Abstract: In one example, an apparatus comprises: a memory array having an array of memory elements arranged in rows and columns, each memory element being configured to store a data element; and a memory access circuit configured to: perform a row write operation to store a first group of data elements at a first row of the array of memory elements; perform a column read operation at a first column of the array of memory elements to obtain a second group of data elements; and perform a column write operation to store a third group of data elements at the first column of the array of memory elements to replace the second group of data elements.

Abstract: A system receives one or more transcripts of communications between entities. The system identifies a requested action in the communications based at least in part on a mapping between the requested action and an application programming interface. The system identifies one or more statements eliciting information, based on parameters to the application programming interface. The system generates a definition of an artificial agent based, at least in part, on the requested action and the one more statements eliciting information.

Abstract: Systems and methods are provided for bypassing or enforcing constraints regarding actions on network-based services. An account manager can store or maintain an account type of standard or special as account information associated with an account. When a user operating under an account tries to perform an action on a network-based service, the network-based service may retrieve or receive the account type associated with the account from the account manager. If the account type is standard, the network-based service can enforce constraints on the action. Alternatively, if the account type is special, the network-based service can bypass constraints on the action.

Abstract: Techniques for chatbot utterance routing in a provider network include jointly training a service classifier and a plurality of auxiliary classifiers based on a mixed service set of labeled chatbot utterance training examples to yield a trained service classifier. When a particular chatbot user utterance is received, the trained service classifier can be used to determine if the utterance is in-scope or out-of-scope, and if in-scope, to determine which service of a set of services in the provider network to which to route the utterance for further processing. By jointly training the service classifier with the auxiliary classifiers, the accuracy of the in-scope/out-of-scope determination by the trained service classifier is improved as well as its accuracy in routing the utterance to the appropriate service for processing the utterance as intended by the user.

Abstract: Natural language-based question answering systems and techniques are generally described. In some examples, a natural language processing system, first natural language data. The natural language processing system may determine first slot data included in the first natural language data. A set of content items associated with the first slot data may be determined. A first machine learning model may use the first natural language data to generate prediction data associated with a first attribute among a list of attributes of the set of content items. In some examples, a first value associated with the first attribute for a first content item of the set of content items may be determined. Second natural language data may be generated based at least in part on the first value. The second natural language data may include a response to the first natural language data.

Abstract: Systems, devices, and methods are provided for implementing a cloud-based mainframe service. A cloud-based mainframe service may utilize various resources, including an operating system that is provisioned with an authorization interceptor that uses a first set of security policies stored in a policy database to determine whether to grant or deny access to resources managed by the operating system. The authorization interceptor may use the security policies of the policy database to determine whether to grant access to operating system resources. A database management system may use a second set of security policies stored in the policy database to determine whether to grant or deny access to resources managed by the database system. Security policies for a mainframe service may be centrally stored in a policy database managed by a policy management service.

Abstract: Techniques are generally described for highlight recovery for image sensors. In some examples, a first frame of image data comprising a first pixel may be received. The first pixel may include a value Ci,c where c?(R, G, B, IR). A white balance gained value Ci,cwb may be determined using Ci,c·WBc, where WBc represents a per-color channel white balance gain. A color-corrected value Ciccm may be determined based at least in part on Ci,cwb and a color correction matrix. An interpolation coefficient ? may be determined based at least in part on a clipped representation of the color-corrected value Ciccm. A second value Ciout may be determined for the first pixel by interpolating between an unclipped representation of the first pixel and an unclipped representation of Ciccm using ?.

Abstract: A global segmenting and analysis service of a provider network may receive documents (e.g., posts, product reviews) from different applications. The service may analyze the documents to identify target entities and sentiment. The service may generate different levels of sentiment data and store data into a segmented database. For example, the service may store within-document level sentiment, document-level sentiment, and multi-document level sentiment for a target entity. The service may also update the entity taxonomy automatically or with only a small number of sample documents. The client may query the service for the segmented sentiment data.

Abstract: The transcoding of a contribution feed into a plurality of output feeds in various formats can be monitored to ensure that unnecessary data is not transmitted in the contribution feed. Each output feed can be transcoded using respective values for a set of video format parameters. These values can be aggregated and analyzed to determine the lowest values for individual parameters that are being used for the various output feeds. A video encoder for the contribution feed can then dynamically modify the video format parameters used to encode the contribution feed in order to avoid encoding and transmitting data that is not actually used for these output streams, which can conserve resources such as network bandwidth, or enable those resources to be used more advantageously to send data that will actually result in higher quality video presentation in the output formats for current limitations or conditions.

