Abstract: Systems and methods for identifying correlations for adaptive noise reduction. The system obtains a set of network performance metrics measured during a period of time from a sensor module communicatively coupled to a set of sensors deployed at remote devices. The system may input the set of network performance metrics into a trained model to obtain an optimal set of network performance metrics to be measured during a next period of time, wherein the model is configured to determine a next set of metrics to be measured based on a previous set of metrics. The system identifies a second set of sensors for usage and may generate one or more commands configured to effectuate activation of one or more sensors of the second set of sensors, and deactivation of one or more sensors of the first set of sensors.

Abstract: At a high level, the technology disclosed herein relates to base station antenna tilt adjustments based on using one or more auto-tilt optimization models. In embodiments, baseline tilt data and user equipment (UE) measurement data may be used for generating training data for the one or more auto-tilt optimization models. The baseline tilt data and UE measurement data may correspond to a plurality of sectors for a first base station, a plurality of sectors for a second base station, etc. The one or more auto-tilt optimization models may provide key performance indicator (KPI) predictions associated with base station antenna tilt adjustments, and the KPI predictions can be updated as feedback is received upon tilt adjustments made to one or more antennas.

Abstract: System and methods for selecting customers to notify of a service interruption or outage are described. A network may receive an interruption notification indicating when a service interruption is to occur. In some examples, the interruption notification may also include a particular set of equipment, such as a base station, that will experience the interruption, such as being taken offline for maintenance or a power outage affecting the base station. A notification module determines which customers will be affected by the service interruption and transmits a notification to the customers, informing the customers of the service interruption.

Abstract: Techniques related to identifying a component causing an issue in a wireless telecommunication network are disclosed. In one example aspect, a method for wireless communication includes creating a knowledge graph representing dependencies among components in the wireless telecommunication network. In response to obtaining an indication of the issue in the wireless telecommunication network, the indication of the issue and the knowledge graph are provided to a trained machine learning (ML) model, and the trained ML model indicates a particular component among the components in the wireless telecommunication network that is likely causing the issue.

Abstract: A method of initiating a KPI backfill operation for a cellular network based on detecting a network data anomaly, where the method includes receiving, by an anomaly detection and backfill engine (ADBE) executed by a computing device, a data quality metric that is based on a KPI of the cellular network; detecting, by the ADBE, the network data anomaly based on the data quality metric being more than a threshold amount different than a predicted value for the data quality metric, where the network data anomaly indicates that at least a portion of a data stream from which the KPI is calculated was unavailable for a previous iteration of the KPI; and providing, by the ADBE and based on detecting the network data anomaly, a backfill command to a backfill processing pipeline to perform the backfill operation by reaggregating the KPI when the portion of the data stream becomes available.

Abstract: System and methods for selecting customers to notify of a service interruption or outage are described. A network may receive an interruption notification indicating when a service interruption is to occur. In some examples, the interruption notification may also include a particular set of equipment, such as a base station, that will experience the interruption, such as being taken offline for maintenance or a power outage affecting the base station. A notification module determines which customers will be affected by the service interruption and transmits a notification to the customers, informing the customers of the service interruption.

Abstract: A method for managing an ontology of a data handling facility of a communication system. The method includes discovering connections between some of a plurality of electrically powered components of the data handling facility based on time-series data obtained from a plurality of electrical power monitors, and forming a physical ontology layer of the data handling facility and a logical ontology layer of the data handling facility. In addition, the method includes receiving a query from a user concerning a hypothetical modification to the operation of the data handling facility, and forecasting change in the operation of the data handling facility based on the query received from the user.

Abstract: A method is disclosed for providing a database of subscriber interactions with a network. The database is configured for geographical searching. A provider record relating to a transaction between a network node and a subscriber is received. A first period provider records collection including provider records from a first predetermined time period is stored. The first period provider records collection is converted into a first period unified records collection. A second period unified records collection is stored having unified records from a second predetermined time period. The second predetermined time period is longer than the first predetermined time period. The second period unified records collection includes sub-period partitions with unified records having timestamps in a first time sub-period.

Abstract: Systems and methods for identifying correlations for adaptive noise reduction. The system obtains a set of network performance metrics measured during a period of time from a sensor module communicatively coupled to a set of sensors deployed at remote devices. The system may input the set of network performance metrics into a trained model to obtain an optimal set of network performance metrics to be measured during a next period of time, wherein the model is configured to determine a next set of metrics to be measured based on a previous set of metrics. The system identifies a second set of sensors for usage and may generate one or more commands configured to effectuate activation of one or more sensors of the second set of sensors, and deactivation of one or more sensors of the first set of sensors.

Abstract: Systems and methods are provided for providing a zero touch operations and self-healing data platform. Initially, data corresponding to a mobile communications network is received from a plurality of sources. A machine learning model is trained to detect anomalies corresponding to the data. Upon detecting an anomaly, self-remediation is initiated. Upon determining the self-remediation exceeds a configurable number of attempts or duration, an action is automatically performed. In various aspects, the action includes automatically initiating an alert corresponding to the anomaly or automatically creating a ticket corresponding to the anomaly.

