Verizon Patent and Licensing Inc.

Said Soulhi

Bryan Christopher Larish

Patricia Ruey Chang

Haim Shalom Ner

Bjorn Olof Erland Kalderen

Edward Michael Jack

Gagan Puranik

Adam Barron

A computer device may include a memory configured to store instructions and a processor configured to execute the instructions to train a generator neural network to simulate a network entity using a discriminator neural network that discriminates output associated with the network entity from output generated by the generator neural network. The computer device may be further configured to receive a set of input parameters associated with the simulated network entity; use the generator neural network to generate output for the simulated network entity based on the received set of input parameters; and apply the generated output for the simulated network entity to manage a communication network.

Method and system for virtual emulation and self-organizing network control using deep generative models

In some implementations, a network device may receive a set of key performance indicators (KPIs) associated with a network node. The network device may determine a KPI embedding associated with the set of KPIs. The network device may receive an image of a radio frequency (RF) environment associated with the network node. The network device may determine an RF environment embedding associated with the image of the RF environment. The network device may determine a joint embedding based on the KPI embedding and the RF environment embedding. The network device may determine at least one configuration parameter for the network node based on the joint embedding.

Systems and methods for configuring a network node based on a radio frequency environment

A device may generate neural network encryption models based on a dataset descriptor, a dataset geometry, and selected neural network types, and may generate obfuscation features based on the dataset descriptor, a noise type, an obfuscation model type, and noise and model parameters. The device may train the neural network encryption models, with a dataset and the obfuscation features, to generate model weights, a latent space, and noising and denoising models, and may generate an intelligent decryption model based on the model weights, the latent space, and the noising and denoising models. The device may receive an encrypted dataset associated with a target environment, and may determine whether the target environment is valid according to immune rules. The device may process, based on determining that the target environment is valid, the encrypted dataset, with the intelligent decryption model, to generate a decrypted dataset.

Systems and methods for providing enhanced multi-layered security with encryption models that improve zero-trust architectures

A system described herein may train an explanation model based on a set of images and a set of explanation labels. The system may receive input data, and may provide the input data to the explanation model and a second model. The second model may provide a set of output labels, which may include performing unknown or 'black box' processing on the input data. The explanation model may generate one or more images based on the input data, compare the images to the set of images based on which the explanation model was trained, and accordingly identify one or more explanation labels with bounding boxes associated with the generated one or more images. The system may output, in response to the input data, the set of output labels provided by the second model as well as the identified explanation labels.

Systems and methods for explanation of models using image analysis techniques

A system described herein may provide a technique for the augmentation of one or more models based on target distributions of values associated with the models. Such models may be associated with a wireless network, and may associate a set of states to a set of values and/or to a set of actions. Such states may include particular sets of configuration parameters associated with the wireless network, the values may include Key Performance Indicators ('KPIs') associated with a randomness state model or other metrics associated with the wireless network, and the actions may include one or more actions to perform with respect to the wireless network when given a particular state and/or set of KPIs.

Systems and methods for machine learning model augmentation using target distributions of key performance indicators in a wireless network

One or more computing devices, systems, and/or methods for system parameter identification and network component configuration are provided. A state comprising a system parameter combination, a traffic model, and a channel assignment may be generated. A network traffic scenario is executed through a virtualized testbed using the state. A reward for the system parameter combination may be generated based upon key performance indicators output by the network traffic scenario. A reward policy and rewards generated for system parameter combinations are used to select a system parameter combination that is used to configure a network component of a communication network.

Virtualized architecture for system parameter identification and network component configuration with reinforcement learning

A system described herein may provide a technique for analyzing metrics, parameters, attributes, and/or other information associated with networks or other devices or systems associated with high-dimensional data in order to determine potential configuration changes that may be made to such networks or other devices or systems in order to optimize and/or otherwise enhance the operation of such networks or other devices or systems. Multiple autoencoders associated with multiple dimensions may be used to calculate reconstruction errors or other features of data (e.g., metrics, parameters, etc.) that may be used to define operating or performance states of the network. Operating or performance states of network components may be mapped to quantum state objects ('QSOs') for analysis using artificial intelligence and/or machine learning techniques or other suitable techniques.

Systems and methods for autonomous network management using deep reinforcement learning

A device may receive network policies of a network, and network performance data identifying KPIs of the network, and may generate an embedded space of reconstructed data that is embedded in an original space that includes the KPIs. The device may calculate reconstruction errors based on differences between the reconstructed data and the network performance data, and may calculate a convex hull of the original space. The device may calculate a convex hull of the embedded space, and may determine reward metrics based on the reconstruction errors, the convex hull of the original space, and the convex hull of the embedded space. The device may define performance baselines associated with portions, and may generate a new reward for a portion based on a particular reconstruction error, a particular convex hull of the embedded space, and a particular performance baseline. The device may perform actions based on the new reward.

Systems and methods for managing network performance based on defining rewards for a reinforcement learning model

Method and system for virtual network emulation and self-organizing network control using deep generative models

A method, a device, and a non-transitory storage medium provide a validation and anomaly detection service. The service includes quantitatively assessing latent space data representative of network performance data, which may be generated by a generative model, based on quantitative values pertaining to quantitative criteria. The quantitative criteria may include Hausdorff distances, divergence, joint entropy, and/or total correlation. The service further includes generating geogrid data for services areas of deployed network devices and service areas for prospective and new deployments based on selected latent space data and corresponding network performance data.

Method and system for anomaly detection and network deployment based on quantitative assessment

A method, a device, and a non-transitory storage medium provide a validation and anomaly detection service. The service includes quantitatively assessing latent space data representative of network performance data, which is generated by a generative model, based on quantitative values pertaining to quantitative criteria. The quantitative criteria may include Hausdorff distances, divergence, joint entropy, and total correlation. The service further includes generating geogrid data for services areas of deployed network devices and service area for prospective and new deployments based on selected latent space data and corresponding network performance data.

A network device predicts end-to-end channel capacity entropy to permit use of optimal throughput settings in a network. The network device stores class definitions for a network condition; identifies multiple input features to correlate with the class definitions; generates a multiclass classification model that produces an importance score for each of the multiple input features, wherein the importance score reflects the contribution of an input feature to the network condition; selects two or more of the multiple input features with highest importance scores as influential features; predicts the behavior of the influential features to identify a current class, from the class definitions, for the network condition over an end-to-end communication channel; and sends an estimated network condition, based on the current class, to a device for traffic optimization.

Growing community of inventors

Boston, MA, United States of America

Said Soulhi