Abstract: Disclosed are various embodiments for evaluating performance metrics (e.g., accuracy, depth precision, curvature accuracy, coverage, data acquisition time, etc.) of sensors (e.g., cameras, depth cameras, color cameras, etc.) according to captured image data. One or more reference boards having different types of reference structures (e.g., three-dimensional shapes, materials, three-dimensional patterns (e.g., waves, steps, etc.), gaps, etc.) that are used to evaluate the performance properties of a sensor. A reference board is attached to a robotic arm and positioned in front of a sensor. The robotic arm positions the reference board in front the sensor in different viewpoints while the sensor captures image data associated with the reference board. The captured image data is compared with ground truth data associated with the reference board to determine performance metrics of the sensor.

Abstract: Images of a hand are obtained by a camera. A pose of the hand relative to the camera may vary due to rotation, translation, articulation of joints in the hand, and so forth. Avatars comprising texture maps from images of actual hands and three-dimensional models that describe the shape of those hands are manipulated into different poses and articulations to produce synthetic images. Given that the mapping of points on an avatar to the synthetic image is known, highly accurate annotation data is produced that relates particular points on the avatar to the synthetic image. An artificial neural network (ANN) is trained using the synthetic images and corresponding annotation data. The trained ANN processes a first image of a hand to produce a second image of the hand that appears to be in a standardized or canonical pose. The second image may then be processed to identify the user.

Abstract: Systems and methods for configuration for contextual response to motion-based events are disclosed. For example, a predefined event may be detected. The system may determine that a subject associated with the event is of a given type of subjects and may determine whether another person is present in the environment associated with the event to assist in providing aid. The presence of the other person may be detected, a type of the other person may be performed, and a notification may be output on a second device associated with the other person based on the type. The system may track when the other person reach an area associated with the event and may cause the first device to perform an action to assist the other person in providing aid.

Abstract: Techniques for authenticating users at devices that interact with the users via voice input. For instance, the described techniques may allow a voice-input device to safely verify the identity of a user by engaging in a back-and-forth conversation. The device or another device coupled thereto may then verify the accuracy of the responses from the user during the conversation, as well as compare an audio signature associated with the user's responses to a pre-stored audio signature associated with the user. By utilizing multiple checks, the described techniques are able to accurately and safely authenticate the user based solely on an audible conversation between the user and the voice-input device.

Abstract: Online block parsing techniques are performed that use anchor tokens. When a log entry of a log stream is received, the log entry may be tokenized into an ordered series of log units. The log units may be matched with anchor tokens in previously initialized log templates to determine a matching log template. A log template based on the matching log template may be provided with the tokenized log entry to a destination for a parsed version of the log stream.

Abstract: A supplemental content retrieval system receives requests for supplemental content from various content providers, and makes respective requests to supplemental content provider(s). Responses to the requests may be subject to a time limit. The requests to the supplemental content provider(s) may be made in parallel or serially, and may sometimes include information based on responses from another content provider. The supplemental content retrieval system receives responses with content items from the supplemental content providers and, for items in the responses that do not specify a category, determines categories of the items (e.g., based on item metadata, interactive links in the content item, item identifiers for the items, etc.). The supplemental content retrieval system applies a constraint (e.g., an inter-item category constraint or similar, which also may be associated with a time limit) to the categorized items to prepare responses to respective content provider requests.

Abstract: Code may be dynamically routed to computing resources for execution. Code may be received for execution on behalf of a client. Execution criteria for the code may be determined and computing resources that satisfy the execution criteria may be identified. The identified computing resources may then be procured for executing the code and then the code may be routed to the procured computing resources for execution. Permissions or authorization to execute the code may be shared to ensure that computing resources executing the code have the same permissions or authorization when executing the code.

