using a common set of compression operators.

Abstract: Availability zone and region recovery are described. For an availability zone (AZ), a recovery availability zone (rAZ) may be identified based on available computing capacity of the recovery availability zone and geographic proximity of the availability zone relative to the recovery availability zone. In an instance in which the availability zone is impacted in which at least one of hardware and software of the availability zone is not fully operational, a virtual private cloud (VPC) is generated that establishes a peered connection between the availability zone and the recovery availability zone. A service is executed in the recovery availability zone, permitting any other services executing in the availability zone to invoke the service and become partially or fully operational.

Abstract: Systems and methods are provided for implementing a multi-service file system for a hosted computing instance via a locally-addressable secure compute layer. Software within the instance can submit file operations to the secure compute layer, which the secure compute layer can translate into calls to one or more network-accessible storage services. To provide a multi-service file system, the secure compute layer can obtain mapping data mapping file system objects within the virtualized file system to different network-accessible storage services. On receiving a file operation, the secure compute layer can determine one or more network-accessible storage services corresponding to the file operation, and submit appropriate calls to the one or more network-accessible storage services.

Abstract: Techniques for change data capture (CDC) log augmentation are described. In some examples, a user configures CDC log augmentation by indicating which data should be included in a CDC log, and the database, when generating a CDC log associated with this configuration, can obtain the associated data and augment the CDC log by inserting this data into it. The augmented data can include one or more fields from a record in a separate database table, where the record can be identified based on the changed record represented by the CDC log.

Abstract: Various embodiments of systems and methods for providing virtualized (e.g., serverless) broker clusters for a data streaming service are disclosed. A data streaming service uses a front-end proxy layer and a back-end broker layer to provide virtualized broker clusters, for example in a Kafka-based streaming service. Resources included in a virtualized broker cluster are monitored and automatically scaled-up, scaled-down, or re-balanced in a way that is transparent to data producing and/or data consuming clients of the data streaming service.

Abstract: A technique for merging, via lattice surgery, a color code and a surface code, and subsequentially decoding one or more rounds of stabilizer measurements of the merged code is disclosed. Such a technique can be applied to bottom-up fault-tolerant magic state preparation protocol such that an encoded magic state can be teleported from a color code to a surface code. Decoding the stabilizer measurements of the merged code requires a decoding algorithm specific to the merged code in which error correction involving qubits at the border between the surface and color code portions of the merged code is performed. Error correction involving qubits within the surface code portion and within color code portion, respectively, may additionally be performed. In some cases, the magic state is prepared in a color code via a technique for encoding a Clifford circuit design problem as an SMT decision problem.

Abstract: Techniques for building, validating, and utilizing custom database software in a shared management environment are described. A database customization service allows users to provide database installation files as part of a configuration to generate a custom database engine version. The creation of custom database engine versions may be governed by policy, and the database customization service may provide recommendations for creating custom database engine versions according to various criteria. The custom database engine version can be built, and its functionality and correctness verified. Users may launch database instances using these custom database engine versions that can be managed under a shared management scheme where both users and the database customization service can perform host-level management.

Abstract: An object-based data storage service receives a request to store a data object in a first location corresponding to a first data storage tier. The request may specify a parameter to enable transitioning of the data object to another data storage tier. In response to the request, the object-based data storage service stores the data object in the first location and monitors access of the data object to determine usage data associated with the data object. The object-based data storage service processes the usage data to determine that the data object is to be transitioned to a second data storage tier. As a result of this determination, the object-based data storage service transitions the data object to a second location corresponding to the second data storage tier.

Abstract: Methods, systems, and computer-readable media for auto-tuning permissions using a learning mode are disclosed. A plurality of access requests to a plurality of services and resources by an application are determined during execution of the application in a learning mode in a pre-production environment. The plurality of services and resources are hosted in a multi-tenant provider network. A subset of the services and resources that were used by the application during the learning mode are determined. An access control policy is generated that permits access to the subset of the services and resources used by the application during the learning mode. The access control policy is attached to a role associated with the application to permit access to the subset of the services and resources in a production environment.

