Abstract: Disclosed are systems and methods to compare two or more machine learning models to determine the comparative performance of those models. Markers may be assigned to data items and data item marker scores generated for those data items, independent of the machine learning models. Each of the machine learning models to be compared may then process the data items and generate respective model scores for those data items. A sub-set of the data items may then be generated for each machine learning model based on the model scores assigned to the data items by the respective model. A model marker score may then be computed for each machine learning model based on the marker scores assigned to each of the data items of the sub-set of data items determined for each model. Finally, the model marker scores may be compared to determine which machine learning model has the highest performance.

Abstract: This disclosure describes an electronic device that can be installed within an access control system in parallel with a legacy interface device. The electronic device may perform steps of determining a set of configuration settings associated with an access control system to which the electronic device is electrically connected, receiving, by the electronic device, a first signal in a format associated with the access control system, configuring, based on the set of configuration settings, one or more signal pathways within the electronic device, generating, by routing the first signal through the one or more signal pathways, a second signal in a second format, and providing the second signal to at least one second electronic device.

Abstract: Techniques for selecting a skill component to process a natural language input are described. When a natural language input is received, natural language understanding (NLU) output data representing the natural language input is generated, and skill components (capable of processing the NLU output data) are determined. Thereafter, rules (for preventing the invocation of skill components) are implemented in a tiered manner, resulting in the determination of a subset of the skill components. The subset of skill components is ranked using a machine learning model(s), and the top-ranked skill component is called to process the NLU output data.

Abstract: Methods and apparatus are described for high-resolution video encoding using spatial lapping to reduce visual discontinuities within video frames. A plurality of frame regions may be encoded to form a higher-resolution video frame. For example, a 7680×4320 video frame may be constructed from four 4096×2160 frame regions arranged into a 2×2 grid. The encoding of the frame regions may be done in parallel. The encoded frame regions may be arranged such that the frame regions lap one another at one or more overlapping regions. Encoding of border regions adjacent to the overlapping regions may be done based on an encoding parameter (e.g., quantization parameter) of pixels of an overlapping region to determine an encoding parameter of pixels of the border regions to reduce visual artifacts at seams between frame regions and overlapping regions.

Abstract: HTTP/HTTPS communication between non-terrestrial objects and ground stations are enabled by splitting common HTTP/HTTPS proxy functions across two sides of a radio frequency (RF) communication link. An HTTP/HTTPS request may be generated and then transmitted across an RF communication downlink to a ground location, where the HTTP/HTTPS request may be analyzed to determine an appropriate endpoint. A connection is established with the endpoint to permit transmission of the request. In response to the request, the ground location may then take an HTTP/HTTPS response and transmit across an RF communication uplink to the non-terrestrial object.

Abstract: Autonomous ground vehicles that are outfitted with onboard cameras and configured to travel within an area or region are programmed with a previously generated visual map of the area or region. The visual map includes a plurality of map points from which images were previously captured and processed to determine visual features of such images. In the event of a loss of one or more position signals, such as GPS signals, the autonomous ground vehicles capture images using one or more of the cameras, and process the images to determine visual features from such images. The visual features determined from such images are compared to the visual map. Positions of one or more of the map points are identified as most likely corresponding to a position of the autonomous ground vehicle where visual features associated with such map points are similar to the visual features determined from such images.

Abstract: Disclosed are various embodiments for on-demand application-driven network slicing. In one embodiment, first data is transmitted to or from a first application in a particular computing device using a first network slice in a communications network. A request is sent from the particular computing device to configure a second network slice in the communications network based at least in part on a requirement of a second application. Second data is transmitted to or from the second application using the second network slice.

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

Abstract: Techniques are described for enabling users to create and configure request-based policies used by a data storage system to generate custom responses to particular types of requests for users' stored data objects. The request-based policies include rules used to identify types of requests for data objects and corresponding types of responses to be generated responsive to detecting such requests. The types of responses that can be specified in such policies include responses other than simply returning a requested data object or denying a request. For example, a rule for certain types of responses can instruct the data storage system to return a dynamically generated data object, another rule can instruct the system to return a different data object than the one requested, and so forth.