Abstract: This disclosure describes a natural language question (NLQ) query service within a service provider network that provides row level security (RLS) for autocomplete during entry of NLQs and fuzzy matching in NLQ answering. The rules take the form of per-user predicates such as Tim can only see rows with region=US. In configurations a complex extraction and preprocessing pipeline to extract distinct combinations of values against RLS predicate “rule keys” is used. Those distinct values are indexed along with grouped rule keys to enable pushing down predicates at auto-complete time. This enables pushing part of RLS rule handling to ingestion time of a dataset rather than handling all RLS rule handling at query time, enabling meeting of latency goals. In configurations, a single logical document of unique cell values is split into multiple documents with a subset of rule keys to handle scalability limits.

Abstract: Systems and techniques for performing a calibration test of an obstacle detection sensor of a drive unit are described. An example system includes a calibration region and at least one drive unit. The calibration region is disposed in an environment and includes a calibration apparatus. The at least one drive unit is operable to autonomously move throughout the environment. The at least one drive unit includes a sensor configured to detect one or more objects within the environment. The at least one drive unit is configured to autonomously transition to the calibration region upon detecting at least one predetermined condition, and perform a calibration test of the first sensor at the calibration region using the calibration apparatus.

Abstract: Systems and methods are disclosed for rapid PCR testing. Example embodiments may include a PCR testing module that includes a housing having a PCR machine disposed therein; a sample input station on the housing, wherein the sample input station is configured to receive a sample collection device (SCD) comprising a biological specimen sample provided by the patient; an SCD processing mechanism configured to transfer a lysed microportion of the biological specimen sample into a PCR sample tube attached to the SCD; at least one mechanism configured to separate the PCR sample tube from the SCD and transfer the PCR sample tube to the PCR machine; and a controller configured to (i) use the PCR machine to conduct a PCR test on contents of the PCR sample tube, and (ii) generate results of the PCR test.

Abstract: The systems and methods disclosed herein can increase the speed and ease with which a user can discover compatible replacement parts for a specified model of a product by automatically determining when the user is interested in replacement parts. A replacement part system can conclude that a user is looking for replacement parts based on the user's search queries, the user's browsing behavior, or the user's prior purchases. By automatically deducing an intention to purchase a replacement part, the replacement part system can make it easier for a user to locate suitable replacement parts. In addition, the replacement part system can be configured to tailor listing set of replacement parts search results it provides to a user. The set of search results can be tailored based on a number of criteria to increase the speed with which the user finds a compatible replacement part.

Abstract: A subscription storage service of a provider network may be used to select a particular datastore to store a topic subscription record for an IoT device (e.g., after the IoT device subscribes to an MQTT topic). The service may select the particular datastore based on the type of the subscription (e.g., non-wildcard vs. wildcard subscription) and one or more criteria associated with the topic (e.g., subscribe operation TPS for the topic). This may allow the service to store wildcard subscriptions to a different structure and/or datastore (e.g., a “trie” structure of an in-memory data store), while offloading the storage of non-wildcard subscriptions to another type of datastore (e.g., a key-value store) that has lower performance and/or operational cost to store data.

Abstract: A file system manager implemented at a provider network identifies a storage device of a first group of storage devices of a provider network as an initial location of a file system object. Based on an access metric associated with the object, the file system manager initiates a transfer of contents of the object to a second storage device of a different storage device group, without receiving a client request specifying the transfer. In response to an access request received via a file system programmatic interface, contents of the object are provided from the second storage device. Based on a second access metric, the object is transferred back to the first group of storage devices.

Abstract: Described are systems and methods directed to determining a topology for the efficient allocation and distribution of a service to customers that are distributed over a geographic area. For example, the topology can facilitate the planning, scheduling, and allocation of resources for the efficient provisioning of services to customers at locations distributed over the geographic area. Certain input parameters and/or constraints associated with the provisioning of the service may be processed to define a cost function that may represent a total cost of providing the service while ensuring that none of the constraints are exceeded. The cost function may be optimized to determine a lowest relative cost for providing the services to the customers using the various mechanisms and/or types of resources, and a topology associated with the optimized cost function can be generated to specify how the service should be provided to satisfy the demand.

