Abstract: AM and LM parameters to be used for adapting an ASR model are derived for each audio segment of an audio stream comprising multiple audio programs. A set of identifiers, including a speaker identifier, a speaker domain identifier and a program domain identifier, is obtained for each audio segment. The set of identifiers are used to select most suitable AM and LM parameters for the particular audio segment. The embodiments enable provision of maximum constraints on the AMs and LMs and enable adaptation of the ASR model on the fly for audio streams of multiple audio programs, such as broadcast audio. This means that the embodiments enable selecting AM and LM parameters that are most suitable in terms of ASR performance for each audio segment.

Abstract: According to an aspect, there is provided a power supply unit, PSU, configured for use in a first radio access network, RAN, node in a communication network. The PSU includes a power supply input interface configured to enable the PSU to be connected to, and receive electrical power from, an electrical power supply; and a power line communication, PLC, unit connected to the power supply input interface and configured to connect to a first baseband, BB, processing unit in the first RAN node. The first BB processing unit is configured for processing baseband signals in the first RAN node. The PLC unit is configured to operate according to a PLC protocol to enable communications between the first BB processing unit and a second BB processing unit in a second RAN node that is also connected to the electrical power supply.

Abstract: Methods and devices enable connecting devices to cellular networks using the devices' hardware identifiers. Subscriber records include a hardware identifier assigned when the devices are manufactured. A target hardware identifier included in an attachment request is associated with an International Mobile Subscriber Identity, IMSI, available to the cellular network if, according to subscriber records, the device is registered.

Abstract: Methods, systems, and apparatuses are presented for grouping worker nodes in a machine learning system comprising a master node and a plurality of worker nodes, the method comprising grouping each worker node of the plurality of worker nodes into a group of a plurality of groups based on characteristics of a data distribution of each of the plurality of worker nodes, subgrouping worker nodes within the group of the plurality of groups into subgroups based on characteristics of a worker neural network model of each worker node from the group of the plurality of groups, averaging the worker neural network models of worker nodes within a subgroup to generate a subgroup average model, and distributing the subgroup average model.

Abstract: (57) Abstract: A method, performed by a first node (111), for handling parameters to configure a second node (112). The first node (111) determines (207) parameters to configure the second node (112). The determining (207) is based on an analysis by a Generative Adversarial Network comprising performing iteratively: i) generating a set of parameters estimated to be linked to a performance of a first group of nodes (121) being above a threshold, and ii) discriminating between the generated set of parameters and a first set of parameters observed to be linked to the performance of the first group of nodes (121) being above the threshold to obtain a score for every parameter. The score indicates how different the generated parameters and the first set of parameters are. The determined parameters have the score resulting from the discriminating being lower than anotherthreshold. The first node (111) also outputs an indication comprising the determined parameters.

Abstract: Methods, systems, apparatuses and computer programs are presented for developing machine-learning models. A method for decentralized machine learning in a target worker node comprises: receiving a plurality of adapted neural network models from a plurality of worker nodes, wherein each of the adapted neural network models is generated by training a worker node neural network using local data of the worker node from among the plurality of worker nodes; selecting, from the plurality of adapted neural network models, a set of adapted neural network models that satisfy performance criteria when local data of the target worker node is input; and averaging the set of adapted neural network models to generate an average model.

Abstract: Methods and devices optimize backup power management for network devices connected to backup power sources including two different batteries. Machine learning techniques are used to predict upcoming power outages affecting the network device based on power-related historic information and current conditions. A backup-power operation plan prescribing usage of the backup power sources during the predicted power outages is then generated. The backup-power operation plan schedules a battery among the backup power devices to be used at least twice without being recharged.

Abstract: A method of operating a radio network node to adjust power consumption of a telecommunications network is provided. The method includes determining a traffic prediction representing how each node of a set of nodes will interact with the radio network node over a period of time using a combined traffic model based on a traffic model of each node in the set of nodes. The method further includes determining to enable or disable at least one power related feature of the radio network node based on the traffic prediction.

Abstract: Embodiments described herein provided methods and apparatus for controlling switching a power supply of a first network node between an electrical power grid and an alternative power supply in which the first network node is configured to transmit data to a second network node over a data transmission link. The method includes receiving first information indicative of a predicted decrease in a transmission rate of data transmitted by the first network node over the data transmission link. Responsive to receiving the first information, a first request is transmitted to the first network node to switch the power supply of the first network node from the electrical power grid to the alternative power supply.

Abstract: A method (300) for using federated learning to develop a machine-learning model is disclosed. The method, performed by a management function, comprises developing a seed version of the machine-learning model using a machine-learning algorithm (310) and communicating the seed version of the machine-learning model to a plurality of distributed nodes, each of the plurality of distributed nodes being associated with a local data set (320).