Abstract: Evaluation data sets may be programmatically generated for code generation models. An evaluation data set is obtained that includes items that correspond to different evaluation tests for a code generation system. The individual items of the evaluation data set maybe converted, including the conversion of a function signature for the items, the test statements for the items and using a code generation system to generate the body of the function.

Abstract: Systems and methods are provided to implement a database connection manager that dynamically routes database query statements to database servers storing replicas of the database. In embodiments, the database connection manager may implement a data access interface that mimics the interface of the database servers, so as to appear to clients as just another database server. In embodiments, various health or performance metrics of the servers are monitored, and the routing may be performed based on the metrics to load balance the servers. In embodiments, the routing may distinguish between read and write statements, so that read statements are routed to servers hosting read-only replicas, and write statements are routed to servers hosting writable replicas. These and other disclosed features improve the functioning of the database, to allow, among other things, the database servers to be dynamically load balanced, and the servers' identities to be hidden from clients.

Abstract: A system for improved natural language understanding (NLU) provides pre-feature input to a named entity recognition (NER) component. Pre-features may include no-textually derived information associated with the circumstances describing a user command (such as time, location, etc.). A domain and/or intent may also be determined prior to NER processing and may be passed to the NER component as an input. The pre-features and/or domain or intent information may assist the NER processing by providing context to a textual input, thereby improving NER processing such as semantic tagging, which in turn may improve overall NLU processing quality.

Abstract: Domain-specific parameters may be used for tuning speech processing. A pre-trained transformer-based language model may train domain-specific parameters using domain-specific unlabeled text data. This domain-specific parameters can then be appended to candidate texts produced by a speech model on received speech data and input to the transformer-based language model to score the candidate texts. The scores of the candidate texts determined using the pre-trained transformer-based language model can then be used to select a candidate text for further speech processing.

Abstract: Techniques for an optimization service of a service provider network to help optimize the selection, configuration, and utilization, of virtual machine (VM) instance types to support workloads on behalf of users. The optimization service may implement the techniques described herein at various stages in a life cycle of a workload to help optimize the performance of the workload, and reduce underutilization of computing resources. For example, the optimization service may perform techniques to help new users select an optimized VM instance type on which to initially launch their workload. Further, the optimization service may monitor a workload for the life of the workload, and determine new VM instance types, and/or configuration modifications, that optimize the performance of the workload. The optimization service may provide recommendations to users that help improve performance of their workloads, and that also increase the aggregate utilization of computing resources of the service provider network.

Abstract: A host device may include an interconnect, a host memory, and a set of processor cores. A processor core may execute a VM assigned to a cryptographic key and may send a request to access a physical address in the host memory toward the interconnect. An enforcer device may receive the request and extract a key identifier from the request. The enforcer device may determine whether to allow the request to access the physical address via the interconnect based on the key identifier and a list of allowed keys stored on the enforcer device. If the enforcer device determines to not allow the request to access, the enforcer device may modify the physical address and/or the key identifier of the request.

Abstract: In one embodiment of the disclosure, packets are routed according to a VRF (Virtual Routing and Forwarding) domain that is represented by a VRF identifier in a VRF-identifier (ID) field of packet headers. In one embodiment of the disclosure, a VRF identifier may be added to a VRF-ID field of a packet header so that the packet is routed according to a VRF domain that the packet belongs to.

Abstract: Devices, systems, and methods are provided for managing advertisement exposure. The method may include receiving, by a first device, from a second device, an audio fingerprint based on presentation of first media content using a first media channel at a first time; determining, using automatic content recognition, that the audio fingerprint represents a portion of a first advertisement; determining that the first advertisement was presented at a second time using a second media channel and the second device; determining that a first number of times that the first advertisement has been presented to the user exceeds a first threshold number and that a second number of times that the first advertisement has been presented to the user does not exceed a second threshold number; refraining from selecting the first advertisement at a third time; and selecting the first advertisement for presentation at a fourth time.

Abstract: Disclosed are various embodiments for using unmanned vehicles to manage radio-based network infrastructure. In one embodiment, an unmanned vehicle is instructed to deliver a radio unit to a location to add coverage to a radio-based network.

Abstract: A computing resource services provider receives user diagnostics comprising analysis operations to be performed on one or more computing resources. The computing resource services provider monitors resources indicated by the user diagnostics. When one or more computing resources exceed a threshold value or otherwise indicate abnormalities in computing resource usage, the computing resource services provider performs one or more analysis operations to determine one or more causes for increased computing resource usage. Results of the one or more analysis operations are presented to a user once the computing resource services provider completes the one or more analysis operations.