Abstract: Remote Triggered Black Holes (RTBHs) can be precisely placed on networks that are not directly physically connected to a target of an attack. A network source of a potential attack can be determined. A path between the network source and the target can be identified, and a determination can be made as to which networks along that path subscribe to an attack mitigation service. From multiple identified subscriber networks, a subscriber network can be identified that is determined to be appropriate for placement of a black hole to mitigate the attack. Once selected, the identified network can receive attack information and acknowledge placement of the black hole. The subscriber network can then begin discarding traffic for the attack target. A subscriber-owned list of network prefixes can be reviewed before allowing RTBH injection for a corresponding address space.

Abstract: Systems and methods for sender profile and/or recipient profile disambiguation and/or confirmation are disclosed. In instances where a sender profile is not indicated by a user sending a communication from a communal device, heuristic data may be utilized to infer the sender profile. Similar heuristic data may also be used when selection of the sender profile is associated with a low confidence level. Heuristic data may also be used to infer the recipient profile when the user does not indicate the recipient profile or when selection of the recipient profile is associated with a low confidence. Various confirmations may result from the sender and recipient profile disambiguation.

Abstract: When a first search query including an image of an item is received to search for items associated with similar images, a second search query that includes text based on the image is generated. The text may be based on previous queries associated with the depicted item, visual features of the image, or text that is present in the image. The results from the first search query are scored based on their correspondence with the image of the item. Results having a score greater than a threshold are presented first in the output, followed by a selected number of results from the second search query. Results from the first search query that are associated with a score less than the threshold may be presented after the results from the second search query. This presentation increases the likelihood that items presented earlier in the output are relevant to the initial query.

Abstract: A first configurable address decoder can be coupled between a source node and a first interconnect fabric, and a second address decoder can be coupled between the first interconnect fabric and a second interconnect fabric. The first address decoder can be configured with a first address mapping table that can map a first set of address ranges to a first set of target nodes connected to the first interconnect fabric. The second address decoder can be configured with a second address mapping table that can map a second set of address ranges to a second set of target nodes connected to the second interconnect fabric. The second address decoder can be part of the first set of target nodes. The first address decoder and the second address decoder can be configured or re-configured to determine different routes for a transaction from the source node to a target node in the second set of target nodes via the first and second interconnect fabrics.

Abstract: Described is a multiple-camera system and process for detecting, tracking, and re-verifying agents within a materials handling facility. In one implementation, a plurality of feature vectors may be generated for an agent and maintained as an agent model representative of the agent. When the object being tracked as the agent is to be re-verified, feature vectors representative of the object are generated and stored as a probe agent model. Feature vectors of the probe agent model are compared with corresponding feature vectors of candidate agent models for agents located in the materials handling facility. Based on the similarity scores, the agent may be re-verified, it may be determined that identifiers used for objects tracked as representative of the agents have been flipped, and/or to determine that tracking of the object representing the agent has been dropped.

Abstract: Systems and methods are described for merging customer profiles, such as may be implemented by a computer-implemented contact center service. In some aspects, a subset of profiles may be determined that satisfy merging criteria, where individual profiles include a plurality of data fields. At least one value in a first data field that conflicts between at least two profiles may be identified. Next a merged value may be selected for the first data field based on data deduplication criteria, where the data deduplication criteria includes at least one indicator of accuracy of values of the plurality of data fields. As a result of a determination that at least the subset of profiles of the group of profiles meet the merging criteria, at least the subset of profiles may be combined into a combined profile using the merged value.

Abstract: Disclosed herein are techniques for classifying data with a data processing circuit. In one embodiment, the data processing circuit includes a probabilistic circuit configurable to generate a decision at a pre-determined probability, and an output generation circuit including an output node and configured to receive input data and a weight, and generate output data at the output node for approximating a product of the input data and the weight. The generation of the output data includes propagating the weight to the output node according a first decision of the probabilistic circuit. The probabilistic circuit is configured to generate the first decision at a probability determined based on the input data.

Abstract: Disclosed are various embodiments for implementing passenger profiles for autonomous vehicles. A passenger of the autonomous vehicle is identified. A passenger profile corresponding to the passenger and comprising a passenger preference is identified. The passenger preference is identified. A configuration setting of the autonomous vehicle corresponding to autonomous operation of the autonomous vehicle is then adjusted based at least in part on the passenger preference.

