T-mobile Usa, Inc.

Vijay Veggalam

Suchit Satpathy

Florendo Atienza

Jeremy Babb

Kendall Bush

Nirmal Chandrasekaran

Prabha Jayaram

Jonathan Silberlicht

Sakshi Shori

Machine learning systems and techniques are described herein for monitoring application servers within the computing infrastructure of a wireless network. A machine learning system may train and execute models based on performance metrics, error data, and failures from application servers configured to perform various functions relating to communication session authorization, monitoring, and charging for voice, data, and messaging communication services over the wireless network. Machine learning models may be configured to predict service degradations, detect performance anomalies, and determine root causes, and various models may output particular application servers, components, and/or predicted times of certain events. Based on the outputs of the machine learning models, the machine learning system may perform various actions for the wireless network infrastructure, including data outputting and/or executing processes on the application servers to analyze, restart, and/or repair particular components.

Machine learning monitoring of wireless network infrastructure application servers

Machine-learning based techniques are described herein for determining and modifying handover parameters within multilayer wireless networks. Various communication session data, such as key performance indicators, may be analyzed and compared at multiple frequency layers to determine sets of custom parameters associated with one or more wireless networks. The sets of custom parameters and network performance data may be used to train one or more machine-learned models to improve and/or optimize the handover parameters used by the network nodes. In some examples, different trained models may be associated with different network performance metrics, such as throughput optimization, network speed, and/or dropped call minimization, etc. A trained machine-learned model may be used to analyze the session data from a set of network nodes, and to determine or tune the handover parameters used by the network nodes.

Machine-learning frequency layer management and handover parameter tuning

A service management system can obtain a service request message indicating an event that affects a user device of a telecommunications network. The system can store an indication of the event at a memory device that stores historical information of events and can process the event with a machine learning (ML) model that is trained based on the historical information to simulate the effect of the event on the telecommunications network. The system generates an output for display on a computing device. The output includes a recommendation and identifies an attribute of the event and a predicted value configured to reduce a workload or increase operational efficiency. The recommendation is configured to make the event actionable. The system receives feedback that a reviewer has acted on the recommendation. In response to the feedback, the system can re-train the ML model and configure the process for handling the service request.

Recommendation engine with machine learning for guided service management, such as for use with events related to telecommunications subscribers

This disclosure describes techniques for optimizing operational performance of femtocell devices deployed within a telecommunications network. More particularly, a femtocell Optimization (FCO) system is described that monitors a performance state of femtocell devices within a telecommunications network, detects a change in performance state of a particular femtocell device, and preemptively deploys a first set of service resolution data to detect a service issue. In response to detecting a service issue, a second set of service resolution data is deployed to preemptively resolve the service issue.

Femtocell provisioning and service issue optimization

Techniques are described herein for evaluating customer support sessions to identify active listening. The techniques include receiving a recording of a customer service support session from a customer support terminal, which is associated with a representative identifier of a customer service representative. The customer service support session includes one or more segments, and a plurality of active listening factor scores is calculated and averaged to determine an active listening factor score for each of the one or more segments. The plurality of active listening factor scores corresponds to different active listening attributes demonstrated by the customer service representative. The active listening factor score for each of the one or more segments are averaged to determine a session score, which is associated with the representative identifier. The session score can be used to present one or more support suggestions for the customer service representative in a support tool.

Active listening using artificial intelligence for communications evaluation

This disclosure describes techniques for optimizing operational performance of femtocell devices deployed within a telecommunications network. More particularly, a femtocell Optimization (FCO) system is described that detects service issues that affect femtocell device(s). In some instances, the FCO system may automatically deploy a service resolution for known service issues or alerts network representatives for unresolved service issues. Moreover, the FCO system may be further configured to optimize resource utilization of a femtocell device network by preemptively correcting femtocell provisioning issues, based on patterns of historic service issues. The FCO system may further optimize resource usage of femtocell devices by selectively intercepting and terminating connectivity requests (i.e. rejecting issuance of Internet Protocol (IP) addresses) from client devices, or client device types, that have not historically overwhelmed femtocell network resources. The FCO system may further generate a femtocell analysis model to preemptively deploy service resolutions that protect against probable degradations in performance.

Growing community of inventors

Morris Plains, NJ, United States of America

Vijay Veggalam