Abstract: Devices and techniques are generally described for detection of abusive computational nodes. In various examples, first input data describing a first plurality of computational nodes and first data identifying a dimension along which to parse the first plurality of computational nodes may be received. A first computing device may generate input graph data representing the first plurality of computational nodes. The computational nodes of the first plurality of computational nodes may share a same value for the dimension are connected to one another in the input graph data. In various examples, a first graph machine learning model and at least one known abusive computational node may be used to determine a first set of candidate computational nodes for further evaluation. In some cases, network access of a first computational node of the first set of candidate computational nodes may be terminated.

Abstract: Systems, devices, and methods are provided for cloud-based privacy controls. User content is encrypted using a content encryption key (CEK). The CEK may be double-encrypted by the data producer—the inner envelope is encrypted using keys associated with privacy domains that are authorized to access the user content. The outer envelope is encrypted using a cloud privacy control's public key. When a data consumer requests access the user content, the cloud privacy control evaluates privacy policies and determine whether access should be permitted. If permitted, the cloud privacy control decrypts the outer envelope and provides the inner envelope with CEK to the requestor. Upon receiving the inner envelope, the data consumer may then decrypt the inner envelope with its privacy domain private key to obtain the CEK. The CEK may then be used to perform a decryption and obtain the user content.

Abstract: A computing system includes a rack having standard slots for computing devices and computing devices coupled to the rack. One or more of the computing devices includes a chassis, a circuit board assembly in a primarily horizontal orientation, and one or more processors coupled to the circuit board assembly. One or more stacks of hard disk drives are coupled to the chassis. The chassis has a width that is equal to or less than half of the width of one of the standard slots of the rack.

Abstract: A boot process for a computing device, such as integrated circuit, includes security features that are inaccessible during certain operation modes. An image including permission to access those security features is received during the boot process and may be verified using one or more keys. In operation, access to the security features is permitted during the operation modes after the image is verified. Such an approach enables a boot process to permit access to certain features after receipt and verification of different images.

Abstract: Systems, methods, and computer-readable media are disclosed for determining virtual product placement opportunities in a media content and determining product candidates for virtual insertion into the media content. The product placement system may determine shot segments from the media content and for each shot segment may determine candidate product placement locations. The product placement system may determine contextual information from the shot segments and from the contextual information determine candidate products suitable for the product placement locations. The product placement system may determine total screen time for each product placement opportunity as well as quality of each opportunity. For each product and product placement opportunity, the product placement system may determine an expected revenue and a projected insertion cost.

Abstract: Various mounting systems for mounting heat sink apparatus to bare die processors are disclosed. The mounting systems include an upper plate, which may include a heat transfer portion, positioned in proximity to the upper surface of bare die processors to provide heat conduction away from the processors. The disclosed mounting systems secure the upper plate to the processors with balanced and centralized forces to inhibit tilting of the upper plate and reduce the risk of damaging the processors.

Abstract: A resource-based authorization control may be added to a data processing service. A control-plane action requested by an account or entity creates a resource-based authorization policy. The resource-based authorization policy pertaining to respective resources allows recurrent data-plane actions to maintain a data flow in a data pipeline to be authorized independent of a separate authorization service.

Abstract: A unique identifier of an expression which refers to a cell of a data sheet is stored in a repository. In an application comprising an operation based on evaluating the expression, the expression is indicated by the unique identifier. In response to detecting a write directed to the cell, a modified version of contents of the cell is stored without re-evaluating the expression. The expression is re-evaluated in response to determining that the operation of the application is to be implemented.

Abstract: Techniques are described for managing access of executing programs to non-local block data storage. In some situations, a block data storage service uses multiple server storage systems to reliably store network-accessible block data storage volumes that may be used by programs executing on other physical computing systems. A group of multiple server block data storage systems that store block data volumes may in some situations be co-located at a data center, and programs that use volumes stored there may execute on other physical computing systems at that data center. If a program using a volume becomes unavailable, another program (e.g., another copy of the same program) may in some situations obtain access to and continue to use the same volume, such as in an automatic manner in some such situations.

