Abstract: A virtual network provider system supports a virtual network including virtual machines that are each assigned to an underlay address of an underlay addressing scheme. The virtual network provider system further includes multiple routing domains each defined to include a different subset of the virtual machines. Each of the routing domains is assigned to a range of overlay addresses of an overlay addressing scheme. For each routing domain, the assigned range of overlay addresses is allocated among the subset of the virtual machines in the routing domain. The system further includes a virtual network host configured to use addresses of the overlay addressing scheme to selectively route messages between endpoints on select pairs of the virtual machines assigned to a same routing domain of the plurality of routing domains.

Abstract: Systems and methods for providing scene understanding can include obtaining a plurality of images, stitching images associated with the scene, detecting objects in the scene, and providing information associated with the objects in the scene. The systems and methods can include determining filter tags or query tags that can be selected to filter the plurality of objects, which can then be provided as information to the user to provide further insight on the scene. The information may be provided in an augmented-reality experience via text or other user-interface elements anchored to objects in the images.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: Described embodiments provide systems and methods for resource appropriation in a multi-tenant environment using risk and value modeling. A resource server can provide a plurality of applications access to a plurality of resources in response to requests from clients based in part on risk scores and value scores. The resource server can generate and execute a risk model and a value model to determine a risk score and a value score for each of the applications. The resource server can use the risk and value scores to determine access to a particular resource for a requested application. The resource server can assign a first allocation of resource tokens to an application. The resource tokens can correspond to access privileges to plurality of resources. The resource server can dynamically modify the resource allocation for applications responsive to changes to a risk score or value score of a respective application.

Abstract: Described embodiments provide systems and methods for resource appropriation in a multi-tenant environment using risk and value modeling. A resource server can provide a plurality of applications access to a plurality of resources in response to requests from clients based in part on risk scores and value scores. The resource server can generate and execute a risk model and a value model to determine a risk score and a value score for each of the applications. The resource server can use the risk and value scores to determine access to a particular resource for a requested application. The resource server can assign a first allocation of resource tokens to an application. The resource tokens can correspond to access privileges to plurality of resources. The resource server can dynamically modify the resource allocation for applications responsive to changes to a risk score or value score of a respective application.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: Methods, systems, and apparatus for content item auction bidding. In one aspect, a method includes receiving a request for a content item, the request including request feature values and a device identifier, the device identifier being included in a remarketing list; obtaining a predicted performance measure for a remarketing content item associated with the remarketing list based on a first timestamp and a second timestamp, the first timestamp being included in the remarketing list and associated with the device identifier included in the request, and the second timestamp being for the request; determining a bid adjustment value based on the first timestamp and the second timestamp; obtaining a remarketing bid for the remarketing content item, the remarketing bid specifying an amount a content item provider is willing to pay for distribution of the remarketing content item; and adjusting the remarketing bid based on the bid adjustment value.

Abstract: Described embodiments provide systems and methods for resource appropriation in a multi-tenant environment using risk and value modeling. A resource server can provide a plurality of applications access to a plurality of resources in response to requests from clients based in part on risk scores and value scores. The resource server can generate and execute a risk model and a value model to determine a risk score and a value score for each of the applications. The resource server can use the risk and value scores to determine access to a particular resource for a requested application. The resource server can assign a first allocation of resource tokens to an application. The resource tokens can correspond to access privileges to plurality of resources. The resource server can dynamically modify the resource allocation for applications responsive to changes to a risk score or value score of a respective application.

Abstract: A method for receiving training data for training a neural network (NN) to perform a machine learning (ML) task and for determining, using the training data, an optimized NN architecture for performing the ML task is described. Determining the optimized NN architecture includes: maintaining population data comprising, for each candidate architecture in a population of candidate architectures, (i) data defining the candidate architecture, and (ii) data specifying how recently a neural network having the candidate architecture has been trained while determining the optimized neural network architecture; and repeatedly performing multiple operations using each of a plurality of worker computing units to generate a new candidate architecture based on a selected candidate architecture having the best measure of fitness, adding the new candidate architecture to the population, and removing from the population the candidate architecture that was trained least recently.

Abstract: Methods and apparatus for reverse link acknowledgement in a wireless local area network. A method includes receiving, at a first node, a data communication over a common channel, the data communication being decodable by other nodes. The method also includes determining transmission resources from the data communication, wherein the transmission resources are different for each node, and transmitting a response over the common channel using the determined transmission resources. An apparatus includes a transmitter configured to transmit to a plurality of nodes a data communication over the common channel, and a receiver configured to receive responses from the plurality of nodes, wherein each response was sent using different transmission resources determined from the data communication.