Abstract: A technique for processing strong ordered transactions in a direct memory access engine may include retrieving a memory descriptor to perform a strong ordered transaction, and delaying the strong ordered transaction until pending write transactions associated with previous memory descriptors retrieved prior to the memory descriptor are complete. Subsequent transactions associated with memory descriptors following the memory descriptor are allowed to be issued while waiting for the pending write transactions to complete. Upon completion of the pending write transactions, the strong ordered transaction is performed.

Abstract: Accelerated convolution of neural networks can be performed by executing N computing engines (CEs) of a neural network processor in parallel. An input dataset can be divided spatially into N chunks such that a respective last portion of each chunk overlaps with a respective first portion of a subsequent chunk. Portions of each chunk can be processed by a respective CE to generate a respective portion of an output dataset. The overlapping intermediate states computed by each CE from processing the overlapping portion can be stored locally for sharing with a subsequent CE using an on-chip bus.

Abstract: An apparatus includes a hub, a carrier structure and a first plurality of connective elements. The carrier structure surrounds the hub and holds a plurality of rollers that rotate within the carrier structure about an axis that is orthogonal to an axis of rotation of the hub. The first plurality of connective elements are coupled to the hub and the carrier structure. A connective element of the first plurality of connective elements elastically deforms to absorb an impact experienced by the carrier structure such that the rollers can rotate within the carrier structure while the carrier structure experiences the force.

Abstract: A group masked autoencoder may be implemented for anomaly detection. An autoencoder network model may be trained without supervision and applied to output an estimated joint probability distribution of normality for a group of frames of time series data. The estimated joint probability distribution may be used to determine an anomaly score for the time series data. An anomaly may be detected according to the anomaly score and a result that indicates a detected anomaly may be provided.

Abstract: This disclosure describes, in part, techniques for identifying interactions and events associated with inventory locations. For instance, system(s) may receive image data representing a user interacting with an inventory location. The system(s) may then generate heatmap data indicating a first portion of the image data that represents the inventory location and feature data indicating a second portion of the image data that represents the user. Next, the system(s) may analyze the heatmap data with respect to the feature data to determine that the second portion of the image data corresponds to the first portion of the image data. As such, the system(s) may determine that the user is interacting with the inventory location. Based on the determination, the system(s) may analyze the first portion of the image data to identify an event that occurs at the inventory location, such as the user removing an item.

Abstract: Systems and methods are disclosed for a shuttle track system having a calibration shuttle for determining portions of the track having non-conforming geometry for optimum use of the shuttle track system. For example, the shuttle track system may include a track having an electromagnetic propulsion system (e.g., linear synchronous motor) embedded between two sides of the track. The calibration shuttle may move to various locations on the track and may use imaging systems for generating image data of a portion of the track. A trained model may process the image data to determine edges and surfaces of the components of the track and may determine distances between reference points on the track. The distances may be compared to threshold values to determine if they are in conformity with desired track geometry.

Abstract: Methods and systems are disclosed. The methods and systems receive partner metadata and third party metadata, and generate lifecycle data and catalog data based on the partner metadata and the third party metadata. The methods and systems also receive event data and generate broadcastable consumer data based at least in part on the partner metadata, the third-party metadata, or the event data. The broadcastable consumer data includes partner metadata modified by the computing device.

Abstract: An aerial vehicle is configured to calculate ranges to objects around the aerial vehicle when operating within indoor spaces, using a LIDAR sensor or another range sensor. The aerial vehicle calculates ranges within a plurality of sectors around the aerial vehicle and identifies a minimum distance measurement for each of the sectors. Sets of adjacent sectors having distance measurements above a threshold are identified, and bearings and minimum distance measurements of the sets of adjacent sectors are determined. When the aerial vehicle detects an object within a flight path, the aerial vehicle selects one of the sets of adjacent sectors based on the minimum distance measurements and executes a braking maneuver in a direction of the selected one of the sets of adjacent sectors.