Abstract: Techniques for managing access to sensitive data in transcriptions are described. A method for managing access to sensitive data in transcriptions may include receiving a request to generate a redacted transcript of content, obtaining a transcript of the content, sending at least a portion of the transcript to a model endpoint to identify sensitive entities in the transcript, receiving an inference response identifying one or more sensitive entities in the transcript, and generating the redacted transcript based at least one the transcript and the inference response.

Abstract: A materialized view management service (MVMS) is capable of monitoring resource allocation changes of a source data object at a source data store and responsively generating resource change alerts to the owner of a target data object (the materialized view) in the target data store. Resource allocation changes may include autoscaling changes to the source data object's partition scheme, throughput limit, storage limit, and the like. The MVMS generates resource change alerts in response to these detected events and pushes the alerts to interested subscribers. Depending on the embodiment, the alerts may be pushed to human administrators, or the target data store itself, which may be configured to automatically adjust the resource allocation of the target data object to match the source data object. Advantageously, the disclosed alerts allow view owners to gain real time visibility of resource auto-scaling at the data source and appropriately react to such changes.

Abstract: Techniques are described for providing customizable sign-on functionality, such as via an access manager system that provides single sign-on functionality and other functionality to other services for use with those services' users. The access manager system may maintain various sign-on and other account information for various users, and provide single sign-on functionality for those users using that maintained information on behalf of multiple unrelated services with which those users interact. The access manager may allow a variety of types of customizations to single sign-on functionality and/or other functionality available from the access manager, such as on a per-service basis via configuration by an operator of the service, such as co-branding customizations, customizations of information to be gathered from users, customizations of authority that may be delegated to other services to act on behalf of users, etc.

Abstract: Techniques are described for managing communications between multiple computing nodes, such as computing nodes that are separated by one or more physical networks. In some situations, the techniques may be used to provide a virtual network between multiple computing nodes that are separated by one or more intermediate physical networks, such as from the edge of the one or more intermediate physical networks by modifying communications that enter and/or leave the intermediate physical networks. 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 (e.g., users of a program execution service). The managing of the communications may include determining whether communications sent to managed computing nodes are authorized, and providing the communications to the computing nodes only if they are determined to be authorized.

Abstract: Systems and methods are disclosed for securing fabric to speaker-based devices. An example device may include a housing having a first portion devoid of apertures, and a second portion having a first number of apertures. The device may include a fabric disposed about the housing, and an adhesive disposed between the fabric and the housing. The adhesive may include a first portion disposed between a first portion of the fabric and the first portion of the housing, and a second portion disposed between a second portion of the fabric and the second portion of the housing. The first portion of the adhesive may have a patterned non-liquid material that is bonded, and the second portion of the adhesive may have the patterned non-liquid material that is unbonded.

Abstract: In one example, a pallet jack cap system includes a stabilization outrigger secured around steer wheels of a pallet jack, a bonnet in an elevated position over forks of the pallet jack, an anchor base for attachment to the pallet jack, and a vertical extension system between the anchor base and the bonnet. The vertical extension system supports the bonnet in the elevated position over the forks of the pallet jack. The vertical extension system can include an elevating vertical support assembly, an elastic counterpoise to assist with positioning the bonnet at a particular elevation based on an extension of the vertical support assembly, and an extension lock to secure the vertical support assembly at the particular elevation. The bonnet can be lowered down over a unit load on a pallet transported by the pallet jack, to secure the unit load and items in transport by the pallet jack.

Abstract: Idle periods may be for management actions at processing clusters for managed databases. A leader node of a processing cluster for a managed database may monitor a network endpoint at a proxy service associated with a database managed by the database service. An idle period for the database may be detected. A management action for the processing cluster may be determined to be performed during the detected idle period. The leader node may cause the determined management action to be performed.

Abstract: An operating system, or operating system update, management service uses a shared read-only multi-attach volume of a block-based storage service to distribute operating systems or operating system updates to a set of virtualized computing instances. Also, to store launch specific information, that is specific to a given virtualized computing instance, additional writable volumes are used, wherein a write volume is attached to each of the computing instances of the set. This eliminates the need to provide a full copy of an OS volume to each of the computing instances.

