Abstract: Facilitating artificial intelligence enabled dynamic threshold-based cell and/or carrier switching for energy efficiency in advanced communication networks is provided. A method includes determining traffic load switching thresholds for respective cells of a group of cells of a communications network. The method also includes determining respective results of application of a utility function to the respective cells. Based on the traffic load switching thresholds and the respective results of the utility function, the method includes determining that a selected switching policy for a single cell of the group of cells satisfies a parameter of the utility function. In addition, the method includes facilitating implementing the selected switching policy for the single cell. Respective switching policies of other cells of the group of cells, other than the single cell, are not implemented during the implementing of the selected switching policy for the single cell.

Abstract: Energy efficient link adaptation (e.g., using a computerized tool), is enabled. For example, a system can comprise at least one processor and at least one memory that stores executable instructions that, when executed by the processor, facilitate performance of operations. The operations can comprise, for a communication link between network equipment and a user equipment, modeling network equipment power consumption applicable to the network equipment and modeling user equipment power consumption applicable to the user equipment, based on at least one model resulting from the modeling of the network equipment power consumption and the modeling of the user equipment power consumption, determining a link energy efficiency applicable to the communication link, and based on the link energy efficiency, performing link adaptation comprising modifying a set of link transmission parameters applicable to the communication link, wherein the modifying has been determined to increase the link energy efficiency.

Abstract: The technologies described herein are generally directed to distributing inference with a language model to a pipeline of computing equipment. For instance, a system can, based on a resource availability table, identify a pipeline of compute nodes that process an inference request, based on different layers of a large language model, with the resource availability table being generated based on a first compute node of the pipeline of compute nodes. The system may, based on the resource availability table, communicate an input representation to the first compute node, and after processing the input representation, the first compute node communicates an intermediate representation to a second compute node of the pipeline. The system may further, based on an output representation, communicate a response to the inference request, with the output representation being received from the second node after processing the intermediate representation using a second layer of the large language model.

Abstract: The technologies described herein are generally directed to using machine learning to allocate auxiliary data collection resources. For instance, a system can identify an anomaly in operation of a computing node of computing nodes of a network. The system may further identify, by employing a machine learning model, respective locations of telemetry information representative of operations of computing nodes that have dependencies upon the computing node, with the machine learning model being trained based on data representative of past anomalies in computing nodes of the network. Further, the system may, based on the telemetry information, identify and direct a network failure mitigation resource to perform a mitigation action to mitigate effects of the anomaly.

Abstract: The technologies described herein are generally directed to managing a network using multi-agent, adaptive machine learning. For instance, a system can, based on a performance objective of a geographic area, a global policy model resulting in model training data. The example system may further include communicating the model training data to a base station of the base stations of the geographic area. The example system may further include, receiving from the base station, a model gradient value applicable to modifying the global policy model, with the model gradient value being identified based on an action implemented by the base station that was selected based on a base station maintained by the base station, that was trained based on the model training data and a local performance objective applicable to the base station.

Abstract: The technologies described herein are generally directed to using machine learning to prioritize wireless reassociation. For instance, a system can identify state information corresponding to a candidate computing device of a network, wherein the candidate computing device is hosting a candidate communication session to be handed over from a first access point of the network. The system may further, based on the state information, change, employing a machine learning model, a handover priority score of the candidate communication session, wherein the machine learning model was trained based on data representative of past network events in past communication sessions of the computing devices of the network, resulting in a changed handover priority score. The system may further, based on the changed handover priority score, cause a handover of the candidate communication session from the first access point to a second access point of the network.

Abstract: The technologies described herein are generally directed to using machine learning to identify, analyze, and select storage resources for auxiliary data associated with a system event. For instance, a system can identify a network event in an operation of a computing node of a network. The example method may further include identifying, by employing a machine learning model, respective auxiliary data including respective sources of auxiliary data representative of operations of dependent computing nodes of the computing nodes, different from the computing node, that have dependencies upon the operation of the computing node. In the example method, the machine learning model was trained based on data representative of past network events. The example method may further include, based on the respective classifications of the respective auxiliary data, facilitating, by the computing system, storage of the respective auxiliary data at a storage location of a hierarchical storage system.

