Abstract: A method (700) for distributed machine learning, ML, using a set of N CDs. The method includes obtaining first topology information, wherein, for each CD included in the set of N CDs, the first topology information identifies other CDs included in the set of N CDs to which the CD is connected, thereby identifying a set of CD pairs. The method also includes obtaining first network state information, wherein, for each CD included in the set of N CDs, the first network state information comprises a network state vector for the CD, wherein the network state vector for the CD comprises, for each other CD to which the CD is indicated as being connected by the first topology information, a determined network state value. The method further includes determining a consensus weight matrix (W) using the obtained first network state information and the first topology information.

Abstract: A method, system and apparatus are disclosed. According to one or more Examples, an infrastructure network node (16) for providing a virtual multiple-input multiple-output. MIMO, cellular network supporting a plurality of network nodes (16) associated with a plurality of service providers. SPs is provided. The infrastructure network node (16) is configured to determine a downlink coordinated precoder for online multi-cell MIMO wireless network virtualization. WNV, where the downlink coordinated precoder is based at least on imperfect channel state information. CSI, and to optionally indicate the downlink coordinated precoder.

Abstract: A method and network node for distributed coordinated downlink precoding for multi-cell multiple input multiple output (MIMO) wireless network virtualization are disclosed. According to one aspect, a network node is configured to, in each of a plurality of successive transmission time periods: transmit to each of a plurality of SPs a corresponding set of channel information; receive from each of the plurality of SPs a service demand and a normalized precoding matrix, the normalized precoding matrix being determined by the corresponding SP based on the corresponding set of channel information; allocate a virtual transmit power to each of the plurality of SPs based at least in part on the received service demands and normalized precoding matrices; and determine a downlink precoding matrix to transmit messages to WDs, the downlink precoder matrix being based at least in part on the received service demands and normalized precoding matrices.

Abstract: A method and network node configured to perform wireless network virtualization to allocate resources of an infrastructure provider (InP) among a plurality of service providers (SPs) are provided. According to one aspect, a method in a network node includes, in each of a plurality of successive time slots in a time interval, T: receiving precoding feedback information from each SP, allocating a virtual transmit power to each cell served by an SP based at least in part on the received precoding feedback information and determining a precoding matrix for each SP. The method also includes minimizing a loss function based at least in part on the received precoding feedback information received in multiple previous time slots.

Abstract: Methods, systems and computer program products for distributed machine learning are provided. Such methods, systems and products may comprise, or may comprise instructions operable to configure one or more processors to perform, a set of acts. Such acts may comprise a plurality of nodes performing a set of training round activities, and a server performing a set of network topology design activities. The network topology design activities may comprise the server generating update data based on data from the plurality of nodes. The plurality of nodes may use that data to update control values used to exchange and combine machine learning models. After the set of training round activities and the set of network topology design activities have been repeated one or more times, the plurality of nodes may send machine learning models to the server, and the server may use them to create an aggregated machine learning model.

Abstract: A method (700) for distributed machine learning, ML, using a set of N CDs. The method includes obtaining first topology information, wherein, for each CD included in the set of N CDs, the first topology information identifies other CDs included in the set of N CDs to which the CD is connected, thereby identifying a set of CD pairs. The method also includes obtaining first network state information, wherein, for each CD included in the set of N CDs, the first network state information comprises a network state vector for the CD, wherein the network state vector for the CD comprises, for each other CD to which the CD is indicated as being connected by the first topology information, a determined network state value. The method further includes determining a consensus weight matrix (W) using the obtained first network state information and the first topology information.

Abstract: A method, system and apparatus are disclosed. An edge node configured to communicate with a plurality of wireless devices (WDs) is described. The edge node includes a communication interface configured to receive a plurality of signal vectors from the plurality of WDs, where the plurality of signal vectors is based on a plurality of updated local models associated with the plurality of WDs. The edge node also includes processing circuitry in communication with the communication interface, where the processing circuitry is configured to update a global model based at least on the plurality of signal vectors; and cause at least one transmission of the updated global model to the plurality of WDs.

Abstract: A method, system and apparatus are disclosed. According to one or more Examples, an infrastructure network node (16) for providing a virtual multiple-input multiple-output. MIMO, cellular network supporting a plurality of network nodes (16) associated with a plurality of service providers. SPs is provided. The infrastructure network node (16) is configured to determine a downlink coordinated precoder for online multi-cell MIMO wireless network virtualization. WNV, where the downlink coordinated precoder is based at least on imperfect channel state information. CSI, and to optionally indicate the downlink coordinated precoder.

Abstract: A method, system and apparatus are disclosed. According to one aspect, a network node configured to communicate with a wireless device (WD), includes processing circuitry configured to perform downlink wireless network virtualization by minimizing an expected deviation of received signals at WDs subject to network node power constraints.

Abstract: A method, network node and processor for processing uplink signals transmitted from a wireless device (WD) to provide a combination of quantize-forwarding and decode-forwarding relayed signals in massive multiple input multiple output (MIMO) heterogeneous networks (HetNets). According to one aspect, whether quantize-forwarding or decode-forwarding is used depends on a ratio of a quality of a WD signal received at a Small Cell Base Station (SCBS) to a quality of the WD signal received at a Macro Cell Base Station (MCBS). When the ratio is less than a first threshold, signals received at the SCBS may be quantize-forwarded to the MCBS and when the ratio exceeds a second threshold larger than the first threshold, signals received at the SCBS may be decode-forwarded to the MCBS. When the ratio lies between the first and second threshold, signals received at the SCBS may be quantize-forwarded to the MCB.