Abstract: Techniques for an electronic device to perform power-stealing techniques to harvest energy from a power-control circuit to power components of the electronic device. In some examples, the electronic device may be connected in the power-control circuit between a power supply and a relay that is selectively configured to activate a power load. According to the techniques described herein, the electronic device may include voltage-drop circuitry that is connected in the power-control circuit such that a voltage drop is produced across electrical components of the electronic device while the relay is in the activated, or triggered, state. In this way, the electronic device may perform power-stealing from the power-control circuit while the relay is maintained in the activated state.

Abstract: Techniques for conversational-based searching are described. A system may receive a first spoken user input, and may determine that the first spoken user input corresponds to a request for information associated with an entity. The system may retrieve item results corresponding to the entity. The system may determine a suggested user input based on the retrieved item results, and may determine output data corresponding to the suggested user input. The system may send output data to a user device, where the output data includes the item results and the suggested user input. The system may receive a second spoken user input, and may determine that the second spoken user input corresponds to the suggested user input. In response, the system may send the previously determined output data to the device.

Abstract: Techniques are described for pre-processing frames to apply a temporal filter. A P-frame may be temporally filtered based on neighboring reference frames. The temporal filter may be applied to blocks of the P-frame to improve alignment with dependent frames that are encoded based on blocks of the P-frame.

Abstract: When a customer places an order for a first item and a second item, delivery systems utilize an aerial vehicle to transport the first item from a first source of the first item to a second source of the second item. At the second source, the first item and the second item are transferred into a carrier vehicle and transported toward a destination of the order. Once the carrier vehicle arrives within a vicinity of the destination, an autonomous ground vehicle departs the carrier vehicle with the first item and the second item, and delivers the first item and the second item to the destination. The delivery systems effectively expand the capacity of sources of items, and simplify processes for delivering orders for items, by enabling items from multiple sources to be combined at a single source prior to delivery, and transported to a destination from the single source.

Abstract: A quantum computing service includes a quantum circuit compilation module that uses received information about quantum processing units (QPUs) along with information about a quantum circuit to be compiled in order to generate an optimized ordered list of compilation passes to be performed to compile the quantum circuit for execution on a given QPU.

Abstract: Techniques for using multiple machine learning (ML) models, with varying compute costs, for ASR processing is described. The system may include an arbitrator component configured to determine which ML model is to be used to process an audio frame from a sequence of audio frames representing a spoken natural language input. The arbitrator component may switch between the ML models, on a frame-by-frame basis, to reduce an overall compute cost for the entire spoken natural language input. The outputs of the different ML models may be combined to determine the final output for the entire spoken natural language input.

Abstract: A set of alternative vocal input styles for specifying a parameter of a dialog-driven application is determined. During execution of the application, an audio prompt requesting input in one of the styles is presented. A value of the parameter is determined by applying a collection of analysis tools to vocal input obtained after the prompt is presented. A task of the application is initiated using the value.

Abstract: An aerial vehicle configured for operation within indoor spaces has a meshed construction with a housing defined by upper and lower sections having meshes provided above and below propellers and motors of the aerial vehicle. The aerial vehicle also includes a suite of sensors such as LIDAR sensors, time-of-flight sensors, cameras, ultrasonic sensors, or others. Meshes of the upper and lower sections include central openings along with spokes and concentric rings provided about the central openings. Meshes of the lower section have substantially larger central openings than meshes of the upper section, but feature more dense spokes or concentric rings beneath tips of the rotating propellers, which may be hinged or foldable in nature. Data captured by sensors of the aerial vehicle may be utilized for any purpose, such as to generate environment maps of an indoor space, or to monitor the indoor space for adverse conditions or events.

