Metadata aggregation using a trained entity matching predictive model

1. A metadata aggregation system comprising:

a computing platform including a hardware processor and a system memory;

a software code stored in the system memory, the software code including a trained entity matching predictive model trained using training data obtained from a reference database;

the hardware processor configured to execute the software code to:

obtain a plurality of metadata inputs from a plurality of sources;

conform the plurality of metadata inputs to a common format;

match, using the trained entity matching predictive model, at least some of the conformed plurality of metadata inputs to a same entity to generate a plurality of matches;

determine, using the trained entity matching predictive model, a confidence score for each of the plurality of matches;

send a confirmation request to at least one human editor for confirmation of each of the plurality of matches having a respective confidence score greater than a first threshold score and less than a second threshold score; and

update the reference database, in response to receiving a confirmation of at least one of the plurality of matches as a confirmed match from the at least one human editor, to include the confirmed match in the reference database.

2. The metadata aggregation system of claim 1 , wherein the hardware processor is further configured to execute the software code to:

obtain an updated training data from the updated reference database; and

retrain, using the updated training data, the trained entity matching predictive model.

3. The metadata aggregation system of claim 2 , wherein the hardware processor is further configured to execute the software code to:

receive a plurality of entity queries from at least one client of the metadata aggregation system;

identify, using the trained entity matching predictive model, a plurality of entity matching failures based on null-results in response to at least some of the plurality of entity queries; and

prioritize the retraining of the trained entity matching predictive model to reduce the plurality of entity matching failures.

4. The metadata aggregation system of claim 1 , wherein the hardware processor is further configured to execute the software code to:

perform an automated identification of each of the plurality of matches having a respective confidence score greater than or equal to the second threshold score as another confirmed match; and

further update the updated reference database to include the another confirmed match in the updated reference database.

5. The metadata aggregation system of claim 4 , wherein the hardware processor is further configured to execute the software code to:

obtain a further updated training data from the further updated training database; and

retrain, using the further updated training data, the trained entity matching predictive model.

6. The metadata aggregation system of claim 5 , wherein the hardware processor is further configured to execute the software code to:

receive a plurality of entity queries from at least one client of the metadata aggregation system;

identify, using the trained entity matching predictive model, a plurality of entity matching failures based on null-results in response to at least some of the plurality of entity queries; and

prioritize the retraining of the trained entity matching predictive model to reduce the plurality of entity matching failures.

7. The metadata aggregation system of claim 1 , wherein the hardware processor is further configured to execute the software code to:

perform an automated identification of each of the plurality of matches having a respective confidence score less than or equal to the first threshold score as a mismatch; and

prevent the reference database from being updated to include the mismatch.

8. The metadata aggregation system of claim 1 , wherein at least one of the plurality of sources of the plurality of metadata inputs comprises at least one of a knowledge base or a news source.

9. The metadata aggregation system of claim 1 , wherein at least one of the plurality of sources of the plurality of metadata inputs comprises a social media platform.

10. The metadata aggregation system of claim 1 , wherein the same entity is one of an athlete, a sports team, a sports league, a movie, a movie franchise, a video game, a linear television program or series, a web television program or series, an actor, or a fictional character.

11. A method for use by a metadata aggregation system including a computing platform having a hardware processor and a system memory storing a software code, the software code including a trained entity matching predictive model trained using training data obtained from a reference database, the method comprising:

obtaining, by the software code executed by the hardware processor, a plurality of metadata inputs from a plurality of sources;

conforming, by the software code executed by the hardware processor, the plurality of metadata inputs to a common format;

matching, by the software code executed by the hardware processor and using the trained entity matching predictive model, at least some of the conformed plurality of metadata inputs to a same entity to generate a plurality of matches;

determining, by the software code executed by the hardware processor and using the trained entity matching predictive model, a confidence score for each of the plurality of matches;

sending, by the software code executed by the hardware processor, a confirmation request to at least one human editor for confirmation of each of the plurality of matches having a respective confidence score greater than a first threshold score and less than a second threshold score; and

updating the reference database, by the software code executed by the hardware processor in response to receiving a confirmation of at least one of the plurality of matches as a confirmed match from the at least one human editor, to include the confirmed match in the reference database.

12. The method of claim 11 , further comprising:

obtaining, by the software code executed by the hardware processor, an updated training data from the updated reference database; and

retraining, by the software code executed by the hardware processor and using the updated training data, the trained entity matching predictive model.

13. The method of claim 12 , further comprising:

receiving, by the software code executed by the hardware processor, a plurality of entity queries from at least one client of the metadata aggregation system;

identifying, by the software code executed by the hardware processor and using the trained entity matching predictive model, a plurality of entity matching failures based on null-results in response to at least some of the plurality of entity queries; and

prioritizing, by the software code executed by the hardware processor, the retraining of the trained entity matching predictive model to reduce the plurality of entity matching failures.

14. The method of claim 11 , further comprising:

performing, by the software code executed by the hardware processor, an automated identification of each of the plurality of matches having a respective confidence score greater than or equal to the second threshold score as another confirmed match; and

further updating, by the software code executed by the hardware processor, the updated reference database to include the another confirmed match in the updated reference database.

15. The method of claim 14 , further comprising:

obtaining, by the software code executed by the hardware processor, a further updated training data from the further updated training database; and

retraining, by the software code executed by the hardware processor and using the further updated training data, the trained entity matching predictive model.

16. The method of claim 15 , further comprising:

receiving, by the software code executed by the hardware processor, a plurality of entity queries from at least one client of the metadata aggregation system;

identifying, by the software code executed by the hardware processor and using the trained entity matching predictive model, a plurality of entity matching failures based on null-results in response to at least some of the plurality of entity queries; and

prioritizing, by the software code executed by the hardware processor, the retraining of the trained entity matching predictive model to reduce the plurality of entity matching failures.

17. The method of claim 11 , further comprising:

performing, by the software code executed by the hardware processor, an automated identification of each of the plurality of matches having a respective confidence score less than or equal to the first threshold score as a mismatch; and

preventing, by the software code executed by the hardware processor, the reference database from being updated to include the mismatch.

18. The method of claim 11 , wherein at least one of the plurality of sources of the plurality of metadata inputs comprises at least one of a knowledge base or a news source.

19. The method of claim 11 , wherein at least one of the plurality of sources of the plurality of metadata inputs comprises a social media platform.

20. The method of claim 11 , wherein the same entity is one of an athlete, a sports team, a sports league, a movie, a movie franchise, a video game, a linear television program or series, a web television program or series, an actor, or a fictional character.