Abstract: A method and network node for online coordinated multi-cell precoding are provided. According to one aspect, a method includes receiving from each of the plurality of service providers a virtual precoder matrix determined by the corresponding service provider. The method also includes determining a precoder matrix by minimizing a precoding deviation from a virtualization demand of the network subject to at least one power constraint, the virtualization demand being based at least in part on a product of a channel state matrix and a virtual precoder matrix. The method further includes applying the determined precoder matrix to signals applied to a plurality of antennas to achieve a sum of throughputs for the plurality of service providers that is greater than a sum of throughputs for the plurality of service providers achievable when the virtual precoder matrices are applied to the signals.

Abstract: A method for performing online convex optimization is provided. The method includes receiving, from two or more worker nodes, a local decision vector and local data corresponding to each of the two or more worker nodes. The method includes performing a multi-step gradient descent based on the local decision vector and the local data received from the two or more worker nodes. Performing the multi-step gradient descent includes determining a global decision vector and corresponding global information. The method includes sending, to each of the two or more worker nodes, the global decision vector and corresponding global information.

Abstract: A method and network node configured to perform wireless network virtualization to allocate resources of an infrastructure provider (InP) among a plurality of service providers (SPs) are provided. According to one aspect, a method in a network node includes, in each of a plurality of successive time slots in a time interval, T: receiving precoding feedback information from each SP, allocating a virtual transmit power to each cell served by an SP based at least in part on the received precoding feedback information and determining a precoding matrix for each SP. The method also includes minimizing a loss function based at least in part on the received precoding feedback information received in multiple previous time slots.

Abstract: A method and network node for distributed coordinated downlink precoding for multi-cell multiple input multiple output (MIMO) wireless network virtualization are disclosed. According to one aspect, a network node is configured to, in each of a plurality of successive transmission time periods: transmit to each of a plurality of SPs a corresponding set of channel information; receive from each of the plurality of SPs a service demand and a normalized precoding matrix, the normalized precoding matrix being determined by the corresponding SP based on the corresponding set of channel information; allocate a virtual transmit power to each of the plurality of SPs based at least in part on the received service demands and normalized precoding matrices; and determine a downlink precoding matrix to transmit messages to WDs, the downlink precoder matrix being based at least in part on the received service demands and normalized precoding matrices.

Abstract: A method and network node for online coordinated multi-cell precoding are provided. According to one aspect, a method includes receiving from each of the plurality of service providers a virtual precoder matrix determined by the corresponding service provider. The method also includes determining a precoder matrix by minimizing a precoding deviation from a virtualization demand of the network subject to at least one power constraint, the virtualization demand being based at least in part on a product of a channel state matrix and a virtual precoder matrix. The method further includes applying the determined precoder matrix to signals applied to a plurality of antennas to achieve a sum of throughputs for the plurality of service providers that is greater than a sum of throughputs for the plurality of service providers achievable when the virtual precoder matrices are applied to the signals.

Abstract: A method, system and apparatus are disclosed. A network node is configured to communicate with a wireless device, WD, where the network node configured to, and/or comprising a radio interface and/or comprising processing circuitry configured to: receive an indication whether the WD is to cache at least one file based on spatial statistics of a cache-enabled device-to-device, D2D, network collected by the WD, and perform spectrum allocation based on the received indication.

Abstract: Embodiments of a method of operation of a power control coordinator to control transmission power of a plurality of Device-to-Device (D2D) pairs that co-exist with a Cellular User Equipment (CUE) that communicates with a base station of a cellular communications network comprises obtaining, for a particular time slot, delayed Network State Information (NSI) feedback from at least some of the plurality of D2D pairs. The method further comprises computing transmission powers for the D2D pairs, respectively, for the particular time slot using On-Line Convex Optimization (OCO) to solve an optimization problem that maximizes a weighted sum data rate of D2D pairs with a constraint of maximum expected interference to the base station. The method further comprises providing, to each D2D pair, an indication of the computed transmission power for the D2D pair for the particular time slot.

Abstract: A method, network node and processor for processing uplink signals transmitted from a wireless device (WD) to provide a combination of quantize-forwarding and decode-forwarding relayed signals in massive multiple input multiple output (MIMO) heterogeneous networks (HetNets). According to one aspect, whether quantize-forwarding or decode-forwarding is used depends on a ratio of a quality of a WD signal received at a Small Cell Base Station (SCBS) to a quality of the WD signal received at a Macro Cell Base Station (MCBS). When the ratio is less than a first threshold, signals received at the SCBS may be quantize-forwarded to the MCBS and when the ratio exceeds a second threshold larger than the first threshold, signals received at the SCBS may be decode-forwarded to the MCBS. When the ratio lies between the first and second threshold, signals received at the SCBS may be quantize-forwarded to the MCB.

Abstract: A method, system and apparatus are disclosed. According to one aspect, a network node configured to communicate with a wireless device (WD), includes processing circuitry configured to perform downlink wireless network virtualization by minimizing an expected deviation of received signals at WDs subject to network node power constraints.

Abstract: Embodiments of a method of operation of a power control coordinator to control transmission power of a plurality of Device-to-Device (D2D) pairs that co-exist with a Cellular User Equipment (CUE) that communicates with a base station of a cellular communications network comprises obtaining, for a particular time slot, delayed Network State Information (NSI) feedback from at least some of the plurality of D2D pairs. The method further comprises computing transmission powers for the D2D pairs, respectively, for the particular time slot using On-Line Convex Optimization (OCO) to solve an optimization problem that maximizes a weighted sum data rate of D2D pairs with a constraint of maximum expected interference to the base station. The method further comprises providing, to each D2D pair, an indication of the computed transmission power for the D2D pair for the particular time slot.