System, method, and computer program for operating multi-feed of log-data in an AI-managed communication system

Patent metadata

Field	Value
Publication number	US-11271822-B1
Application number	US-201916280367-A
Country	US
Kind code	B1
Filing date	Feb 20, 2019
Priority date	Mar 7, 2018
Publication date	Mar 8, 2022
Grant date	Mar 8, 2022

Abstract

A system, method, and computer program product are provided for operating multi-feed of log data in an AI-managed communication system. In use, an identification of at least one artificial intelligence (AI) system and an identification of at least one AI model of a plurality of AI models used by the AI system are obtained. Additionally, a stream of log data is received, and a log data feed adapted to the AI model is created. Further, the log data feed is communicated using a corresponding AI model of the plurality of AI models.

First claim

What is claimed is:

1. A computer program product comprising computer executable instructions stored on a non-transitory computer readable medium that when executed by a processor instruct the processor to:
obtain an identification of at least one artificial intelligence (AI) system and an identification of at least one AI model of a plurality of AI models used by the at least one AI system;
receive a stream of log data;
create a log data feed adapted to the at least one AI model;
communicate the log data feed using a corresponding AI model of the plurality of AI models;
compute a first confidence level for the at least one AI model, the first confidence level representing at least one of:
a probability of the at least one AI model for detecting a first classifier of the log data feed, or
a probability of the at least one AI model for detecting a network situation,
wherein the first classifier precedes the network situation;

eliminate at least one parameter from the log data feed to form a second log data feed;
analyze the second log data feed using a corresponding second AI model;
compute a second confidence level for the corresponding second AI model, the second confidence level representing at least one of:
a probability of the corresponding second AI model for detecting a second classifier of the second log data feed, or
a probability of the corresponding second AI model for detecting a second network situation;

determine that the second confidence level is higher than the first confidence level; and
eliminate at least one second parameter from the second log data feed to form a third log data feed;
analyze the third log data feed using a corresponding third AI model;
compute a third confidence level for the corresponding third AI model, the third confidence level representing at least one of:
a probability of the corresponding third AI model for detecting a third classifier of the third log data feed, or
a probability of the corresponding third AI model for detecting a third network situation;

determine that the third confidence level is higher than the second confidence level; and
eliminate at least one third parameter from the third log data feed to form a fourth log data feed.

2. The computer program product of claim 1 , wherein the corresponding AI model is configured to detect the first classifier in the log data feed.

3. The computer program product of claim 2 , wherein the log data includes parameters associated with an operation of network entities of a communication network and a time of reporting.

4. The computer program product of claim 1 , wherein the log data feed includes parameters associated with the corresponding AI model.

5. The computer program product of claim 1 , wherein the computer program product is configured to obtain, for the corresponding AI model, a plurality of parameters of the log data which are processed by the corresponding AI model.

6. The computer program product of claim 1 , wherein the computer program product is configured to refine the log data feed to include only parameters processed by the corresponding AI model.

7. The computer program product of claim 1 , wherein the computer program product is configured to simultaneously operate the plurality of AI models, each of the AI models processing a corresponding log data feed.

8. A computer program product comprising computer executable instructions stored on a non-transitory computer readable medium that when executed by a processor instruct the processor to:
obtain an identification of at least one artificial intelligence (AI) system and an identification of at least one AI model of a plurality of AI models used by the at least one AI system;
receive a stream of log data;
create a log data feed adapted to the at least one AI model;
communicate the log data feed using a corresponding AI model of the plurality of AI models;
identify a network situation based on the log data;
identify a resolution level;
initiate, based on the network situation and the resolution level, a feed channel of the log data feed for a first particular AI model, the feed channel including corresponding monitoring rules and corresponding parameter characterization; and
repeat the initiation step for all AI models of the plurality of AI models that correspond to the network situation and the resolution level.

9. A method, comprising:
obtaining an identification of at least one artificial intelligence (AI) system and an identification of at least one AI model of a plurality of AI models used by the at least one AI system;
receiving a stream of log data;
creating a log data feed adapted to the at least one AI model;
communicating the log data feed using a corresponding AI model of the plurality of AI models;
computing a first confidence level for the at least one AI model, the first confidence level representing at least one of:
a probability of the at least one AI model for detecting a first classifier of the log data feed, or
a probability of the at least one AI model for detecting a network situation,
wherein the first classifier precedes the network situation;

eliminating at least one parameter from the log data feed to form a second log data feed;
analyzing the second log data feed using a corresponding second AI model;
computing a second confidence level for the corresponding second AI model, the second confidence level representing at least one of:
a probability of the corresponding second AI model for detecting a second classifier of the second log data feed, or
a probability of the corresponding second AI model for detecting a second network situation;

determining that the second confidence level is higher than the first confidence level; and
eliminating at least one second parameter from the second log data feed to form a third log data feed;
analyzing the third log data feed using a corresponding third AI model;
computing a third confidence level for the corresponding third AI model, the third confidence level representing at least one of:
a probability of the corresponding third AI model for detecting a third classifier of the third log data feed, or
a probability of the corresponding third AI model for detecting a third network situation;

determining that the third confidence level is higher than the second confidence level; and
eliminating at least one third parameter from the third log data feed to form a fourth log data feed.

10. A device, comprising:
a non-transitory memory storing instructions; and
one or more processors in communication with the non-transitory memory, wherein the one or more processors execute the instructions to:
obtain an identification of at least one artificial intelligence (AD system and an identification of at least one AI model of a plurality of AI models used by the at least one AI system;
receive a stream of log data;
create a log data feed adapted to the at least one AI model;
communicate the log data feed using a corresponding AI model of the plurality of AI models;
compute a first confidence level for the at least one AI model, the first confidence level representing at least one of:
a probability of the at least one AI model for detecting a first classifier of the log data feed, or
a probability of the at least one AI model for detecting a network situation,
wherein the first classifier precedes the network situation;

eliminate at least one parameter from the log data feed to form a second log data feed;
analyze the second log data feed using a corresponding second AI model;
compute a second confidence level for the corresponding second AI model, the second confidence level representing at least one of:
a

Assignees

• Amdocs Development Ltd

Inventors

• Hermoni Ofer
• Sandlerman Nimrod
• Felstaine Eyal

Classifications

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title

• H04L41/0803 Primary

  Configuration setting · CPC title

• H04L67/562

  Brokering proxy services · CPC title

• H04L63/1425

  Traffic logging, e.g. anomaly detection · CPC title

• H04L67/1001

  for accessing one among a plurality of replicated servers · CPC title