Abstract: An autonomous mobile device (AMD) moves around a physical space while performing tasks. The AMD may have sensors with fields of view (FOVs) that are forward-facing. As the AMD moves forward, a safe region is determined based on data from those forward-facing sensors. The safe region describes a geographical area clear of obstacles during recent travel. Before moving outside of the current FOV, the AMD determines whether a move outside of the current FOV keeps the AMD within the safe region. For example, if a path that is outside the current FOV would result in the AMD moving outside the safe region, the AMD modifies the path until poses associated with the path result in the AMD staying within the safe region. The resulting safe path may then be used by the AMD to safely move outside the current FOV.

Abstract: Systems and methods are provided to eliminate multiplication operations with zero padding data for convolution computations. A multiplication matrix is generated from an input feature map matrix with padding by adjusting coordinates and dimensions of the input feature map matrix to exclude padding data. The multiplication matrix is used to perform matrix multiplications with respective weight values which results in fewer computations as compared to matrix multiplications which include the zero padding data.

Abstract: A skid-steer delivery autonomous ground vehicle has a drive train and suspension that aids in maneuverability. The AGV has six wheels, each of which is powered by its own motor. The AGV has features that diminish the dragging effect on the wheels, either by choice of wheel features or by taking weight off the front wheels during turning.

Abstract: Systems, devices, and methods are provided for processing images using machine learning. Features may be obtained from an image using a residual network, such as ResNet-101. Features may be analyzed using a classification model such as K-nearest neighbors (K-NN). Features and metadata extracted from images may be used to generate other images. Templates may be used to generate various types of images. For example, assets from two images may be combined to create a third image.

Abstract: A materialization configuration request is received via a programmatic interface from a client of a journal-based multi-data-store database. The request indicates a partitioning rule to be used to select, for respective writes indicated in committed transaction entries of a journal, the materialization node at which the writes are to be stored. A control plane component of the database verifies that a set of materialization nodes corresponding to the partitioning rule has been established, and initiates the propagation of writes from the journal to the materialization nodes by respective write appliers.

Abstract: Techniques for training a machine-learning model are described. In an example, a computer generates a first pseudo-label indicating a first mask associated with a first object detected by a first machine-learning model in a first training image. A transformed image of the first training image can be generated using a transformation. Based on the transformation, a second pseudo-label indicating a second mask detected in the transformed image and corresponding to the first mask can be determined. A second machine-learning model can be trained using the second pseudo-label. The trained, second machine-learning model can detect a third mask associated with a second object based on a second image.

Abstract: A first computing device acquires video data representing a user performing an activity. The first device uses a first pose extraction algorithm to determine a pose of the user within a frame of video data. If the pose is determined to be potentially inaccurate, the user is prompted for authorization to send the frame of video data to a second computing device. If authorization is granted, the second computing device may use a different algorithm to determine a pose of the user and send data indicative of this pose to the first computing device to enable the first computing device to update a score or other output. The second computing device may also use the frame of video data as training data to retrain or modify the first pose extraction algorithm, and may send the modified algorithm to the first computing device for future use.

Abstract: A database management system receives a command defining a view of the database. The view definition is accepted without determining whether references to schema elements within the view definition are resolvable to existing elements of the database schema. A query of the view is received. In response to the query of the view, the database management system resolves references to schema elements in the view definition by determining whether the references correspond to data available for processing the query.

Abstract: Systems and methods for automatically prioritizing computing resource configurations for remediation include receiving information describing configuration issues that may result in impaired system performance or unauthorized access, parsing that information and automatically analyzing configuration details of a user's private computing environment to determine that assets provide an environment in which configuration issues may be exploited to produce undesired results. Such systems and methods can generate assessments indicating the likelihood an issue can be exploited and potential impacts of the issue being exploited. Such systems and methods can use these assessments to generate a report prioritizing remediation of specific configuration issues for specific vulnerable assets based on the actual configuration of the user's computing resources and the data managed using those resources.

Abstract: Devices, systems, and methods are provided for enhanced geographical caching of estimated arrival times. A method may include receiving respective user inputs indicative of respective users being in transit to a destination location from within a geographic region; determining, for the first user and the second user, a first estimated time of arrival from a first geographical area to the destination location, the first geographical area including the first location and the second location; identifying a third location of the first device at a third time, wherein the third location is within the first geographical area; determining that a time-to-live (TTL) of the first estimated time of arrival has not expired at the third time; and refraining from recalculating the first estimated time of arrival.