Abstract: Systems and methods are disclosed for robotic picking assemblies with highly damped suction cups. In one embodiment, an example suction cup for a picking assembly may include a shell formed of an elastomeric material, where the shell forms a first bellow. The suction cup may include a first fluid chamber disposed on an outer portion of the suction cup, a first opening, where fluid from the first fluid chamber passes through the first opening, and an optional second fluid chamber disposed on an inner portion of the suction cup. The suction cup may include an optional second opening formed in the shell, where fluid from the second fluid chamber passes through the second opening. The first fluid chamber and the second fluid chamber may be configured to dampen movement of the suction cup.

Abstract: Techniques and systems can obtain a first private key usable with a classical cryptography algorithm and a second private key usable with a post-quantum cryptography algorithm based on classical and post-quantum public keys hosted by a computer-implemented storage of an online service provider. A plurality of keys to perform a cryptography operation on data hosted by the computer-implemented storage can be generated, the plurality of keys generated based on at least the first and second private keys and a cryptography derivation function identified in the computer-implemented storage. The plurality of keys can be used to perform the cryptography operation on the data hosted by the computer-implemented storage.

Abstract: Techniques for evaluating a natural language understanding (NLU) component and determining an action to resolve an issue processing a user input are described. The system determines which component is invoked by a baseline NLU component is processing the user input, and which component is invoked by an updated NLU component. Based on that information, the system selects the action to resolve the updated NLU component generating an undesired response to the user input.

Abstract: Users in a facility may move about and their authentication status may change accordingly. For example, upon entering a facility, a user may present identification data. At that point, the user is considered an authenticated user with persistent profile data. However, the identification of a user may not be presented at entry or may be unreliable, which causes the user to be considered an unauthenticated user with temporary profile data. In other scenarios, the identification of authenticated users may become unreliable based on their location or their proximity to other users. In cases where users are crowded at a location, the profile data for such users may be aggregated and may be based on common settings of the set of possible users in the group. A user interface may be presented to a user based on their designated profile data.

Abstract: Systems, methods, and computer-readable media are disclosed for modular robotic manipulation systems. In one embodiment, an example modular robot assembly may include a housing having a base, and a robotic manipulator coupled to the base, where the housing is configured to provide the robotic manipulator access to a first side, a second side, and a third side of the modular robot assembly. The module robot assembly may include a first camera system, a controller, and an optional display coupled to the housing. The modular robot assembly may be configured to be coupled to other modular robot assemblies, and the housing may be configured to be secured, such that the modular robot assembly can be independently transported.

Abstract: Systems, methods, and computer-readable media are disclosed for capacitive touch sensing using system-in-package components and batteries. In one embodiment, a device may include a flexible printed circuit, a button cell electrically coupled to the flexible printed circuit, and a system-in-package that is electrically coupled to the flexible printed circuit. The system-in-package may include an electromagnetic interference shield, and a capacitive touch sensor. The capacitive touch sensor may be configured to detect a change in capacitance via a change in electric field at the electromagnetic interference shield.

Abstract: A quantum computing service may use multiple quantum computers to execute a same quantum computing algorithm to improve the quantum computational accuracy. The quantum computing service may instruct individual ones of the multiple quantum computers to execute the quantum computing algorithm repeatedly for a number of times. The quantum computing service may obtain a plurality of results from the multiple quantum computers. The quantum computing service may aggregate the plurality of results to generate an ensemble result, and provide the ensemble result to a customer as a final result of the quantum computing algorithm.

Abstract: Systems and methods are described for execution of multiple tasks associated with a set of code in an on-demand network code execution system. A user may provide a set of code that is associated with the multiple tasks. The system may generate a first virtual machine instance for execution of a first task. The system may determine that a second task is associated with the first task and may identify a location of the first virtual machine instance. The system may further identify a second virtual machine instance for execution of the second task based on the location of the first virtual machine instance. For example, the system may identify the first virtual machine instance from a plurality of pre-generated virtual machine instances and/or may generate the first virtual machine instance.