Abstract: Techniques and apparatus for controlling the transmit power of an uplink (UL) signal from a user terminal in a wireless communications system in an effort to achieve some target characteristic, such as a target carrier-to-interference (C/I) ratio, at an access point (AP) are provided. In this manner, such a user terminal may help avoid or compensate for imbalances in received radio frequency (RF) power between UL signals received from multiple user terminals by the AP. For example, the transmit power at each user terminal may be controlled in an effort to achieve a target post-processing C/I ratio of 28 dB per spatial stream in an effort to reduce large power imbalances and optimize throughput per user terminal. The user terminal and the AP may compose part of a multiple-input multiple-output (MIMO) communication system utilizing spatial-division multiple access (SDMA) techniques.

Abstract: Apparatus and methods are directed to calibrating a delay within a wireless access point for determining a position of a mobile station. One method includes receiving an initial packet at an eavesdropping device, receiving a response packet, sent by another entity, at the eavesdropping device, computing a time difference based upon the packet arrival times, and providing the time difference to a position determination entity. Another method includes providing a request to appropriate eavesdropping devices to send information, receiving, from each appropriate eavesdropping device, a time difference which represents a difference in a time of arrival of a packet transmitted by the wireless access point and a time of arrival of a packet transmitted by the mobile station, determining a processing delay estimate based upon the time differences, and determining a position of the mobile station based upon the processing delay estimate and the received time differences.

Abstract: Techniques and apparatus for controlling the transmit power of an uplink (UL) signal from a user terminal in a wireless communications system in an effort to achieve some target characteristic, such as a target carrier-to-interference (C/I) ratio, at an access point (AP) are provided. In this manner, such a user terminal may help avoid or compensate for imbalances in received radio frequency (RF) power between UL signals received from multiple user terminals by the AP. For example, the transmit power at each user terminal may be controlled in an effort to achieve a target post-processing C/I ratio of 28 dB per spatial stream in an effort to reduce large power imbalances and optimize throughput per user terminal. The user terminal and the AP may compose part of a multiple-input multiple-output (MIMO) communication system utilizing spatial-division multiple access (SDMA) techniques.

Abstract: Apparatuses and methods for adjusting wireless-derived positions of a mobile station using a motion sensor are presented. One method includes estimating a position of a mobile station based upon wireless signal measurements and measuring a movement of the mobile station using a relative motion sensor. The method further includes detecting a displacement of the mobile station based upon the measured movement, determining that the displacement is below a threshold and then adjusting the estimated position of the mobile station using information from the relative motion sensor. An apparatus includes a wireless transceiver, a relative motion sensor, a processor coupled to the wireless transceiver and the relative motion sensor, and a memory coupled to the processor. The memory stores executable instructions and data for causing the processor to execute methods for adjusting wireless-derived positions using a motion sensor.

Abstract: Techniques are provided in which a mobile station may determine a first round-trip time (RTT) for a first measurement related communication between the transceiver and a first network device, wherein the first RTT comprises a uniform delay of time applied during the first measurement related communication by the first network device; Determine a second RTT for a second measurement related communication between the transceiver and a second network device, wherein the second RTT comprises the uniform delay of time applied during the second measurement related communication by the second network device; and determine, at least in part, a position of the mobile station with regard to at least the first and second network devices based, at least in part, on the first RTT and the second RTT.

Abstract: An example method for determining a processing time for wirelessly determining a position of a mobile station includes: measuring a round trip time delay to each of multiple wireless access points; estimating an initial processing time for each of the wireless access points; calculating the position of the mobile station based upon the measured round trip time delays and estimated processing times; and updating the initial processing time for each of the wireless access points based upon the calculated position of the mobile station.

Abstract: Techniques are provided in which a mobile station may determine a first round-trip time (RTT) for a first measurement related communication between the transceiver and a first network device, wherein the first RTT comprises a uniform delay of time applied during the first measurement related communication by the first network device; Determine a second RTT for a second measurement related communication between the transceiver and a second network device, wherein the second RTT comprises the uniform delay of time applied during the second measurement related communication by the second network device; and determine, at least in part, a position of the mobile station with regard to at least the first and second network devices based, at least in part, on the first RTT and the second RTT.

Abstract: Methods and apparatus for determining the quality of a communication link transmitting a specified packet type are disclosed. The methodology includes configuring first packets, which share transmission characteristics with a specific packet type, such as a Voice over IP packet, such that the first packets experience similar communication link treatment as the specific packet type. The first packets are then transmitted over the communication link to a quality monitoring server located near a termination equipment of the particular communication link. Second packets, which are transmitted by the server in response to the first packets, are received and evaluated to determine the quality of the communication link based on characteristics of the second packets. Corresponding apparatus are also disclosed.