Abstract: Facilitating dynamic network traffic steering using artificial intelligence in advanced communication networks is provided. A method includes determining respective results of application of a utility function to respective combinations of potential handovers of a specified user equipment from a source cell to respective target cells of a group of target cells. The method also includes, based on the respective results of the application of the utility function, determining that a first combination of the respective combinations increases a value of the utility function as compared to other combinations of the respective combinations, other than the first combination. Further, the method includes, during a defined interval associated with a traffic steering process, facilitating the handover of the specified user equipment from the source cell to the first target cell. Other user equipment other than the specified user equipment within the communication network are not handed over during the defined interval.

Abstract: A method can comprise, as part of initially deploying a second radio frequency (RF) transceiver, transferring, by a system, a core model to the second RF transceiver, the core model having been trained based on a training process comprising training an artificial intelligence model for a first RF transceiver, based on first data that is measured for the first RF transceiver. The method can further comprise applying, by the system, transfer learning on the core model at the second RF transceiver based on second data that is measured for the second RF transceiver, to produce a trained model. The method can further comprise calibrating, by the system, the second RF transceiver based on an output of the trained model to produce a calibrated second RF transceiver. The method can further comprise transmitting, by the system, RF information via the calibrated second RF transceiver.

Abstract: A method can comprise allocating, by a system, an adaptive cell-specific bandwidth part for facilitation of cellular network communications, wherein the adaptive cell-specific bandwidth part comprises a group of bandwidth sizes that enable different energy consumption by the cellular network. The method can further comprise transitioning, by the system, from a first bandwidth size to a second bandwidth size, wherein the second bandwidth size is smaller, based on determining that the second bandwidth size is sufficient to serve a predicted amount of network traffic. The method can further comprise, after transitioning to the second bandwidth size, transitioning, by the system, from the second bandwidth size to the first bandwidth size, based on determining that a second criterion has been satisfied. The method can further comprise, after transitioning from the second bandwidth size to the first bandwidth size, facilitating, by the system, cellular network communications with the first bandwidth size.

Abstract: Facilitating energy aware multi-cell admission control in advanced communication networks is provided herein. A method includes, based on notification of a denial of an admission request by a first cell, facilitating a transmission of the admission request to a second cell and at least a third cell. The second cell and the third cell are neighbor cells of the first cell. The admission request can be received from a user equipment. The method also includes, based on receipt of an acceptance of the admission request from the second cell and the third cell, and based on a determination that a first utility of the second cell is higher than a second utility of at least the third cell, selecting the second cell as an admission cell for the user equipment. Further, the method includes facilitating admission of the user equipment to a cell cluster via the second cell.

Abstract: The technologies described herein are generally directed toward using differential checkpoints to generate synthetic wireless traffic data. For instance, a system can obtain training data based on a time series data representation generated based on measured wireless traffic in a wireless coverage area. The system can further, based on the training data and an output of a discriminator, train a generator to generate synthetic wireless traffic data, resulting in a trained generator. Further, the system can, based on an output of the generator, train the discriminator to detect the synthetic wireless traffic data, with a generative adversarial network being deployed for the wireless coverage area that may include the generator and the discriminator.

Abstract: Facilitating concurrent energy aware admission control, dynamic load balancing, and traffic steering in advanced communication networks is provided. A method includes facilitating energy efficiency aware load balancing of already served user equipment to distribute the already served user equipment among a group of cells of a communication network. The method also includes facilitating controlling of admission of incoming user equipment to the communication network and facilitating traffic steering of the incoming user equipment among the group of cells. Further, the method includes facilitating conflict mitigation among respective policies associated with facilitating the energy efficiency aware load balancing, facilitating of the controlling of the admission, and facilitating of the traffic steering.

