Network operators may control many thousands of devices and, as a consequence, they may be bombarded with notifications (e.g., posture assessments and exceptions for devices that do not comply with standards or which require remediation actions to circumvent security issues) and can be discouraged by unimportant or irrelevant information. To address the challenge that was described above, techniques are presented herein that support a process for intelligently filtering and prioritizing notifications to reduce noise and deliver recommendations that will provide the greatest impact to an environment. Aspects of the presented techniques encompass a smart recommendation notification system that prioritizes network actions based on multiple embedding spaces and dimensions; the use of business and financial logic, a persona, and a network operating state for reducing network actions into a prioritized output; the use of click-through and sequence mining to establish a ground truth of event prioritization of a network operator; an adaptive learning method for tracking proposed network recommendations to the final action that may be executed by a network operator; and a method for reducing multi-step recommendations based on an identification of the most efficient sequence of events based on a network operator implementation.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Sun, Pengfei; Shao, Qihong; and Engi, Derek W., "DEEPSORTING: CUSTOMER-CENTRIC SEQUENCE-TO-SEQUENCE NOTIFICATION PRIORITIZATION", Technical Disclosure Commons, (January 30, 2023)