Abstract: On-circuit utilization monitoring may be performed for a systolic array. A current utilization measurement may be determined for processing elements of a systolic array and compared with a prior utilization measurement. Based on the comparison, a throttling recommendation may be provided to a management component to determine whether to perform the throttling recommendation.

Abstract: Systems and methods to generate synthetic lip synchronization may generate source facial keypoints based on a source video, generate target facial keypoints based on a target audio, determine distances between the source facial keypoints and target facial keypoints, and transform or warp the source facial keypoints and associated surfaces to the target facial keypoints. In this manner, target video having synthetic lip synchronization that matches the target audio may be generated, and the target video may substantially preserve or maintain surface textures or features from the source video in the target video, thereby generating natural and believable synthetic lip synchronization corresponding to the target audio.

Abstract: This disclosure describes techniques for updating planogram data associated with a facility. The planogram may indicate, for different shelves and other inventory locations within the facility, which items are on which shelves. For example, the planogram data may indicate that a particular item is located on a particular shelf. Therefore, when a system identifies that a user has taken an item from that shelf, the system may update a virtual cart of that user to indicate addition of the particular item. In some instances, however, a new item may be stocked on the example shelf instead of a previous item. The techniques described herein may use sensor data generated in the facility to identify this change and update the planogram data to indicate an association between the shelf and the new item.

Abstract: An audio/video (A/V) device includes a first housing and a second housing rotatably coupled to the first housing. The first housing includes a camera oriented in a first direction, one or more microphones oriented in the first direction, one or more first lighting elements oriented in the first direction, one or more second lighting elements oriented in a second direction, one or more third lighting elements oriented in a third direction, and one or more first contacts. The second housing includes a loudspeaker oriented in a fourth direction and one or more second contacts configured to engage with the one or one first contacts. An engagement between the one or more first contacts and the one or more second contacts communicatively couples the loudspeaker to computing components with the first housing.

Abstract: A visualization tool for machine learning models obtains metadata from a first training node at which a multi-layer machine learning model is being trained. The metadata includes a parameter of an internal layer of the model. The tool determines a plurality of metrics from the metadata, including respective loss function values corresponding to several training iterations of the model. The tool indicates the loss function values and the internal layer parameter values via a graphical interface.

Abstract: An input/output (I/O) device can initiate data migration of a virtual machine (VM) instance from a source device to a target device. The data migration of the VM instance may include migrating the data for the VM instance and tag data associated with the data. The data for the VM instance and the tag data may be stored together in a source memory. A first read request from the I/O device can enable a memory controller in the source device to read the data for the VM instance and the tag data together, store the tag data in a tag data buffer, and transmit the data for the VM instance to the target device. A second read request from the I/O device can read the stored tag data from the tag data buffer and transmit to the target device. The target device can write the data for the VM instance together with the tag data in the target memory.

Abstract: When creators generate media content in accordance with media programs, the media content is evaluated to identify any number of violations of policies, and to generate scores representing a level of risk that the creators will violate one or more of the policies in the future. Subsequently, media content of the creators is transmitted to listeners in accordance with the scores. In addition to audio data of creators or transcripts of the audio data, scores may be generated based on images associated with the creators, titles or summaries of media programs, or reports received from listeners. Scores calculated for creators may increase or decrease over time, depending on numbers of violations of policies by such creators, or other factors, and be utilized with a goal of protecting listeners against exposure to harmful content.

Abstract: Techniques for detecting regressions with respect to the accuracy of an anomaly detection compute service in detecting anomalies in users' time series data. The techniques include providing an instrumented time series instrumented with a set of one or more anomalies to the anomaly detection service. The anomaly detection service detects a set of one or more anomalies in the instrumented time series. The precision and recall of the detected anomalies with respect to the instrumented anomalies is computed. From the computed precision and recall, an anomaly detection accuracy is computed as an F-score or F-measure. It is then determined whether a regression in anomaly detection accuracy has occurred by comparing the computed accuracy score to a threshold. If a regression has occurred, an alert can be generated or a recent change to the anomaly detection service can be rolled back.

