Detecting data anomalies using artificial intelligence
The systems and methods disclosed herein receive a dataset including an observed set of values for a set of variables. The system can use a first set of AI models to identify a set of anomalies in the observed set of values by comparing an observed set of patterns against multiple reference patterns. The system can use a second set of AI models to evaluate the identified anomalies by comparing an observed set of association rules with an expected set of association rules. The system can use a third set of AI models to generate reconfiguration commands to remove the identified anomalies. The reconfiguration commands can be automatically executed to modify the observed association rules to align with the expected association rules.
1. A computer-implemented method for improving data quality using an artificial intelligence (AI) model, the method comprising:
receiving a structured dataset including an observed set of values for each variable in a variable set;
identifying, using a first AI model set, an anomaly set from the observed set of values in the structured dataset by:
determining multiple reference patterns that correspond to an expected set of values for the variable set, and
comparing an observed pattern set from the observed set of values against the multiple reference patterns;
using a second AI model set to evaluate the identified anomaly set by:
dynamically generating an observed association rule set configured to cause the second AI model set to generate the received observed set of values in the structured dataset, and
comparing the observed association rule set with an expected association rule set to determine one or more observed association rules corresponding to the anomaly set;
using a third AI model set, generating a reconfiguration command set configured to modify the one or more observed association rules corresponding to the anomaly set into alignment with one or more expected association rules by:
identifying a portion of the observed set of values that corresponds to the one or more observed association rules of the observed association rule set that corresponds to the anomaly set, and
mapping the portion of the observed set of values to one or more expected association rules of the expected association rule set, the one or more expected association rules being configured to adjust the portion of the observed set of values to a corresponding expected set of values; and
automatically executing the reconfiguration command set on the structured dataset to modify the one or more observed association rules corresponding to the anomaly set into alignment with the one or more expected association rules.
2. The computer-implemented method of claim 1 , further comprising:
selecting the third AI model set from multiple AI models using a respective set of performance metric values of each of the multiple AI models.
3. The computer-implemented method of claim 1 , further comprising:
receiving an unstructured dataset;
using a fourth AI model set to evaluate the unstructured dataset against a predefined criterion set by:
extracting a set of information from the unstructured dataset, wherein the set of information includes a set of entities, and
evaluate the set of information against a threshold value set of the predefined criterion set;
using the evaluation to identify a portion of the extracted information failing to satisfy the threshold value set; and
generating a set of actions to increase a degree of satisfaction of the extracted information against the predefined criterion set.
4. The computer-implemented method of claim 3 , wherein the reconfiguration commands include one or more of:
instructions to automatically execute a set of workflows for a first type of anomaly, or
triggering one or more alerts for a second type of anomaly.
5. The computer-implemented method of claim 1 , wherein at least one model of the first AI model set, or at least one model of the second AI model set, or at least one model of the third AI model set are the same.
6. The computer-implemented method of claim 3 , wherein the set of information is evaluated against the threshold value set by measuring one or more of:
a degree of completeness of the set of extracted information,
a degree of accuracy of the set of extracted information, or
a degree of alignment of the set of extracted information with specific formats of the predefined criterion set.
7. The computer-implemented method of claim 3 , further comprising:
displaying a compliance report indicating (i) the identified anomaly set, (ii) the set of actions, and (iii) the increased degree of satisfaction of the unstructured dataset against the predefined criterion set.
8. One or more non-transitory, computer-readable storage media comprising instructions thereon, wherein the instructions, when executed by at least one data processor of a system, cause the system to:
receive a structured dataset including an observed set of values for each variable in a variable set;
identify, using a first AI model set, an anomaly set from the observed set of values in the structured dataset by comparing an observed pattern set from the observed set of values against multiple reference patterns;
use a second AI model set to evaluate the identified anomaly set by:
dynamically generating an observed association rule set configured to cause the second AI model set to generate the received observed set of values in the structured dataset, and
comparing the observed association rule set with an expected association rule set to determine one or more observed association rules corresponding to the anomaly set;
use a third AI model set to generate a reconfiguration command set configured to modify the one or more observed association rules corresponding to the anomaly set into alignment with one or more expected association rules by:
identifying a portion of the observed set of values that corresponds to the one or more observed association rules of the observed association rule set that corresponds to the anomaly set, and
associating the portion of the observed set of values to one or more expected association rules of the expected association rule set, the one or more expected association rules being configured to adjust the portion of the observed set of values to a corresponding portion of the expected set of values; and
execute the reconfiguration command set on the structured dataset to modify the one or more observed association rules corresponding to the anomaly set into alignment with the one or more expected association rules.
