Expanding semantic classes via user feedback
The present technology extends to methods, systems, and computer program products for expanding semantic classes via user feedback. Aspects of the technology learn how a set of labels can be expanded from user-generated tags. Text labels applied by human reviewers to digital content can be inspected and compared to one another. When a threshold of human-generated text tags contain similar terminology, the set of labels can be expanded to define a representation of the similar terminology. Similar terminology can include terms that originate from the same base term, are synonyms, are more specific terms related to a general term category, etc. Similar terminology can be consolidated into a defining term that is used to generate a new (more granular) label or a new top level label. Accordingly, new semantic classes can be discovered from user-generated feedback. New semantic classes can provide a more granular representation of content item classification.
1. A computer-implemented method comprising:
detecting that a threshold number of content items in a corpus of content items include related terminology, the corpus of content items being associated with initial semantic classes that classify the corpus of content items;
based on the detection, generating, via a semantics class computer module, a consolidated term for a tag of a new semantic class based on the related terminology, the new semantic class being in addition to the initial semantic classes classifying the corpus of content items; and
expanding the initial semantic classes into an expanded set of semantic classes by adding the tag corresponding to the new semantic class.
2. The computer-implemented method of claim 1 , further comprising:
receiving a classification of a number of the content items into the new semantic class.
3. The computer-implemented method of claim 1 , wherein generating the new semantic class comprises generating a new semantic sub-class that defines a subset of terminology refining an existing semantic class.
4. The computer-implemented method of claim 3 , wherein the expanding the initial semantic classes further comprises:
resolving a database including the tags into a supplemental semantic class comprising the initial semantic classes coupled with a set of labels including the consolidated term as a new label, wherein the new label is associated with the new semantic class.
5. The computer-implemented method of claim 4 , wherein the resolving the database expands the initial semantic classes by adding the new label that serves as a subset classification of the initial semantic classes.
6. The computer-implemented method of claim 4 , further comprising:
in addition to the new label, adding a catch-all label to the set of labels associated with other tags that do not contain the related terminology.
7. The computer-implemented method of claim 4 , wherein the new semantic class classifies a different subset from other subsets of the initial semantic classes.
8. The computer-implemented method of claim 4 , further comprising:
detecting that another threshold number of second content items of the corpus of content items include a second set of tags containing another set of related terminology;
generating a second consolidated term based on the second set of tags;
resolving the database including the second set of tags into a revised supplemental semantic class comprising the initial semantic classes coupled with a revised set of labels including the second consolidated term as a second new label; and
receive a classification of a number of the content items to be classified into the revised supplemental semantic class.
9. A system comprising:
storage configured to store instructions; and
one or more processors configured to execute the instructions and cause the one or more processors to:
detect that a threshold number of content items in a corpus of content items include related terminology, the corpus of content items being associated with initial semantic classes that classify the corpus of content items;
generate, via a semantics class computer module, a consolidated term for a tag of a new semantic class based on the related terminology, the new semantic class being in addition to the initial semantic classes classifying the corpus of content items; and
expand the initial semantic classes into an expanded set of semantic classes by adding the tag corresponding to the new semantic class.
10. The system of claim 9 , wherein the one or more processors is configured to execute the instructions and cause the one or more processors to:
receive a classification of a number of the content items into the new semantic class.
11. The system of claim 9 , wherein generating the new semantic class comprises generating a new semantic sub-class that defines a subset of terminology refining an existing semantic class.
12. The system of claim 11 , wherein the one or more processors is configured to execute the instructions and cause the one or more processors to:
resolve a database including the tags into a supplemental semantic class comprising the initial semantic classes coupled with a set of labels including the consolidated term as a new label, wherein the new label is associated with the new semantic class.
13. The system of claim 12 , wherein the resolving the database expands the initial semantic classes by adding the new label that serves as a subset classification of the initial semantic classes.
14. The system of claim 12 , wherein the one or more processors is configured to execute the instructions and cause the one or more processors to:
in addition to the new label, add a catch-all label to the set of labels associated with other tags that do not contain the related terminology.
15. The system of claim 12 , wherein the new semantic class classifies a different subset from other subsets of the initial semantic classes.
16. The system of claim 12 , wherein the one or more processors is configured to execute the instructions and cause the one or more processors to:
detect that another threshold number of second content items of the corpus of content items include a second set of tags containing another set of related terminology;
generate a second consolidated term based on the second set of tags;
resolve the database including the second set of tags into a revised supplemental semantic class comprising the initial semantic classes coupled with a revised set of labels including the second consolidated term as a second new label; and
receive a classification of a number of the content items to be classified into the revised supplemental semantic class.
17. A non-transitory computer-readable medium comprising instructions, the instructions, when executed by a computing system, cause the computing system to:
detect that a threshold number of content items in a corpus of content items include related terminology, the corpus of content items being associated with initial semantic classes that classify the corpus of content items;
generate, via a semantics class computer module, a consolidated term for a tag of a new semantic class based on the related terminology, the new semantic class being in addition to the initial semantic classes classifying the corpus of content items; and
expanding the initial semantic classes into an expanded set of semantic classes by adding the tag corresponding to the new semantic class.
18. The non-transitory computer-readable medium of claim 17 , wherein the non-transitory computer-readable medium further comprises instructions that, when executed by the computing system, cause the computing system to:
receive a classification of a number of the content items into the new semantic class.
19. The non-transitory computer-readable medium of claim 17 , generating the new semantic class comprises generating a new semantic sub-class that defines a subset of terminology refining an existing semantic class.
20. The non-transitory computer-readable medium of claim 19 , wherein the non-transitory computer-readable medium further comprises instructions that, when executed by the computing system, cause the computing system to:
resolve a database including the tags into a supplemental semantic class comprising the initial semantic classes coupled with a set of labels including the consolidated term as a new label, wherein the new label is associated with the new semantic class.