Abstract: Described technologies generate a data structure corresponding to values sequenced based on a plurality of timestamps associated with the values. The data structure can include a first section identifying a first timestamp associated with the plurality of timestamps and a number representing how many timestamps are associated with the plurality of timestamps, and a second section including at least a value linked to the first timestamp, and an additional value representing an encoding type associated with the second section. The data structure can be stored in computer-implemented storage.

Abstract: A device including a mount, a first arm coupled to the mount, and a second arm coupled to the mount. The first arm includes first teeth and the second arm includes second teeth. A first light assembly rotationally couples to the first arm, and includes a first coupler that engages with the first teeth. A second light assembly rotationally couples to the second arm, and includes a second coupler that engages with the second teeth. A camera assembly pivotably couples to the mount, and includes a camera, a first passive infrared (PIR) sensor, and a second PIR sensor.

Abstract: A system to manage database queries including storage devices to implement a data store to store database data and computing devices to implement a query engine. The query engine is configured to receive, from a client, a database query and initiate performance of the query at the data store. The query engine is configured to compare a performance time of the query with a performance time threshold. Based on the performance time exceeding the performance time threshold: send a query identifier for the query and a token indicating that the query has not been completed; and receive, from the client, an additional query comprising the token and the query identifier. Based on a determination that the performance time does not exceed the performance time threshold, send a response to the query to the client, the response comprising data requested by the query.

Abstract: Techniques for configuring device-specific skills as top-level skills are described. When a system receives a user input the system performs NLU processing to determine an intent of the user input. In some instances, the system may identify a device-specific skill associated with the device interacted with by the user. At least partially in parallel to performing NLU processing to determine the intent of the user input, the system may also perform NLU processing to determine a likelihood that the user input corresponds to an intent actionable by the device-specific skill. Once the system has finished NLU processing, the system may implement one or more prioritization rules to determine whether the user input should be sent to the device-specific skill or another skill of the system.

Abstract: Methods and systems for automatically processing a document may include classifying a document, such as a medical document, as one or more document types based at least in part on one or more machine learning models and one or more tokens extracted from the medical document, determining a token contribution weight of each token towards the classification, modifying the medical document based on the token contribution weights of the one or more tokens, and displaying the modified medical document on a display to a user.

Abstract: This disclosure describes, in part, devices and methods for deactivating tags. For instance, an electronic device may include antennas that transmit signals and/or fields for deactivating the tags. In some examples, the electronic device includes a first gate that includes a first portion of the antennas and a second gate that includes a second portion of the antennas. To deactivate the tags, the electronic device may detect locations of an object relative to the electronic device. The electronic device may then activate a respective tag and/or a respective group of tags based on the location of the object. For example, the electronic device may activate a first tag and/or a first group of tags when the object is at a first location, activate a second tag and/or a second group of tags when the object is at a second location, and/or so forth.

Abstract: Techniques are described for monitoring and analyzing input/output (I/O) messages for patterns indicative of ransomware attacks affecting computer systems of a cloud provider, and for performing various remediation actions to mitigate data loss once a potential ransomware attack is detected. The monitoring of I/O activity for such patterns is performed at least in part by I/O proxy devices coupled to computer systems of a cloud provider network, where an I/O proxy device is interposed in the I/O path between guest operating systems running on a computer system and storage devices to which I/O messages are destined. An I/O proxy device can analyze I/O messages for patterns indicative of potential ransomware attacks by monitoring for anomalous I/O patterns which may, e.g., be indicative of a malicious process attempting to encrypt or otherwise render in accessible a significant portion of one or more storage volumes as part of a ransomware attack.

Abstract: A computer-implemented method for providing cryptographic services, including providing key pairs. A key management service receives a web service application programming interface or other such request to generate a key pair. To respond to the request, the key management service obtains a pregenerated key pair that is securely stored and provides the key pair in response to the request.