Abstract: Systems and methods are provided for managing provision of network-stored data sets among instances of function code executing in an on-demand manner. Provision of the data sets may be managed such that the data sets are available efficiently and are able to be accessed within function invocations in a manner that is similar or identical to accessing locally-stored data sets. Moreover, the data sets may, when desired, be made available persistently across function invocations.

Abstract: The present disclosure describes a system and method for implementing a virtual shoe try-on feature. The method includes predicting, using a machine learning model, a first plurality of pixels including pixels of an image corresponding to a foot in the image and adding a first virtual cylindrical object to a three-dimensional model of a shoe such that a portion of the first virtual cylindrical object is positioned within the three-dimensional model of the shoe. The method also include determining a set of pixels including pixels of the first plurality of pixels that intersect with the first virtual cylindrical object and removing, from the three-dimensional model of the shoe, a first portion of the three-dimensional model of the shoe that intersects with the set of pixels.

Abstract: Neck worn devices and related methods monitor posture of a user, activity of the user, and/or an environment of the user and provide feedback. A neck worn monitoring device includes a support frame, a feedback assembly, one or more sensors, and a control unit. The feedback assembly is operable to output feedback. The one or more sensors are configured to generate one or more sensor outputs indicative of at least one of a posture of the user, an activity of the user, or an environment of the user. The control unit is configured to process the one or more sensor outputs to monitor at least one of the posture, the activity, or the environment, and operate the feedback assembly to transmit a feedback indicative of at least one of the posture, the activity, or the environment.

Abstract: An encoding of a cryptographic key is obtained in a form of an encrypted key. Request is provided to a service provider including a fulfillment involving performing a cryptographic operation on data. Upon fulfillment of the request, a response is then received which indicates the fulfillment of the request.

Abstract: A storage service supports attachment of multiple clients to a distributed storage object and further supports persistent reservations that govern types of access the respective clients are granted with respect to the distributed storage object. In order to efficiently distribute reservation state changes to multiple partitions of the distributed storage object hosted by different data storage units/servers, existing connections are used between the data storage units/servers hosting the partitions of the distributed storage object and the connected clients to propagate reservation state changes.

Abstract: Systems and techniques are disclosed for a buffer management system that dynamically and automatically adjusts operational buffers in resources based on resource utilization based on buffer parameters. Buffer parameters for a resource can include a target utilization, a buffer capacity, a minimum buffer capacity, and a maximum buffer capacity. The buffer management detects changes in utilization, such as a number of operations performed in a time period and determines whether a candidate buffer capacity determined using the buffer capacity parameter and the current utilization is between the minimum buffer capacity and the maximum buffer capacity. If so, the buffer management system sets the buffer capacity to the candidate buffer capacity, otherwise the buffer management system sets the buffer capacity to the minimum or maximum buffer capacity as applicable.

Abstract: Described herein is a system for embedding audio watermarks. To improve performance without a user perceiving the audio watermark, a system embeds audio watermarks in audio data using scaling factors that are calculated based on a spectral masking level for each frame of the audio data. The scaling factors may vary over time and correspond to an amplitude of the audio watermark across a series of watermark frames. The system processes the audio data to determine a spectral mask, which represents an amount of energy perceived in a first frequency range that is caused by energy represented in neighboring frequency ranges. By selecting scaling factor values that keep an amplitude of the audio watermark below the threshold indicated by the spectral mask, the system may embed the audio watermark in the first audio data without the audio watermark being audible to the user.

Abstract: Described are automated systems and methods for employing certificate authority meta-resources to facilitate automatic renewal and/or rotation of certificates and/or certificate authorities in a PKI hierarchy. For example, embodiments of the present disclosure can provide creating a certificate authority meta-resource, which can maintain and monitor certain information to facilitate automatic renewal and rotation of certificates and/or certificate authorities in a PKI hierarchy. The certificate authority meta-resource can also keep track of the active certificate authorities and certificates to ensure that trust is maintained without manual configuration of the PKI hierarchy.

Abstract: An interposer card and a virtualization offloading card are provided for installation in a third-party server to integrate the third-party server into a cloud service provider network. The interposer card includes a baseboard management controller that interfaces with a management console of the cloud service provider network. This allows the third-party server to be converted into a server controlled by the cloud service provider network. Additionally, the baseboard management controller of the interposer card acts as a firewall between the third-party server and a management control network of the cloud service provider network. The interposer card and the virtualization offloading card are installed in a chassis of the third-party server via an expansion slot without requiring modification of the hardware or firmware of the third-party server.

