Abstract: Systems and methods may train neural networks (NNs) and determine when to stop training to not waste computing or other resources when improvement is not no longer likely. After training period for a NN, a model trained using training data from other NNs may return a a probability of improvement in the loss of the NN or a probability that the likely best loss of the NN is lower than the best loss of the other NNs for which hyperparameters have been chosen. Training may be stopped if the probability is less than a threshold, or a wait value is greater than a wait threshold.

Abstract: Systems and methods may train neural networks (NNs) and determine when to stop training to not waste computing or other resources when improvement is not no longer likely. After training period for a NN, a model trained using training data from other NNs may return a a probability of improvement in the loss of the NN or a probability that the likely best loss of the NN is lower than the best loss of the other NNs for which hyperparameters have been chosen. Training may be stopped if the probability is less than a threshold, or a wait value is greater than a wait threshold.

Abstract: An aspect includes identifying, by a processor, a plurality of data sources associated with a defined geographic area. A plurality of scheduling data associated with a plurality of users is gathered from the data sources across a communication network. One or more shared activities and interests are identified based at least in part on the scheduling data and/or overlapping and similar proximity of user locations. One or more community networks that link two or more of the data sources are created based at least in part on the one or more shared activities and interests. One or more notifications associated with the one or more shared activities and interests are distributed across the one or more community networks.