Abstract: The present disclosure provides computer-implemented systems and processes for augmenting user ratings of items, such as a rating of a song playing on a user device, by analyzing contextual information, such as user and/or device activity data associated with the device. The contextual information may be used to determine an associated intensity of the rating or feedback. The determined intensity levels can be used to weigh the associated rating events and improve the quality of item recommendations that are based on such ratings. Contextual information may indicate whether the user transitioned from another application to rate an item, device status information, and so on. In one embodiment, contextual information can be used to assess how intensely the user feels about a music item and to provide improved music recommendations, such as songs to provide in a playlist or radio station, based on the intensity.

Abstract: Multi-modal natural language processing systems are provided. Some systems are context-aware systems that use multi-modal data to improve the accuracy of natural language understanding as it is applied to spoken language input. Machine learning architectures are provided that jointly model spoken language input (“utterances”) and information displayed on a visual display (“on-screen information”). Such machine learning architectures can improve upon, and solve problems inherent in, existing spoken language understanding systems that operate in multi-modal contexts.

Abstract: A technology is described for selecting a distributed network service based at least in part on consistency, availability, and partition tolerance (CAP) metrics. An example method may include receiving a client request for a listing of distributed network services and associated CAP metrics that are within a range of at least one CAP specification included in the client request. In response to the client request, a directory service may be queried for a set of distributed network services having the CAP metrics that are within the range of the least one CAP specification included in the client request. A listing of distributed network services that includes the CAP metrics for the distributed network services may be generated from the set of distributed network services, and the listing of distributed network services and CAP metrics may be returned in response to the client request.

Abstract: A beamformer system includes an in-ear audio device, such as an earbud, that has three microphones. Two microphones may be disposed on an external face of the audio-device; one microphone may be disposed in or near the ear canal of a user. Data from two microphones is phase-adjusted and combined; a reference signal is generated by phase-adjusting and removing data from the two microphones from data from a third microphone. The combined data is filtered using the reference signal to remove any residual echo, and the resulting data may be used for communications, speech processing, or other uses.

Abstract: A system (and method) includes a plurality of compute devices configured to execute an endpoint node and a provisioning service. The endpoint node is configured to establish an encrypted communication channel over a public network. The provisioning service is configured to retrieve configuration parameters from a database. The configuration parameters define a security association for the encrypted communication channel and include an encryption key and an identifier of an encryption algorithm. The provisioning service is configured to transmit the configuration parameters to the endpoint node for use in implementation of a security association for the encrypted communication channel.

Abstract: The disclosure relates to provisioning honeypot computing services using a simulation state database to simulate a set of computing resources. One example includes a system that receives a mutating request associated with honeypot credentials, updates a simulation state database associated with the honeypot credentials at least based on the mutating request and generates a simulated mutating response based at least on the simulation state database that simulates a response to the mutating request. The system can also receive a query request associated with the honeypot credentials, query the simulation state database at least based on the query request, and generate a simulated query response based at least on the simulation state database that simulates a response to the query request.

Abstract: A data-collecting device acquires data associated with a real-time data stream and transmits the data to a data-consuming service hosted on a server computer system in the form of a multipart response. The multipart response includes one or more data content parts and at least one authentication content part. Each of the one or more data content parts contains data representing part of the real-time data stream. Each authentication content part includes authentication information usable to verify the integrity of the data transmitted in the data content parts transmitted prior to the authentication content part.

Abstract: A system and method of deploying applications to networked computing environments is disclosed. Application packages and installation instructions are received by a deployment manager. The installation instructions include test instructions. Application packages are distributed in response to requests by developers and installation instructions executed on a target computing environment. Failure of an installation test provides an indication of issues with the application in the target environment. Records of failures received are stored for use by application developers.

Abstract: In some examples, a vision system includes multiple time of flight (ToF) cameras and a single illumination source. The illumination source and the multiple ToF cameras may be synchronized with each other, such as through a phase locked loop based on a generated control signal. A first one of the ToF cameras may be co-located with the illumination source, and a second one of the ToF cameras may be spaced away from the illumination source and the first ToF camera. For instance, the first ToF camera may have a wider field of view (FoV) for generating depth mapping of a scene, while the second ToF camera may have a narrower FoV for generating higher resolution depth mapping of a particular portion of the scene, such as for gesture recognition.