Abstract: Systems, methods, and computer-readable media are disclosed for shuttle lift systems to facilitate simultaneous movement of multiple shuttles across a multi-level storage structure that stores packages. The shuttle lift system may include the multi-level storage structure and a tower lift system at each end of the multi-level storage structure. One tower lift system may raise shuttles to a level of the multi-level storage structure. The other tower lift system may lower shuttles to a level of the multi-level storage structure and/or to a delivery area. The tower lift system may include cables (e.g., tethers) that are supported by a bottom portion and an upper portion of the tower lift system. Shelves may be attached to the cables on either side of the tower lift system and may receive and move shuttles up/down the tower lift system. The tower lift system may have high shuttle throughput and a reduced footprint.

Abstract: Variable reclamation of data copies can be performed. Copies of data may be stored in a data store. The copies of the data may be available to restore or return the data to a state described in or based on one of the copies. One or more copies, or portions of a copy, may be selected for invalidation according to a variable rate. The variable rate may be determined based on accesses to the data. Selected copies, or portions of a copy, may then be invalidated.

Abstract: Approaches in accordance with various embodiments provide an overlay network comprised of interconnected peer instances, which can be used for purposes such as data multicasting, high performance computing, or massive parallel processing. The overlay network can sit on top of a routing layer or fabric of interconnected peer devices that can support large message sizes and/or provide for equal cost, multipath routing. Such an overlay network can provide for high bandwidth data transmission in a cloud or shared-resource environment, as well as for isolated networks that may have otherwise limited gateways.

Abstract: Methods, systems, and computer-readable media for a certificate management system with forced certificate renewal are disclosed. The certificate management system may receive a request to renew a digital certificate. The request may be received at a selected time prior to an automatic renewal date for the certificate, and the automatic renewal date may be stored by the certificate management system. The certificate management system may acquire, based at least in part on the request to renew the certificate, a renewed certificate from a certificate authority. The renewed certificate may be obtained prior to the automatic renewal date. The renewed certificate may be exported from the certificate management system and bound to a computing resource (e.g., a server) prior to the automatic renewal date.

Abstract: A semiconductor package can include a capacitance die. The package can have multiple dice (e.g., logic die, memory die) mounted on a substrate. Each die can include a power domain. The dice can be distributed on the substrate such that an extra space is present on the substrate between at least some of the dice. For example, an extra space may be present between two dice, at a corner of the substrate, or other locations. The extra space can disrupt a coplanarity of the semiconductor package. The capacitance die can be located in the extra space so as to establish the coplanarity with the other dice. The capacitance die can include a capacitor array electrically coupled to multiple power domains of the plurality of dice.

Abstract: Systems, methods, and computer-readable media are disclosed for systems and methods for automated identification and mapping of objects in video content. Example methods may include determining a first set of frames in video content, determining, using one or more object recognition algorithms, a first object present in the first set of frames, determining that a first product corresponding to the first object is present in a product catalog comprising a set of product images, associating a first product identifier of the first product with a video identifier of the video content, and causing presentation of a set of product identifiers associated with the video identifier.

Abstract: Systems and methods described herein provide for novel configuration features for setting up a user device automatically to connect to a network and register the user device to a user account. A mapping of user devices to cryptographic keys for the user devices may be maintained by a computer system. The computer system may receive information that specifies network information and user account information for a particular user device. A mapping of the network information and user account information to the particular user device may be generated. A machine-readable code that includes the network information for connecting the particular user device to the network and a token that includes credentials for the user account information may be generated and transmitted to a assistant configuration device.

Abstract: A system configured to use keywords to augment audio playback with visual effects and/or other effects to provide an immersive audio experience. For example, a device can detect a keyword and control a color and intensity of external lights to provide visual feedback. In addition to visual effects, the device can trigger additional effects using smart plugs or other smart devices. In a listening enhancement mode in which the device outputs audio content, the device performs keyword detection by monitoring playback audio data for preconfigured keywords. In a storytelling mode in which a user reads a book out loud, the device may perform keyword detection by monitoring microphone audio data for the preconfigured keywords. Controlling the visual effects in response to keyword detection is enabled by a new pipeline that sends information back from a wakeword engine to an audio processor.

