Abstract: A method is provided. The method comprises obtaining local observation data of a group of one or more agents. The local observation data indicates performance of each agent included in the group. The method further comprises obtaining global state data indicating collective performance of the group and based on the obtained local observation data and the obtained global state data, determining a discount factor for each agent included in the group. The discount factor is a weight value of a future expected reward for each agent included in the group.

Abstract: In one aspect, a method includes obtaining a first and second set of word embeddings from a first local machine learning (ML) model and a second local ML model. The method includes generating first and second latent space representations by processing the first and second sets of word embeddings using an artificial neural network (ANN) trained with the first and second local ML models, wherein the first and second latent space representations comprises a plurality of first and second contexts associated with the first and second set of word embeddings. The method includes correlating the first and second sets of word embeddings based on the plurality of first and second contexts. The method includes aggregating, based on the correlating, the first set of word embeddings and the second set of word embeddings into a global Machine Learning (ML) model of word embeddings.

Abstract: A method performed by a virtual reality, VR, system for play of a VR game in a self-driving vehicle is provided. The method includes receiving a schedule of predicted acceleration for a future time period for the self-driving vehicle operating in a real world environment. The method further includes, responsive to the schedule of predicted acceleration, adjusting play of the VR game based on the schedule of predicted acceleration. Methods performed by a vehicle system, and related VR systems and vehicle systems are also provided.

Abstract: A method for distributed machine learning (ML) at a central computing device is provided.

Abstract: Embodiments described herein relate to methods and apparatuses for generating one or more answers relating to a machine learning, ML, model. A method in a first node comprises obtaining one or more queries relating to a first output of the ML model, wherein the first output of the machine learning, ML, model is intended to fulfil one or more requirements in an environment; for each of the one or more queries performing a reinforcement learning process.

Abstract: The embodiments herein relate to a method performed by a cloud node. The cloud node obtains measurements from at least some of the sensor nodes. The cloud node mathematically determines a minimum number of sensor nodes and their optimal locations. Based on the obtained measurements and the mathematically determined optimal locations, the cloud node graphically determines an optimal location for each of the minimum number of sensor nodes. The cloud node compares the mathematically and the graphically determined optimal locations. When the comparison indicates that the mathematically and graphically determined optimal locations are the same, the cloud node determines a minimum number of fog nodes. Based on the optimal location of sensor nodes, the cloud node determines an optimal location for each of the minimum number of fog nodes.

Abstract: A method of performing multi-agent reinforcement learning in a system including a master node and a plurality of agents that execute actions on an environment based on respective local policies of the agents is provided. The method includes generating a ranking of the plurality of agents based on levels of variability of stochastic processes underlying the behavior of respective ones of the plurality of agents, sequentially updating the local policies of the agents in order based on the ranking, wherein the local policy of a selected agent is updated conditioned on an expected next state of at least one previously selected agent, simultaneously executing actions by agents based on their updated local policies, and updating the ranking of the plurality of agents in response to executing the actions.

Abstract: Embodiments herein may e.g. relate to a method performed by a network node (12) for handling one or more operations in a communications network comprising a plurality of computing devices (10,11) performing one or more tasks. The network node (12) obtains an indication of a failure of an operation in the communications network; and obtains one or more parameters to resolve the failure. The one or more parameters relate to resources of the plurality of computing devices (10,11) and the communications network (1), wherein the one or more parameters are structured in an hierarchic manner and defined by a task of a capability, a resource used for the task, and a service level for the task. The network node (12) generates a plan by taking an aimed service level into account as well as the obtained one or more parameters; and executes one or more operations using the generated plan.

Abstract: A computer-implemented method for building a machine learning (ML) model is provided. The method includes training a ML model using a set of input data, wherein the ML model includes a plurality of layers and each layer includes a plurality of filters, and wherein the set of input data includes class labels; obtaining a set of output data from training the ML model, wherein the set of output data includes class probabilities values; determining, for each layer in the ML model, by using the class labels and the class probabilities values, a working value for each filter in the layer; determining, for each layer in the ML model, a dominant filter, wherein the dominant filter is determined based on whether the working value for the filter exceeds a threshold; and building a subset ML model based on each dominant filter for each layer, wherein the subset ML model is a subset of the ML model.

Abstract: A computer-implemented method includes: obtaining loss functions including: a first loss function associated with a first reinforcement learning, RL, model performed by a first local agent, wherein the first loss function is a function of one or more first parameters; and a second loss function associated with a second RL model at a second local agent, wherein the second loss function is a function of one or more second parameters; determining a combined loss function based on the loss functions; minimizing the combined loss function with respect to the first parameters and the second parameters to determine updated values for the first parameters and updated values for the second parameters; initiating execution of a first updated action by the first local agent based on the updated values of the first parameters; and initiating execution of a second updated action by the second local agent based on the updated values of the second parameters.

