Abstract: Embodiments are disclosed for a method for machine-learning model accuracy. The method includes generating prediction training data based on training predictions and corresponding probabilities of the training predictions. A classifier of a machine-learning model generates the training predictions. The method also includes training a prediction accuracy model to determine whether the training predictions generated by the machine-learning model are correct. Additionally, the method includes generating predictions in response to corresponding client transactions for the machine-learning model. Further, the method includes determining whether the predictions are accurate using the prediction accuracy model. Also, the method includes providing client predictions corresponding to the client transactions based on the determination.

Abstract: A computer-implemented method for refining dataset to accurately represent output of an artificial intelligence model includes generating a plurality of data points used to interpret a decision of an artificial intelligence model. A subset of data points from the generated plurality of data points satisfying one or more constraints is identified. A linear model is applied on the identified subset of data points satisfying the one or more constraints. One or more insights illustrating the decision of the artificial intelligence model is generated.

Abstract: A computer-implemented method for refining dataset to accurately represent output of an artificial intelligence model includes generating a plurality of data points used to interpret a decision of an artificial intelligence model. A subset of data points from the generated plurality of data points satisfying one or more constraints is identified. A linear model is applied on the identified subset of data points satisfying the one or more constraints. One or more insights illustrating the decision of the artificial intelligence model is generated.

Abstract: Embodiments are disclosed for a method for machine-learning model accuracy. The method includes generating prediction training data based on training predictions and corresponding probabilities of the training predictions. A classifier of a machine-learning model generates the training predictions. The method also includes training a prediction accuracy model to determine whether the training predictions generated by the machine-learning model are correct. Additionally, the method includes generating predictions in response to corresponding client transactions for the machine-learning model. Further, the method includes determining whether the predictions are accurate using the prediction accuracy model. Also, the method includes providing client predictions corresponding to the client transactions based on the determination.

Abstract: Various technologies described herein pertain to configuring an access point in a local network for operation in a community Wi-Fi network. The community Wi-Fi network includes access points in differing local networks that share respective network bandwidth capacities with member computing devices of the community Wi-Fi network. A request to join the access point to the community Wi-Fi network is transmitted from the access point to a community aggregation system. Responsive to the request, the community aggregation system transmits a firmware update to the access point. The access point installs the firmware update in memory of the access point. Moreover, responsive to installation of the firmware update, the access point can share network bandwidth capacity with the member computing devices of the community Wi-Fi network. Further, the community aggregation system supports transparent handoffs between access points in to the community Wi-Fi network.

Abstract: Embodiments produce a set of instructions for connecting to a network through a network access point based on data crowdsourced from mobile computing devices. The crowdsourced data describes interactions between the mobile computing devices and the network access point when establishing a connection to the network. A cloud service analyzes the crowdsourced data to identify a set of instructions for association with the network access point. The mobile computing devices replay the set of instructions when subsequently attempting to connect to the network access point.

Abstract: Embodiments provision captive portals on network access points (NAPs). In response to a request from a NAP owner, a cloud service associates a NAP with the NAP owner and provides a control channel for configuring the NAP. The NAP owner inputs configuration settings describing the captive portal. The cloud service processes the configuration settings to generate configuration data. The cloud service transmits the configuration data to the NAP. The NAP applies the configuration data to configure the NAP. Some embodiments configure Wi-Fi hotspots with captive portals for gated access to the Internet.

Abstract: Embodiments produce a set of instructions for connecting to a network through a network access point based on data crowdsourced from mobile computing devices. The crowdsourced data describes interactions between the mobile computing devices and the network access point when establishing a connection to the network. A cloud service analyzes the crowdsourced data to identify a set of instructions for association with the network access point. The mobile computing devices replay the set of instructions when subsequently attempting to connect to the network access point.

Abstract: Embodiments apply a set of sequenced instructions to connect to a network through a captive portal. A computing device detects a network access point and obtains the instruction set corresponding to the network access point. The instruction set is derived by a cloud service from crowdsourced data describing interactions between mobile computing devices and the network access point. Applying the instruction set includes performing actions such as navigating web pages to accept terms and conditions, provide user or device information, and more.

Abstract: Various technologies described herein pertain to sharing Wi-Fi credentials based upon relationships in a computer-implemented social network. At least one server computing device of a Wi-Fi credential sharing service receives an identifier from a mobile device and credentials for a Wi-Fi network from the mobile device. The identifier identifies a user of the mobile device. Moreover, the credentials for the Wi-Fi network are desirably shared by the user. Further, contacts of the user are retrieved from the social network. The credentials for the Wi-Fi network are retained in respective accounts of the contacts of the user from the social network. Further, an account of a contact includes credentials for Wi-Fi networks shared with the contact.