Abstract: In a resource-on-demand environment, dynamically created server instances are allowed to boot from encrypted boot volumes. Access keys to the boot volumes are provided from a key provider that authenticates new instances based on possession of a security token that has been previously shared between the key provider and the new instance through an out-of-band communication.

Abstract: A system and method for a wearable device capable of detecting a wake gesture for purposes of capturing and forwarding audio data corresponding to a spoken utterance. The device may wake for purposes of capturing utterance audio data in response to a combination of a wake gesture and wakeword. The wake gesture may enable a wakeword detector. The device may also attempt to detect a wakeword utterance in a noisy environment. In response to determining the noisy environment, the device may receive motion data from a motion sensor; determining the motion data corresponds to a wake gesture, and send the audio data corresponding to an utterance to a remote device for processing. The device may also wake based on a combined confidence of wakeword and wake gesture detection.

Abstract: A distributed voice controlled system has a primary assistant and at least one secondary assistant. The primary assistant has a housing to hold one or more microphones, one or more speakers, and various computing components. The secondary assistant is similar in structure, but is void of speakers. The voice controlled assistants perform transactions and other functions primarily based on verbal interactions with a user. The assistants within the system are coordinated and synchronized to perform acoustic echo cancellation, selection of a best audio input from among the assistants, and distributed processing.

Abstract: According to one or more embodiments of the disclosure, a method is provided. The method may include receiving, by at least one server comprising one or more processors, from a user device, a request for a web page comprising product detail information associated with a focus product identifier. The method may also include receiving an advertisement request associated with the focus product identifier. Furthermore, the method may include determining, based at least in part on the focus product identifier, a product sample identifier associated with a product sample to offer with the product. Additionally, the method may include selecting, based at least in part on the product sample identifier, a product sample advertisement.

Abstract: Shared electronic resources can be allocated to customers by auction when there is contention among the customers for the resources. Each customer can receive a bid pool for a shared electronic resource. A customer may prioritize a request by withdrawing a bid amount from the customer's bid pool and submitting the bid amount with a request for the shared resource. A resource provider may assess the capacity of the shared resource to process requests and conduct an auction at various times, such as during periods of congestion, to determine the requests that the shared resource will process at a given time. Customers can demand a refunded bid amount when the auction price is less than the customer's bid amount, and the customer can issue a repudiation challenge if no auction was held or the customer did not receive access to the shared resource.

Abstract: Systems and methods are described to enable health checking of computing devices within a virtual private cloud (VPC) networking environment, without requiring that the devices be accessible via a public network address. An endpoint is placed within the VPC, which enables interaction with an external health checking system via a substrate network. The endpoint handles communications between the heath checking system and the VPC, and can modify data originating from the health checking system such that it appears to originate from the endpoint. From the viewpoint of the VPC, the endpoint itself may appear to be conducting health checking. Thus, external health checking can be used on a VPC without compromising the security of the VPC by requiring that a portion of the VPC be externally addressable.

Abstract: The present disclosure is directed to the generation of a compact artificial neural network by removing individual nodes from the artificial neural network. Individual nodes of the artificial neural network may be deactivated randomly and/or selectively during training of the artificial neural network. In some embodiments, a particular node may be randomly deactivated approximately half of the time during processing of a set of training data inputs. Based on the accuracy of the results obtained when the node is deactivated compared to the accuracy of the results obtained when the node is activated, an activation probability may be generated. Nodes can then be selectively removed from the artificial neural network based on the activation probability.

Abstract: The present disclosure is directed toward the use of two or more autonomous vehicles, working in cooperation, to deliver an item between a source location and a destination location. For example, an autonomous ground based vehicle may transport an item from a source location to a transfer location and an autonomous aerial vehicle will transport the item from the transfer location to the destination location. The transfer location may be at any location along a navigation path between the source location and the destination location. In some examples, the transfer location may be adjacent the destination location such that the autonomous aerial vehicle is only transporting the item a short distance.

Abstract: A certificate authority receives a request to issue a digital certificate from a customer. In response to the request, the certificate authority determines a network endpoint to be specific to the digital certificate that is to serve information usable to determine whether the digital certificate is valid. The certificate authority issues, to the customer, a digital certificate that specifies a network address for the network endpoint and records information about requests made to the network endpoint to obtain the information usable to determine whether the digital certificate is valid.