Abstract: A method can comprise determining, by a system, a first sleep mode action for a first group of cellular base station signal metrics for a cellular base station over a first time period, wherein the first group of cellular base station signal metrics is determined based on a first layer of communications of the cellular base station. The method can further comprise determining, by the system, a second sleep mode action for a second group of cellular base station signal metrics for the cellular base station over a second time period. The method can further comprise determining, by the system, an arbitrated sleep mode action based on the first sleep mode action and the second sleep mode action. The method can further comprise sleeping, by the system, at least part of the cellular base station based on the arbitrated sleep mode action.

Abstract: The technology described herein is directed towards accurate low-complexity decision tree-based traffic predictor models, such as decision tree regressor models for use by base stations. Each model is rapidly retrained per base station using data relevant to the base station. To improve accuracy, statistically significant feature data is selected by performing hypothesis testing on candidate features to filter out features that cannot satisfy a statistical significance threshold (e.g., p-value). The decision tree regressor model is recursively grown based on the selected features' feature values and their traffic level labels. Predicted traffic level data is determined by traversing the trained decision tree to reach a leaf node associated with the prediction data. Resource allocation can be based on the prediction. In addition to time-series training data, spatial training data can be used. Real time traffic monitoring by a radio unit for operating in an autonomous management mode is also facilitated.

Abstract: Use of a data-driven approach that assimilates historical signal to interference-plus noise ratio (SINR) and channel estimation data along with location-map of the cell in which base station equipment is situated to better define the relationship between SINR and the user-channel environmental map and spatio-temporal changes to it to achieve more granular, cell site-specific modeling is disclosed herein. This data-driven approach estimates SINR using variational autoencoders. Variational encoders typically consist of two sections, an encoder section and decoder section. The encoder section learns the distribution on the low-dimensional latent space over the input data samples. The decoder section is a generative model that learns the joint distribution of the latent variables and input data.

Abstract: Use of a data-driven approach that assimilates historical signal to interference-plus noise ratio (SINR) and channel estimation data along with location-map of the cell in which base station equipment is situated to better define the relationship between SINR and the user-channel environmental map and spatio-temporal changes to it to achieve more granular, cell site-specific modeling is disclosed herein. This data-driven approach estimates SINR using variational autoencoders. Variational encoders typically consist of two sections, an encoder section and decoder section. The encoder section learns the distribution on the low-dimensional latent space over the input data samples. The decoder section is a generative model that learns the joint distribution of the latent variables and input data.

Abstract: Facilitating cell and carrier switch off for energy awareness in advanced communication networks is provided herein. A method includes determining a first result of a utility function associated with a first configuration of a set of carriers that service a group of user equipment in a communication network. The first configuration is based on respective activation states of carriers of the set of carriers. The method also includes, based on respective traffic patterns of user equipment of the group of user equipment, evaluating respective results of the utility function for respective configurations of a group of configurations, other than the first configuration, for the set of carriers. Further, the method includes selecting a second configuration from the group of configurations based on a second result of the utility function for the second configuration being determined to be a higher value than a value of the first result.

Abstract: A system can produce a first graph that represents components of a broadband cellular network, wherein the first graph comprises first nodes and first edges. The system can process the first graph with a graph neural network to produce a feature embedding matrix. The system can pool information of nodes of the first graph based on the feature embedding matrix, to produce a second graph, wherein the second graph comprises second nodes and second edges. The system can adjust a parameter of a network controller of the broadband cellular network based on the second graph.

Abstract: Facilitating Multiple-Tenant Energy Efficient Radio Access Network Sharing In Advanced Communication Networks is provided. A method includes facilitating, by a system comprising at least one processor, network energy savings in a communications network that is deployed in a shared radio access network architecture. The facilitating includes implementing, at a network operator level of the communications network according to a defined energy efficiency criterion, energy efficient scheduling of user equipment within the communications network. The facilitating also includes implementing, at an infrastructure provider level of the communications network, network energy savings actions based on respective measured quality of service levels of the user equipment being retained at or above a defined quality of service level. In an example, network equipment comprised by the communications network is configured to operate according to at least a fifth generation radio network communication protocol.