Abstract: Using a first set of machine learning models, a communication from a user of a restaurant is analyzed at an order coordinator linked via a network to resources of an order management service at a provider network. A response to the communication is prepared using another set of models at the provider network and presented to the user. An order of the user for one or more restaurant menu items is fulfilled, based at least partly on analysis of a second communication received from the user after the response is presented.

Abstract: Tidiness data associated with an inventory location, such as a shelf, indicates whether an item returned to the inventory location is designated for storage at the inventory location. When an item is returned to the inventory location and a user is determined to be near the inventory location, the weight change detected at the inventory location is compared to a corresponding weight of each item in the custody of the user, such as items within cart state data associated with the user. Differences in weight between the weight change and the corresponding weight of each item in the custody of the user are calculated. Such differences may then be compared to one or more metrics, such as threshold values, to determine how well the top candidate item in the custody of the user matches the type of item designated for storage at the inventory location.

Abstract: Described implementations obtain a proof of valid attestation data. The attestation data may include configuration data of a host computing system. A prover service may receive the attestation data. The prover service may generate a proof to prove that the attestation data includes valid configuration data of the host computer system, without revealing sensitive or private information of the host computing system. The proof may be a zero-knowledge proof.

Abstract: In some cases, a method may include receiving electricity usage data from a re-ordering device and determining usage information for an appliance that is electrically connected to the re-ordering device based on the electricity usage data. The method may also include determining, based on the usage information and order history information for a customer, an estimated remaining quantity of an item that is associated with an appliance type of the appliance. The method further includes comparing the estimated remaining quantity of the item to a threshold associated with the item. In response to determining that the threshold associated with the item is satisfied, the method includes decrementing the estimated remaining quantity of the item. In response to determining that the threshold associated with the item is not satisfied, the method includes generating order data for automated re-ordering of the item for delivery to the customer.

Abstract: Techniques are described for providing a web cookie firewall application capable of monitoring and controlling the use of web cookies at computing devices. The web cookie firewall application uses web cookie security policies to selectively allow or prohibit the storage or use of individual web cookies or types of web cookies. For example, web cookie security permissions can be defined by such policies to permit or deny the storage or use of web cookies associated with certain website domains, to permit or deny the storage or use of defined types of cookies, or to control access to web cookies based on other attributes. The web cookie firewall application can be implemented as a standalone application, as a web browser plugin or extension, or as any other type of software application that it is capable of monitoring and controlling the use web cookies on a computing device.

Abstract: A system comprises a plurality of computing devices. Requests received by the system are distributed at random among the computing devices. A computing device, in response to receiving a request, stores a record of utilization of the computing device by a source of the request. The computing device determines to throttle requests from the source based, at least in part, on the utilization of the computing device by the source within the time period.

Abstract: Results of applying a set of voting rules to a target corpus of documents are used to obtain a set of derived probabilistic labels indicating the probabilities of the presence of a particular attribute within the documents' constituent objects. A machine learning model is trained to identify a candidate portion of a document from which a value of the attribute is to be extracted. The training data for the model includes learned representations obtained from paths of constituent objects, and the corresponding derived labels. A proposed value for the attribute, obtained based on an assigned attribute value presence probability score for an individual constituent object from a selected candidate portion of a document, is provided.

Abstract: An anomalous period of operation of a database management system is detected by analyzing a time series of data points indicating the number of database queries pending processing by the system. Conditions associated with execution of the pending database queries are recorded and analyzed to identify conditions correlated with the anomalous period of operation. A recommendation for tuning the database is generated based on analysis of the conditions.

Abstract: A biometric identification system acquires a multimodal image. The system is trained to determine three embeddings: a first embedding associated with features present in a first modality, an intersection embedding associated with features present in both the first and a second modality, and an XOR embedding associated with features that are not shared between the first modality and the second modality. Once trained, the system may process a multimodal query image to determine query embedding data. Additional information, such as minutiae depicted in the multimodal query image may also be determined. The query embedding data, and in some implementations the additional information, may be compared with enrolled embedding data associated with a previously enrolled user. If the comparison exceeds a threshold value, an identity may be asserted.

Abstract: A plurality of events is extracted from a set of sources, including an event representing a transfer of a first data set from one data storage stage of a data pipeline to another stage to form a second data set, and another event representing a completion of a computation performed on the second data set. Based on analysis of the plurality of events, a graph is stored; the nodes of the graph represent data sets at respective stages of the data pipeline, and edges represent the events. In response to a request for lineage information pertaining to a particular data set at a particular stage of the pipeline, an indication of a sequence of events represented in the graph is provided, including a particular event which led to the presence of the particular data set at the particular stage.