9. The one or more non-transitory, computer-readable storage media of claim 8 , wherein the instructions further cause the system to:
select the third AI model set from multiple AI models based on a respective set of performance metric values of each of the multiple AI models.
10. The one or more non-transitory, computer-readable storage media of claim 8 , wherein the instructions further cause the system to:
receive an unstructured dataset;
use a fourth AI model set to evaluate the unstructured dataset against a predefined criterion set by:
extracting a set of information from the unstructured dataset, wherein the set of information includes a set of entities, and
evaluating the set of information against a threshold value set of the predefined criterion set;
use the evaluation to identify a portion of the extracted information failing to satisfy the threshold value set; and
generate a set of actions to increase a degree of satisfaction of the extracted information against the predefined criterion set.
11. The one or more non-transitory, computer-readable storage media of claim 10 , wherein the reconfiguration commands includes instructions to one or more of:
automatically execute one or more workflows for a first type of anomaly, or
trigger one or more alerts for a second type of anomaly.
12. The one or more non-transitory, computer-readable storage media of claim 10 , wherein the unstructured dataset includes one or more of: text documents, emails, chat logs, images, or voice recordings.
13. The one or more non-transitory, computer-readable storage media of claim 10 , wherein the information set is evaluated against the threshold value set by measuring one or more of:
a degree of completeness of the set of extracted information,
a degree of accuracy of the set of extracted information, or
a degree of alignment of the set of extracted information with specific formats of the predefined criterion set.
14. The one or more non-transitory, computer-readable storage media of claim 10 , wherein the instructions further cause the system to:
display a compliance report indicating (i) the identified anomaly set, (ii) the action set, and (iii) the increased degree of satisfaction of the unstructured dataset against the predefined criterion set.
15. A system comprising:
at least one hardware processor; and
at least one non-transitory memory storing instructions, which, when executed by the at least one hardware processor, cause the system to:
receive a structured dataset including an observed set of values for each variable in a variable set;
identify, using a first AI model set, an anomaly set from the observed set of values in the structured dataset by comparing an observed pattern set from the observed set of values against multiple reference patterns that correspond to an expected set of values for the variable set;
use a second AI model set to evaluate the identified anomaly set by:
generating an observed association rule set configured to cause the second AI model set to generate the received observed set of values in the structured dataset, and
comparing the observed association rule set with an expected association rule set to determine one or more observed association rules of the observed association rule set that correspond to the anomaly set;
use a third AI model set to generate a reconfiguration command set configured to modify the one or more observed association rules corresponding to the anomaly set into alignment with one or more expected association rules by:
identifying a portion of the observed set of values that corresponds to the one or more observed association rules that correspond to the anomaly set, and
associating the portion of the observed set of values to one or more expected association rules of the expected association rule set, the one or more expected association rules being configured to adjust the portion of the observed set of values to a corresponding expected set of values; and
execute the reconfiguration command set on the structured dataset to modify the one or more observed association rules corresponding to the anomaly set into alignment with the one or more expected association rules.
16. The system of claim 15 , wherein the system is further caused to:
select the third AI model set from multiple AI models based on a respective set of performance metric values of each of the multiple AI models.
17. The system of claim 15 , wherein the system is further caused to:
receive an unstructured dataset;
use a fourth AI model set to evaluate the unstructured dataset against a predefined criterion set by:
extracting a set of information from the unstructured dataset, wherein the set of information includes a set of entities, and
evaluating the set of information against a threshold value set of the predefined criterion set;
use the evaluation to identify a portion of the extracted set of information failing to satisfy the threshold value set; and
generate a set of actions to increase a degree of satisfaction of the extracted set of information against the predefined criterion set.
18. The system of claim 17 , wherein the reconfiguration command set includes instructions to one or more of:
automatically execute one or more workflows for a first type of anomaly, or
trigger one or more alerts for a second type of anomaly.
19. The system of claim 17 , wherein the set of information is evaluated against the threshold value set by measuring one or more of:
a degree of completeness of the extracted set of information,
a degree of accuracy of the extracted set of information, or
a degree of alignment of the extracted set of information with specific formats of the predefined criterion set.
20. The system of claim 17 , wherein the system is further caused to:
display a compliance report indicating (i) the identified anomaly set, (ii) the set of actions, and (iii) the increased degree of satisfaction of the unstructured dataset against the predefined criterion set.