Abstract: Techniques for graduated navigation displayed on an electronic device used in servicing items for users are described herein. For example, the electronic device can receive a first set of directions for moving towards a first geolocation associated with a service address using a first navigation method for presentation. The electronic device can transition from the first set of directions to a second set of directions in response to detecting a transition trigger associated with a second geolocation. The transition can involve ceasing providing the first set of directions for presentation. The transition can also involve providing the second set of directions for moving from the second geolocation towards the service address using a second navigation for presentation.

Abstract: Methods for improving performance of tiered storage of a data processing service by detecting and reducing thrashing of data blocks between warm and cold storage tiers are disclosed. In order to understand the frequency of hits by incoming queries to data blocks that are not currently stored in the warm storage tier, the elapsed time between query hits to the respective data blocks may be tracked using timers. Times below a given amount of time may be used to indicate thrashing. For example, recently evicted data blocks that are subsequently hit by a query within a short amount of time since eviction may indicate thrashing. In scenarios in which thrashing may be occurring, a threshold corresponding to the number of times a given data block in the cold storage tier receives a query hit before being added to the warm storage tier may be turned on.

Abstract: Technologies are disclosed for using hardware-embedded monitors to monitor pages of local memory and detect attribute violations or other unauthorized operations relating to the memory. The attribute violations may include mismatches of attributes (e.g., designating a page as writeable versus executable or vice versa) in entries in a translation buffer that point to a same physical address or other mismatches between designations of attributes for a page in physical and virtual space. Responsive to detecting a violation, an alert or other mitigation protocol, which may include an audit of activities surrounding the violation, may be performed.

Abstract: Techniques for performing a machine learning model based spatial-temporal adaptive shift for end-to-end text-video retrieval are described.

Abstract: A hardware card that can be implemented in a server system is disclosed. The hardware card may be, for instance, a network interface card coupled to a board such as a motherboard or backplane in a computing system. In various instances, the hardware card reports a temperature to a management controller on the board (e.g., a baseboard management controller) that is adjusted due to thermal degradation of the hardware card. For example, the temperature output by the hardware card may be adjusted based on a comparison of a current temperature profile (including transient temperature of the hardware card) to a predetermined temperature profile stored on the hardware card.

Abstract: A container may include a cargo bay for holding items; a lid able to slide between a closed position to an open position relative to the cargo bay; a check strap operatively connected to the lid and able to move vertically in a first direction toward the lid as the lid slides from the closed position to the open position, and to move vertically in a second direction opposite the first direction as the lid slides from the open position to the closed position; and arms operatively attached to rollers and to one or more tension devices, wherein when the check strap moves vertically in the first direction, the check strap separates the rollers and causes the arms to pull the one or more tension devices, and wherein when the check strap moves vertically in the second direction, the one or more tension devices cause second rotation of the arms.

Abstract: Contents of client-initiated handshake messages of a security protocol are obtained at a handshake processing offloader configured for an application. The offloader uses a first security artifact (which is inaccessible from a front-end request processor of the application) and the contents of the handshake messages to generate a second security artifact. The second security artifact is transmitted to the front-end request processor, which uses it to perform cryptographic operations for client-server interactions of the application.

Abstract: Some implementations include methods generating visualization emphasis object for players and may include receiving a video clip associated with a sporting event participated by a plurality of players using a playing field with a play object. The players in a frame of the video clip may be detected. Players who are on the playing field may be identified from the detected players. Each of the players identified to be on the playing field may be associated with a rectangular bounding box that provides an outline of each of the players. A player who has possession of the play object may be identified. A visualization emphasis object may be generated and placed on the player who has possession of the play object. The visual emphasis object may have a size proportional to height of a bounding box associated with the player having the possession of the play object.

