Utilizing machine-learning models to generate identifier embeddings and determine digital connections between digital content items
The present disclosure relates to systems, methods, and non-transitory computer-readable media that utilize machine learning models to generate identifier embeddings from digital content identifiers and then leverage these identifier embeddings to determine digital connections between digital content items. In particular, the disclosed systems can utilize an embedding machine-learning model that comprises a character-level embedding machine-learning model and a word-level embedding machine-learning model. For example, the disclosed systems can combine a character embedding from the character-level embedding machine-learning model and a token embedding from the word-level embedding machine-learning model. The disclosed systems can determine digital connections between the plurality of digital content items by processing these identifier embeddings for a plurality of digital content items utilizing a content management model. Based on the digital connections, the disclosed systems can surface one or more digital content suggestions to a user interface of a client device.
1. A system comprising:
at least one processor; and
at least one non-transitory computer-readable storage medium storing instructions that, when executed by the at least one processor, cause the system to:
identify one or more identifiers corresponding to one or more digital content items;
generate one or more identifier embeddings corresponding to the one or more identifiers by utilizing a first machine learning model;
generate file relation predictions between the one or more digital content items by processing the one or more identifier embeddings utilizing a second machine learning model;
determine, for a first digital content item of the one or more digital content items, a storage location within a file structure for a user account based on a file relation prediction associated with the first digital content item; and
store the first digital content item at the storage location within the file structure for the user account.
2. The system of claim 1 , wherein:
generating the file relation predictions between the one or more digital content items comprises generating a sibling relationship prediction between the first digital content item and a second content item from the one or more digital content items; and
determining the storage location within the file structure for the user account is based at least in part on the sibling relationship prediction between the first digital content item and the second content item.
3. The system of claim 1 , wherein:
generating the file relation predictions between the one or more digital content items comprises generating a parent-child relationship prediction between the first digital content item and a second content item from the one or more digital content items; and
determining the storage location within the file structure for the user account is based at least in part on the parent-child relationship prediction between the first digital content item and the second content item.
4. The system of claim 1 , wherein generating the one or more identifier embeddings comprises:
generating one or more character-level embeddings based on individual characters within the one or more identifiers; or
generating one or more word-level embeddings based on groups of characters within the one or more identifiers.
5. The system of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the system to:
determine an access privilege prediction for the first digital content item, the access privilege prediction comprising a defined access level comprising one or more of view, edit, or share; and
assign the defined access level to the first digital content item based on the access privilege prediction.
6. The system of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the system to:
generate one or more user activity embeddings corresponding to the one or more digital content items; and
wherein generating the file relation predictions between the one or more digital content items is further based on the one or more user activity embeddings.
7. The system of claim 1 , further comprising instructions that, when executed by the at least one processor, cause the system to:
determine the storage location is associated with a shared storage location; and
provide, to one or more additional user account, access to the first digital content item based on storing the first digital content item at the storage location.
8. A non-transitory computer readable medium comprising instructions that, when executed by at least one processor, cause a computing device to:
identify one or more identifiers corresponding to one or more digital content items;
generate one or more identifier embeddings corresponding to the one or more identifiers by utilizing a first machine learning model;
generate file relation predictions between the one or more digital content items by processing the one or more identifier embeddings utilizing a second machine learning model;
identify a file relation prediction associated with a first digital content item; and
store the first digital content item at a storage location within a file structure for a user account based on the file relation prediction associated with the first digital content item.
9. The non-transitory computer readable medium as recited in claim 8 , wherein the first digital content item is newly added to the user account.
10. The non-transitory computer readable medium as recited in claim 8 , wherein identifying the file relation prediction associated with the first digital content item comprises identifying a sibling relationship prediction between the first digital content item and a second content item from the one or more digital content items.
11. The non-transitory computer readable medium as recited in claim 8 , wherein identifying the file relation prediction associated with the first digital content item comprises identifying a parent-child relationship prediction between the first digital content item and a second content item from the one or more digital content items.
12. The non-transitory computer readable medium as recited in claim 8 , wherein generating the one or more identifier embeddings comprises:
generating one or more extension-level embeddings based on file extensions within the one or more identifiers;
generating one or more character-level embeddings based on individual characters within the one or more identifiers; or
generating one or more word-level embeddings based on groups of characters within the one or more identifiers.
13. The non-transitory computer readable medium as recited in claim 8 , further comprising instructions that, when executed by the at least one processor, cause the computing device to assign a defined access level to the first digital content item based on determining an access privilege prediction for the first digital content item.
14. The non-transitory computer readable medium as recited in claim 8 , further comprising instructions that, when executed by the at least one processor, cause the computing device to:
generate one or more user activity embeddings corresponding to the one or more digital content items; and
wherein generating the file relation predictions between the one or more digital content items is further based on the one or more user activity embeddings.
15. A computer-implemented method comprising:
identifying one or more identifiers corresponding to one or more digital content items;
generating one or more identifier embeddings corresponding to the one or more identifiers by utilizing a first machine learning model;
determine file relation predictions between the one or more digital content items by processing the one or more identifier embeddings utilizing a second machine learning model; and
storing a first digital content item of the one or more digital content items at a storage location within a file structure based on a file relation prediction between the first digital content item and a second digital content item from the one or more digital content items.
16. The computer-implemented method of claim 15 , wherein the one or more identifiers corresponding to the one or more digital content items are file names of the one or more digital content items.
17. The computer-implemented method of claim 15 , further comprising identifying that the file relation prediction between with the first digital content item and the second digital content item is a sibling relationship prediction between the first digital content item and the second digital content item from the one or more digital content items.
18. The computer-implemented method of claim 15 , further comprising identifying that the file relation prediction between with the first digital content item and the second digital content item is a parent-child relationship prediction between the first digital content item and the second digital content item from the one or more digital content items.
19. The computer-implemented method of claim 15 , further comprising assigning a defined access level to the first digital content item based on determining an access privilege prediction for the first digital content item, wherein the defined access level comprises one or more of view, edit or share.
20. The computer-implemented method of claim 15 , further comprising:
generate one or more user activity embeddings corresponding to the one or more digital content items; and
wherein determining the file relation predictions between the one or more digital content items is further based on the one or more user activity embeddings.