Abstract: According to aspects herein, methods and systems for enriching location session record data are provided. More particularly, missing data of the location session record, such as an IMSI or MSISDN, is enriched with data from AMF probes. Initially, a location session record (LSR) dataset is received from an LSR probe. At least a portion of data from an access and mobility management function (AMF) dataset is compared to at least a portion of data from an LSR dataset to identify a match between a record in the AMF dataset and a corresponding record in the LSR dataset. Upon identifying the match, the corresponding record in the LSR dataset is enriched with IMSI or MSISDN from the record in the AMF dataset.

Abstract: The system obtains a corpus of data, and extracts triples from the corpus. A first element in a triple indicates a first record in the corpus, a second element in the triple indicates a second record in the corpus, and a third element in the triple indicates a relationship between the first and second records. The system generates grammars representing the triples. A grammar includes concepts and relationships. The concepts include a first and a second concept representing the first and the second record, respectively. The relationships represent the relationship between the first and second records. The system applies each grammar to the triples to obtain an indication of whether each triple is correct. Based on the indication of whether each triple is correct, the system determines an accuracy of each grammar. Based on the accuracy of each grammar, the system selects a grammar having the highest accuracy.

Abstract: A machine learning system automatically diagnoses and resolves issues in a telecommunications network. When a customer reports a network issue using a mobile application, the device performs a diagnostic test, such as a speed test. In addition, network logs or performance metrics during occurrence of the network issue are collected. The results of the diagnostic are used as inputs to a machine learning model in combination with the network logs or metrics to predict the cause of the network issue or to perform a corrective action.

Abstract: System and methods for selecting customers to notify of a service interruption or outage are described. A network may receive an interruption notification indicating when a service interruption is to occur. In some examples, the interruption notification may also include a particular set of equipment, such as a base station, that will experience the interruption, such as being taken offline for maintenance or a power outage affecting the base station. A notification module determines which customers will be affected by the service interruption and transmits a notification to the customers, informing the customers of the service interruption.

Abstract: Techniques related to identifying a component causing an issue in a wireless telecommunication network are disclosed. In one example aspect, a method for wireless communication includes creating a knowledge graph representing dependencies among components in the wireless telecommunication network. In response to obtaining an indication of the issue in the wireless telecommunication network, the indication of the issue and the knowledge graph are provided to a trained machine learning (ML) model, and the trained ML model indicates a particular component among the components in the wireless telecommunication network that is likely causing the issue.

Abstract: A method of initiating a KPI backfill operation for a cellular network based on detecting a network data anomaly, where the method includes receiving, by an anomaly detection and backfill engine (ADBE) executed by a computing device, a data quality metric that is based on a KPI of the cellular network; detecting, by the ADBE, the network data anomaly based on the data quality metric being more than a threshold amount different than a predicted value for the data quality metric, where the network data anomaly indicates that at least a portion of a data stream from which the KPI is calculated was unavailable for a previous iteration of the KPI; and providing, by the ADBE and based on detecting the network data anomaly, a backfill command to a backfill processing pipeline to perform the backfill operation by reaggregating the KPI when the portion of the data stream becomes available.

Abstract: A machine learning system automatically diagnoses and resolves issues in a telecommunications network. When a customer reports a network issue using a mobile application, the device performs a diagnostic test, such as a speed test. In addition, network logs or performance metrics during occurrence of the network issue are collected. The results of the diagnostic are used as inputs to a machine learning model in combination with the network logs or metrics to predict the cause of the network issue or to perform a corrective action.

Abstract: Embodiments for monitoring and rebooting a node of a wireless communications network are provided. Embodiments may include acquiring an incident indication for the node, the incident indication being indicative of a performance drop associated with a communication service provided by the node. Additionally, the incident indication may be analyzed to identify at least one key term or to identify a drop throughput that is comparable to historical incident indications. In some embodiments, the incident indication may include an indication that a reset of the node failed to improve the performance drop. In some embodiments, one or more of the historical incident indications may include an indication that a reset of the node failed to improve the performance drop. Further, a recommendation can be provided based on the incident indication and, in some embodiments, a resolution may be received based on providing the recommendation.

Abstract: A machine learning system automatically diagnoses and resolves issues in a telecommunications network. When a customer reports a network issue using a mobile application, the device performs a diagnostic test, such as a speed test. In addition, network logs or performance metrics during occurrence of the network issue are collected. The results of the diagnostic are used as inputs to a machine learning model in combination with the network logs or metrics to predict the cause of the network issue or to perform a corrective action.

Abstract: In various examples, upon receiving data comprising a plurality of key performance indicators (KPIs), an anomaly may be detected in a first KPI. An engine may apply rules associating two or more KPIs together to confirm, reject, or narrow the anomaly in the first KPI. Through this technology, the mean time for detecting an anomaly in a network may be reduced thereby preventing the potential for more serious anomalies in the network.