System, method, and computer program for preparing a multi-stage framework for artificial intelligence (AI) analysis
A system, method, and computer program product are provided for preparing a multi-stage framework for artificial intelligence (AI) analysis. In use, a first set of monitoring rules used by at least one network entity of a communication network is defined. First event log data of first network activity is collected based on the first monitoring rules, and at least one first network situation is defined. Additionally, at least one first AI model is computed based on the first event log data and the at least one first network situation. A second set of monitoring rules used by the at least one network entity is defined. Second event log data of the first network activity is collected based on the second monitoring rules. Further, at least one second AI model is computed based on the second event log data and the at least one first network situation.
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:
define a first set of monitoring rules used by at least one network entity of a communication network, wherein the first set of monitoring rules cause event log data to be collected with a first level of detail;
collect first event log data of first network activity of the communication network based on the first monitoring rules, the first event log data having the first level of detail;
define a first network situation corresponding to the network activity;
compute a first artificial intelligence (AI) model based on the first event log data having the first level of detail and the first network situation, wherein the first AI model is configured to predict the first network situation from the first event log data with:
a first confidence level that the first network situation will be correctly predicted, and
a first lead-time;
determine that the first confidence level achieves an upper confidence level goal;
determine that a second AI model with a lead-time longer than the first lead-time is to be computed, as a result of determining that the first confidence level achieves the upper confidence level goal;
responsive to determining that the second AI model with the lead-time longer than the first lead-time is to be computed, define a second set of monitoring rules used by the at least one network entity, wherein the second set of monitoring rules cause event log data to be collected with a second level of detail that is lower than the first level of detail;
collect second event log data of the first network activity of the communication network based on the second monitoring rules, the second event log data having the second level of detail; and
compute the second AI model based on the second event log data having the second level of detail and the first network situation, wherein the second AI model is configured for detecting the first network situation in the second event log data with:
a second confidence level that is lower than the first confidence level and that is higher than a minimal predetermined confidence level, and
a second lead-time that is longer than the first lead-time.
2. The computer program product of claim 1 , wherein the computer program product is configured to use the first AI model to detect the first network activity.
3. The computer program product of claim 1 , wherein the computer program product is configured to use the second AI model to detect the first network activity.
4. The computer program product of claim 1 , wherein the first AI model detects a first classifier preceding the first network situation, where the first classifier includes at least one first parameter found in first prior event log data collected before the first network situation by a first time period.
5. The computer program product of claim 4 , wherein the second AI model detects a second classifier preceding the first network situation, where the second classifier includes at least one second parameter found in second prior log data collected before the first network situation by a second time period.
6. The computer program product of claim 5 , wherein the first time period is smaller than the second time period.
7. A method, comprising:
defining a first set of monitoring rules used by at least one network entity of a communication network, wherein the first set of monitoring rules cause event log data to be collected with a first level of detail;
collecting first event log data of first network activity of the communication network based on the first monitoring rules, the first event log data having the first level of detail;
defining define a first network situation corresponding to the network activity;
computing a first artificial intelligence (AI) model based on the first event log data having the first level of detail and the first network situation, wherein the first AI model is configured to predict the first network situation from the first event log data with:
a first confidence level that the first network situation will be correctly predicted, and
a first lead-time;
determining that the first confidence level achieves an upper confidence level goal;
determining that a second AI model with a lead-time longer than the first lead-time is to be computed, as a result of determining that the first confidence level achieves the upper confidence level goal;
responsive to determining that the second AI model with the lead-time longer than the first lead-time is to be computed, defining a second set of monitoring rules used by the at least one network entity, wherein the second set of monitoring rules cause event log data to be collected with a second level of detail that is lower than the first level of detail;
collecting second event log data of the first network activity of the communication network based on the second monitoring rules, the second event log data having the second level of detail; and
computing the second AI model based on the second event log data having the second level of detail and the first network situation, wherein the second AI model is configured for detecting the first network situation in the second event log data with:
a second confidence level that is lower than the first confidence level and that is higher than a minimal predetermined confidence level, and
a second lead-time that is longer than the first lead-time.
8. 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:
define a first set of monitoring rules used by at least one network entity of a communication network, wherein the first set of monitoring rules cause event log data to be collected with a first level of detail;
collect first event log data of first network activity of the communication network based on the first monitoring rules, the first event log data having the first level of detail;
define a first network situation corresponding to the network activity;
compute a first artificial intelligence (AI) model based on the first event log data having the first level of detail and the first network situation wherein the first AI model is configured to predict the first network situation from the first event log data with:
a first confidence level that the first network situation will be correctly predicted, and
a first lead-time;
determine that the first confidence level achieves an upper confidence level goal;
determine that a second AI model with a lead-time longer than the first lead-time is to be computed, as a result of determining that the first confidence level achieves the upper confidence level goal;
responsive to determining that the second AI model with the lead-time longer than the first lead-time is to be computed, define a second set of monitoring rules used by the at least one network entity, wherein the second set of monitoring rules cause event log data to be collected with a second level of detail that is lower than the first level of detail;
collect second event log data of the first network activity of the communication network based on the second monitoring rules, the second event log data having the second level of detail; and
compute the second AI model based on the second event log data having the second level of detail and the first network situation, wherein the second AI model is configured for detecting the first network situation in the second event log data with:
a second confidence level that is lower than the first confidence level and that is higher than a minimal predetermined confidence level, and
a second lead-time that is longer than the first lead-time.
9. The computer program product of claim 1 , further comprising:
determine whether the second confidence level reaches a lower confidence level goal and whether the second lead-time is greater than a minimum reconfiguration time; and
determine that a third AI model is to be computed, as a result of determining that the second confidence level does not reach the lower confidence level goal and that the second lead-time is greater than the minimum reconfiguration time.
10. The computer program product of claim 9 , further comprising:
responsive to determining that the third AI model with the lead-time longer than the first lead-time is to be computed, define a third set of monitoring rules used by the at least one network entity, wherein the third set of monitoring rules cause event log data to be collected with a third level of detail that is lower than the second level of detail;
collect third event log data of the first network activity of the communication network based on the third monitoring rules, the third event log data having the third level of detail; and
compute the third AI model based on the third event log data having the third level of detail and the first network situation, wherein the third AI model is configured for detecting the first network situation in the third event log data with:
a third confidence level that is lower than the second confidence level and that is higher than the minimal predetermined confidence level, and
a third lead-time that is longer than the first lead-time.
11. The computer program product of claim 1 , wherein the lead-time includes a time period between a first time of a detection of a classifier in event log data and a second time that the first network situation is predicted to occur.