Abstract: Some embodiments provide a method that receives a specification of a neural network for execution by an integrated circuit. The integrated circuit includes a neural network inference circuit for executing the neural network to generate an output based on an input, an input processing circuit for providing the input to the neural network inference circuit, a microprocessor circuit for controlling the neural network inference circuit and the input processing circuit, and a unified memory accessible by the microprocessor circuit, the neural network inference circuit, and the input processing circuit. The method determines usage of the unified memory by the neural network inference circuit while executing the neural network. Based on the determined usage by the neural network inference circuit, the method allocates portions of the unified memory to the microprocessor circuit and input processing circuit.

Abstract: Techniques for Meta-Q-Learning (MQL) are described. A method of MQL may include receiving a request from an agent to perform adaptation based at least on task data associated with a new task collected by the agent, identifying a subset of meta-training data corresponding to the task data in a replay buffer, and adapting a policy using the subset of meta-training data and the task data to generate an adapted policy, wherein the adapted policy is used identify a next action for the agent to perform.

Abstract: Methods and apparatuses for improving network packet transmission performance in terms of latency with reduced packet retransmission times and fewer packet drops in congested networks are provided. Packet-switched networks can experience long delays while waiting for out-of-order packets or re-transmissions of lost packets. In addition, network faults such as transmission path failures can result in excessive delay while attempting to find a new route over which to transmit packets. To improve packet transmission performance, application data may be encapsulated into first network packets by a first transport protocol having an interface exposed to the application, in the first network packets may be encapsulated into second network packets according to a second transport protocol. The second transport protocol can enable the second network packets of a same packet flow to be transmitted across multiple paths over the network.

Abstract: Systems and methods are provided for automated modification of delivery parameters. Particularly, computing model that is trained to determine a probability that a delivery defect is likely to occur for a given delivery or set of deliveries. Based on the probability, various limitations associated with the deliveries may be activated or deactivated on a mobile device application used by a delivery driver to perform the deliveries. The systems and methods reduce the number of delivery defects that occur while simultaneously reducing the use of unnecessary guardrails for low-risk deliveries. The model may be queried in real-time such that guardrails for a delivery itinerary may be optimized prior to the delivery driver beginning the delivery route.

Abstract: Approaches are described for image signal processor day-night detection. An exemplary approach for a camera with an infrared light source involves determining a brightness value for image data based on an average luma value for the image data, determining white balance data for the first image data, and determining, based on the white balance data for the image data and white balance data for the infrared light source, a deviation value for the image data. A transition from night mode to day mode can be effected based on comparing the brightness value for the image data to a first threshold, and comparing the deviation value for the image data to a second threshold.

Abstract: A content consumption system or device may implement device-enabled identification for automated user detection. An identifying device may be detected at a content consumption device as within proximity of the content consumption device. An identifying device may be a mobile or wearable computing device, in various embodiments. A user account associated with the identification device may be selected for accessing content at the content consumption device. Access to content may be provided according to the selected user account. In some embodiments, content recommendations or content filtering may be performed based on the automatically determined user account.

Abstract: Technologies for mesh path selection are described. One method of operating a first wireless mesh device includes determining, using first client information about a first link, the first link using a first frequency range, a first metric representing airtime usage on the link, and a second metric representing a first amount of data on a second link, the second link using the first frequency range. The method determines a third metric on a third link, the third link using a second frequency range. The method determines, using the metrics, a first throughput associated with a first path with the second link and a second throughput associated with a second path. The method determines that the third link results in better throughput and sends or receives first data on the third link.

Abstract: Techniques for computer system-based conversations are described. In an example, a system receives, from a first device, first data corresponding to a first interaction in a conversation that requests a function. The system causes the first device to output a first response to the first interaction. Prior to an execution of the function, the system determines that the conversation is to be paused and causes the first device to output a first indication that the conversation is paused. Upon determining that the conversation is to be resumed, the system causes a second device to output a second indication that the conversation is resumed. The second device can be the same or different from the first device. The system receives, from the second device, second data corresponding to a second interaction in the conversation and causes the execution of the function based at least in part on the second data.