Abstract: A method of generating a synthetic training dataset for training a machine learning model using an original training dataset including a plurality of features includes selecting a feature ci of the original training dataset as a target vector yi, selecting remaining features of the original training dataset as a set of training input vectors X\i, where X\i includes all features of the training dataset other than a feature corresponding to the selected feature ci, and training a prediction model f(yi|X\i). The method generates an estimate y?i of the target vector yi by applying the prediction model to the set of training vectors X\i, and inserts a synthetic feature c?i corresponding to the estimate y?i of the target vector yi into a synthetic training dataset.

Abstract: AM and LM parameters to be used for adapting an ASR model are derived for each audio segment of an audio stream comprising multiple audio programs. A set of identifiers, including a speaker identifier, a speaker domain identifier and a program domain identifier, is obtained for each audio segment. The set of identifiers are used to select most suitable AM and LM parameters for the particular audio segment. The embodiments enable provision of maximum constraints on the AMs and LMs and enable adaptation of the ASR model on the fly for audio streams of multiple audio programs, such as broadcast audio. This means that the embodiments enable selecting AM and LM parameters that are most suitable in terms of ASR performance for each audio segment.

Abstract: Embodiments herein disclose e.g. a method performed by a network node (12) for handling operation in a communication network, wherein the network node comprises at least one power supply unit, and one or more additional power units for supplying power to the network node. The network node (12) obtains an output from a computational model; and sets an output voltage from the at least one power supply unit based on the obtained output.

Abstract: Systems and methods are disclosed herein for enhanced feedback for cascaded federated machine learning (ML). In one embodiment, a method of operation of a server comprises, for a training epoch, receiving, from each of client device, information including a local ML model as trained at the client device and an estimated value of each parameter output by the local ML model. The method further comprises aggregating the local ML models to provide a global ML model and training a network ML model based on the estimated values of each of the parameters output by the local ML models and global data available at the server. The method further comprises providing, to each client device, information including the global ML model and feedback information related to a hidden neural network layer of the network ML model. The method further comprises repeating the process for one or more additional training epochs.

Abstract: A method performed by a network computing device in a telecommunications network for adaptively deploying an aggregated machine learning model and an output parameter in the telecommunications network to control an operation in the telecommunications network. The network computing device can aggregate client machine learning models and an output performance metric the client machine learning models to obtain an aggregated machine learning model and an aggregated output performance metric. The network computing device can train a network machine learning model with the aggregated output performance metric and at least one measurement of a network parameter to obtain an output parameter. The network computing device can send to the client computing devices the aggregated machine learning model and the output parameter of the network machine learning model. A method performed by a client computing device is also provided.

Abstract: A method performed by a first network node (111). The method is for handling charging of a wireless device (130). The first network node (111) and the wireless device (130) operate in a wireless communications network (100). The first network node (111) modulates (307) one or more beam forming beams (121) in an antenna array (124) controlled by the first network node (111) with a pulse width modulation. The first network node (111) then charges (308), wirelessly, the wireless device (130), with the modulated one or more beam forming beams (121).

Abstract: A method performed by a mobile device for handling identification of equipment. The mobile device records an image, in a recording direction at a first location, of the equipment. Upon recording the image, the mobile device further obtains one or more radiation indications for determining a direction of radiation from the equipment; and provides the obtained one or more radiation indications associated with the recorded image, to an internal identifying process at the mobile device and/or a network node for identifying the equipment.

Abstract: An operator system of a wireless communication network operator sends, to a regulator system of a regulator, a record that includes information about administration of a subscription identifier associated with the wireless communication network operator. Responsive to sending the record to the regulator system, the operator system receives a response that indicates whether the regulator system approves of or rejects the record being added to a permissioned distributed database that is distributed at least in part between the regulator system and the operator system. The operator system adds or does not add the record to the permissioned distributed database depending on the response.

Abstract: Method and system for Root Cause Analysis (RCA) across multiple network systems. Update information of a first local root cause analysis mechanism is received. An RCA controller generates, based on the update information, a new node to be added to a global root cause decision tree, where the global root cause decision tree is to be shared by at least two of the plurality of network operators. The RCA controller requests storage of the new node in a distributed ledger that is shared by network operators. The RCA controller participates in a verification operation of the new node. In response to determining that the verification operation is successful, the RCA controller adds an entry including the new node to the distributed ledger as part of the global root cause decision tree. Alternatively, when the verification operation is not successful, the new node is not added to the distributed ledger.

Abstract: There is provided mechanisms for handling poor audio quality. A method is performed by a receiving terminal device. The method comprises obtaining an indication of poor audio quality of incoming audio at the receiving terminal device. The incoming audio originates from a transmitting terminal device. The method comprises initiating text conversion of the incoming audio. The method comprises receiving text resulting from automatic speech recognition having been applied to the incoming audio. The method comprises providing a representation of the text to a user interface of the receiving terminal device.