Abstract: A system and method for providing quantum entanglement using a hybrid space-fiber quantum network are described. The hybrid space-fiber quantum network includes a communications hub located proximate to an optical ground station and also includes an aerial entangled particle source, such as an entangled photon source attached to a satellite, drone, aircraft, etc. An atmospheric or free-space channel is used to distribute quantum entanglement between optical ground stations that are separated by geographic distances, via the aerial entangled particle source. Also, fiber optic links are connected to the communications hub located proximate to the optical ground station. The communications hub includes optical switches that enable any of the fiber optic links connected to the communications hub to receive or send distributed quantum entanglement to a remotely located recipient endpoint via the atmospheric or free-space channel.

Abstract: Server-specified subscription filters for long-lived client requests to fetch data in response to events. In one aspect, the techniques encompass a method performed by a set of one or more computing devices. The method includes the step of receiving a long-lived request to fetch data in response to events sent by a client computing device. The method further includes receiving a server-specified subscription filter for the long-lived request and executing the long-lived request. Executing the long-lived request includes creating a persistent function that uses the server-specified subscription filter to map a source event stream to a response event stream. The response event stream is provided to the client computing device. The server-specified subscription filter facilitates filtering of events fetched for the long-lived request in a way that may not be possible or impractical if the subscription client were required to specify the filter in the long-lived request.

Abstract: This disclosure describes systems and methods for using a primary device, communicatively coupled to a remote system, to configure or re-configure a secondary device in the same environment as the primary device. In some instances, the primary device may communicatively couple to the secondary device via a short-range wireless connection and to the remote system via a wireless area network (WAN), a wired connection, or the like. Thus, the primary device may act as an intermediary between the secondary device and the remote system for configuring the secondary device.

Abstract: Implementations detailed herein include description of a computer-implemented method.

Abstract: An acoustic event detection system may employ one or more recurrent neural networks (RNNs) to extract features from audio data, and use the extracted features to determine the presence of an acoustic event. The system may use self-attention to emphasize features extracted from portions of audio data that may include features more useful for detecting acoustic events. The system may perform self-attention in an iterative manner to reduce the amount of memory used to store hidden states of the RNN while processing successive portions of the audio data. The system may process the portions of the audio data using the RNN to generate a hidden state for each portion. The system may calculate an interim embedding for each hidden state. An interim embedding calculated for the last hidden state may be normalized to determine a final embedding representing features extracted from the input data by the RNN.

Abstract: User identities can managed at an organization level, instead of across multiple individual resource accounts. In a resource provider environment, access to various resources and services may require users to have identities with specific resource accounts. Users can instead be associated with organization accounts, or virtual accounts that are not associated with specific resources or services. The organization accounts are attached at the appropriate location(s) in an organizational hierarchy. A user having an organization account can project the identity in any sub-account in the organization hierarchy. This can include any lower-level resource account, or can child accounts under a relevant branch of the hierarchy. A user can validate against the organization account, and receive access to the relevant service or resources using the identity projected in the corresponding resource account.

Abstract: Techniques for reducing the latency of content retrieval from a content delivery network include receiving a request from a client device for media content, parsing the request for attributes associated with the request and the client device, and providing the attributes to a machine learning model to perform server-side prediction of an estimated retrieval time of the media content. A quality level for the media content is determined based on the estimated retrieval time, and the requested media content is provided to the client device at the determined quality level.

Abstract: Systems and methods to measure and affect focus, engagement, and presence of users may include measuring a variety of aspects of users engaged in particular activities. Individual user characteristics or preferences and attributes of activities may be taken into account to determine levels of focus for particular users and activities. A variety of sensors may detect aspects of users engaged in activities to measure levels of focus. In addition, a variety of output devices may initiate actions to affect levels of focus of users engaged in activities. Further, a variety of processing algorithms, including machine learning models, may be trained to identify desired levels of focus, to calculate current levels of focus, and to select actions to change or boost levels of focus. In this manner, activities undertaken by users, as well as interactions between multiple users, may be made more engaging, efficient, and productive.