Abstract: Various technologies described herein pertain to managing and accessing a community Wi-Fi network. The community Wi-Fi network can include sharer computing devices in differing local networks that share respective network bandwidth capacities with consumer computing devices. Feedback information pertaining to network bandwidth capacity available at different geographic locations covered by the community Wi-Fi network can be collected and evaluated to enable generation of maps for visualizing coverage of the community Wi-Fi network. A consumer computing device can display a map that includes visual information pertaining to the geographic coverage of the community Wi-Fi network relative to a geographic location of the consumer computing device.

Abstract: Embodiments personalize a set of network access points (NAPs) acting as captive portals for each user and/or user device. A cloud service, or the user device, accesses context describing the user device. The context identifies at least one NAP, and may include accounting data. One of the NAP is selected for connection based on the context, and an instruction set associated with the selected NAP is provided for replay by the user device to connect through the NAP. In some embodiments, instructions within the instruction set are also personalized to the user and/or user device based on the context.

Abstract: Various technologies described herein pertain to managing and accessing a community Wi-Fi network. The community Wi-Fi network can include sharer computing devices in differing local networks that share respective network bandwidth capacities with consumer computing devices. Feedback information pertaining to network bandwidth capacity available at different geographic locations covered by the community Wi-Fi network can be collected and evaluated to enable generation of maps for visualizing coverage of the community Wi-Fi network. A consumer computing device can display a map that includes visual information pertaining to the geographic coverage of the community Wi-Fi network relative to a geographic location of the consumer computing device.

Abstract: Embodiments produce a set of instructions for connecting to a network through a network access point based on data crowdsourced from mobile computing devices. The crowdsourced data describes interactions between the mobile computing devices and the network access point when establishing a connection to the network. A cloud service analyzes the crowdsourced data to identify a set of instructions for association with the network access point. The mobile computing devices replay the set of instructions when subsequently attempting to connect to the network access point.

Abstract: Various technologies described herein pertain to managing and accessing a community Wi-Fi network. The community Wi-Fi network can include sharer computing devices in differing local networks that share respective network bandwidth capacities with consumer computing devices. Feedback information pertaining to network bandwidth capacity available at different geographic locations covered by the community Wi-Fi network can be collected and evaluated to enable generation of maps for visualizing coverage of the community Wi-Fi network. A consumer computing device can display a map that includes visual information pertaining to the geographic coverage of the community Wi-Fi network relative to a geographic location of the consumer computing device.

Abstract: Embodiments produce a set of instructions for connecting to a network through a network access point based on data crowdsourced from mobile computing devices. The crowdsourced data describes interactions between the mobile computing devices and the network access point when establishing a connection to the network. A cloud service analyzes the crowdsourced data to identify a set of instructions for association with the network access point. The mobile computing devices replay the set of instructions when subsequently attempting to connect to the network access point.

Abstract: Embodiments provision captive portals on network access points (NAPs). In response to a request from a NAP owner, a cloud service associates a NAP with the NAP owner and provides a control channel for configuring the NAP. The NAP owner inputs configuration settings describing the captive portal. The cloud service processes the configuration settings to generate configuration data. The cloud service transmits the configuration data to the NAP. The NAP applies the configuration data to configure the NAP. Some embodiments configure Wi-Fi hotspots with captive portals for gated access to the Internet.

Abstract: Various technologies described herein pertain to configuring an access point in a local network for operation in a community Wi-Fi network. The community Wi-Fi network includes access points in differing local networks that share respective network bandwidth capacities with member computing devices of the community Wi-Fi network. A request to join the access point to the community Wi-Fi network is transmitted from the access point to a community aggregation system. Responsive to the request, the community aggregation system transmits a firmware update to the access point. The access point installs the firmware update in memory of the access point. Moreover, responsive to installation of the firmware update, the access point can share network bandwidth capacity with the member computing devices of the community Wi-Fi network. Further, the community aggregation system supports transparent handoffs between access points in to the community Wi-Fi network.

Abstract: Embodiments provision captive portals on network access points (NAPs). In response to a request from a NAP owner, a cloud service associates a NAP with the NAP owner and provides a control channel for configuring the NAP. The NAP owner inputs configuration settings describing the captive portal. The cloud service processes the configuration settings to generate configuration data. The cloud service transmits the configuration data to the NAP. The NAP applies the configuration data to configure the NAP. Some embodiments configure Wi-Fi hotspots with captive portals for gated access to the Internet.

Abstract: Various technologies described herein pertain to configuring an access point in a local network for operation in a community Wi-Fi network. The community Wi-Fi network includes access points in differing local networks that share respective network bandwidth capacities with member computing devices of the community Wi-Fi network. A request to join the access point to the community Wi-Fi network is transmitted from the access point to a community aggregation system. Responsive to the request, the community aggregation system transmits a firmware update to the access point. The access point installs the firmware update in memory of the access point. Moreover, responsive to installation of the firmware update, the access point can share network bandwidth capacity with the member computing devices of the community Wi-Fi network. Further, the community aggregation system supports transparent handoffs between access points in to the community Wi-Fi network.