Abstract: A method for distributed learning at a local computing device is provided. The method includes: training a local model of a first model type on local data, wherein the local data comprises a first set of labels; testing the local model on a portion of global data pertaining to the first set of labels, wherein the global data comprises a second set of labels and the first set of labels is a strict subset of the second set of labels; as a result of testing the local model on the portion of the global data pertaining to the first set of labels, producing a first set of probabilities corresponding to the first set of labels; and sending the first set of probabilities corresponding to the first set of labels to a central computing device.

Abstract: A method for training a central model in a federated learning system is provide. The method includes receiving a first update from a first local model of a set of local models; receiving a second update from a second local model of the set of local models; enqueueing the first update and the second update in one more queues corresponding to the set of local models; selecting an update from the one or more queues to apply to a central model based on determining that a selection criteria is satisfied, the selection criteria being related to a quality of the central model; and applying the selected update to the central model or instructing a node to apply the selected update to the central model.

Abstract: A method for managing computation load of a fog node is disclosed, wherein a computation capacity of the fog node is predicted to become unavailable to a fog network. The method comprises identifying a candidate set of nodes for computational load transfer from the fog node. The method further comprises obtaining a computation graph representing computation in the fog network, and using a learning model to identify a morphism from the obtained computation graph to a new computation graph, in which the fog node is not included. The identified morphism comprises a sequence of one or more morphing operations that replaces the fog node in the obtained computation graph with a topology of one or more nodes selected from the candidate set. The method further comprises causing computation performed at the fog node to be transferred to one or more nodes of the candidate set.

Abstract: Methods and systems for searching and retrieving information. In one aspect, there is a method of retrieving information using a knowledge base. The method comprises receiving a search query entered by a user and using a first model to identify a category corresponding to the received search query. The method further comprises based on the received search query, a loss function of the first model, and an objective function of a second model, identifying T topics corresponding to the received search query, and performing a search for the received search query only on a part of the knowledge base that is associated with the identified category and/or the identified topics. The method further comprises retrieving one or more files associated with the identified category and/or the identified topics.

Abstract: A method performed by a local client computing device is provided. The method includes training a local model using data from the local client computing device, resulting in a local model update; sending the local model update to a central server computing device; receiving from the central server computing device a first updated global model; determining that the first updated global model does not meet a local criteria, wherein determining that the first updated global model does not meet a local criteria comprises computing a score based on the first updated global model, wherein the score exceeds a threshold; in response to determining that the first updated global model does not meet a local criteria, sending to the central server computing device context information; and receiving from the central server computing device a second updated global model.

Abstract: In one aspect, a method performed by a network node for predicting a probability of state change of a node (e.g., a fog node) in a network is provided. The network node determines a set of weights based on attributes of the node. The network node estimates the probability of state change of the node using the determined set of weights and a set of one or more attribute values related to the node where determining the set of weights includes maximizing an evaluation value associated to the node.

Abstract: A method for detecting and reducing the impact of deficient nodes in a machine learning system is provided. The method includes receiving a local model update from a first local client node; determining a change in accuracy caused by the local model update; determining that the change in accuracy is below a first threshold; and in response to determining that the change in accuracy is below the first threshold, sending a request to the first local client node signaling the first local client node to compress local model updates.

Abstract: A method on a central node or server is provided.

Abstract: The present disclosure relates to a method performed by a cloud node (104) for handling sensor nodes and fog nodes in a communications system (100), wherein the communications system comprises a plurality of sensor nodes (110) located at a plurality of locations, to be handled by the fog nodes (120), the method comprising obtaining (S300) a first number of sensor nodes and their respective locations, out of said plurality of sensor nodes, to monitor the communications system in its entirety, based on measurements from at least some of the plurality of sensor nodes at their respective locations; determining (S310) a second number of said fog nodes and their respective location, based on the first number of sensor nodes and a connectivity capacity of said second number of fog nodes, where the second number of fog nodes is determined to cover said first number of sensor nodes; ranking (S320) said second number of fog nodes according to a conditional probability of failure for the second number of fog nodes, base

Abstract: A method for managing resources includes applying an ensemble model having a plurality of sub-models such that an output of the ensemble model is a weighted average of predictions from the sub-models, and is a prediction of multiple parameters. The method includes determining that an accuracy of the ensemble model is below a first threshold; and as a result, optimizing weights for the predictions from the sub-models. Optimizing weights for the predictions from the sub-models includes: updating the weights selected by the reinforcement learning by looking ahead over a prediction horizon and optimizing the reward function at the given time instance. The method further includes using the prediction of the multiple parameters to manage resources.