CPC H04L 41/16 (2013.01) [G01R 19/2513 (2013.01); G06N 7/01 (2023.01); G06N 20/00 (2019.01); H04L 67/12 (2013.01)] | 21 Claims |
1. A computer-implemented method of training a machine learning model, comprising:
receiving an island-level event data describing an island-level event occurring at an island that is capable of dynamically changing by splitting or merging with another island, wherein:
the island is a part of a network comprising a plurality of islands including the another island and comprises one or more individual components; and
the island-level event data fails to immediately identify a source individual component of the island-level event from the one or more individual components;
estimating a prior probability of each of the one or more individual components causing the island-level event;
performing a posterior inference, using the machine learning model and based on the island-level event data and the prior probability of each of the one or more individual components, to determine a probability estimate of each of the one or more individual components causing the island-level event;
updating the prior probability of each of the one or more individual components based on a result of the performance of the posterior inference; and
executing the receiving the island-level event data, the performing the posterior inference, and the updating the prior probability repeatedly until a difference between the updated and estimated prior probability of each of the one or more individual components is less than a threshold difference,
wherein, after the executing, the machine learning model is trained to receive the island-level event data describing the island-level event occurring at the island, and identify the source individual component as a cause of the island-level event based on the received island-level event data as the island is dynamically changing.
|