Personalized recency and time-aware system
Systems, methods, and computer-storage media provide content item retrieval that is personalized and time aware for a user. Interaction tracking is performed to identify content items with which the user has interacted and time data indicative of when the user interacted with each content item. A machine learning model is trained using the content items and the time data associated with each content item. Once trained, the machine learning model is used to generate a relevance score for each of a number of target content items. At least a portion of the target content items are provided for presentation to the user based on the relevance scores.
1. A computer-implemented method comprising:
identifying a plurality of content items with which a user has interacted and a time stamp for each content item indicative of when the user interacted with the content item;
training a machine learning model using the content items and the time stamp for each content item to generate a trained machine learning model;
generating, using the trained machine learning model, a relevance score for each target content item from a plurality of target content items, the relevance score for each target content item indicative of a relevance of the target content item to the user; and
providing, for presentation on a user device of the user, at least a portion of the target content items based on the relevance scores.
2. The computer-implemented method of claim 1 , wherein the plurality of content items are identified based at least in part on a category associated with each of the content items.
3. The computer-implemented method of claim 1 , wherein training the machine learning model using the content items and the time stamp for each content item comprises:
generating a content item vector representation for each content item; and
training the machine learning model using the content item vector representations.
4. The computer-implemented method of claim 3 , wherein the content item vector representation for a first content item is generated using text of the first content item.
5. The computer-implemented method of claim 3 , wherein training the machine learning model using the content items and the time stamp for each content item further comprises:
generating a time vector representation for the time stamp for each content item; and
training the machine learning model using the time vector representations.
6. The computer-implemented method of claim 1 , wherein generating the relevance score for a first target content item from the plurality of target content items comprises:
generating a vector representation of the first target content item; and
using the trained machine learning model to predict the relevance score for the first target content item using the vector representation of the first target content item.
7. The computer-implemented method of claim 6 , wherein the trained machine learning model predicts the relevance score for the first target content item based on time data associated with the first target content item.
8. One or more computer storage media storing computer-useable instructions that, when used by a computing device, cause the computing device to perform operations, the operations comprising:
identifying a plurality of content items with which a user has interacted and a time stamp for each content item indicative of when the user interacted with the content item;
training a machine learning model using the content items and the time stamp for each content item to generate a trained machine learning model;
generating, using the trained machine learning model, a relevance score for each target content item from a plurality of target content items, the relevance score for each target content item indicative of a relevance of the target content item to the user; and
providing, for presentation on a user device of the user, at least a portion of the target content items based on the relevance scores.
9. The one or more computer storage media of claim 8 , wherein the plurality of content items are identified based at least in part on a category associated with each of the content items.
10. The one or more computer storage media of claim 8 , wherein training the machine learning model using the content items and the time stamp for each content item comprises:
generating a content item vector representation for each content item; and
training the machine learning model using the content item vector representations.
11. The one or more computer storage media of claim 10 , wherein the content item vector representation for a first content item is generated using text of the first content item.
12. The one or more computer storage media of claim 10 , wherein training the machine learning model using the content items and the time stamp for each content item further comprises:
generating a time vector representation for the time stamp for each content item; and
training the machine learning model using the time vector representations.
13. The one or more computer storage media of claim 8 , wherein generating the relevance score for a first target content item from the plurality of target content items comprises:
generating a vector representation of the first target content item; and
using the trained machine learning model to predict the relevance score for the first target content item using the vector representation of the first target content item.
14. The one or more computer storage media of claim 13 , wherein the trained machine learning model predicts the relevance score for the first target content item based on time data associated with the first target content item.
15. A computer system comprising:
a processor; and
a computer storage medium storing computer-useable instructions that, when used by the processor, causes the computer system to perform operations comprising:
identifying a plurality of content items with which a user has interacted and a time stamp for each content item indicative of when the user interacted with the content item;
training a machine learning model using the content items and the time stamp for each content item to generate a trained machine learning model;
generating, using the trained machine learning model, a relevance score for each target content item from a plurality of target content items, the relevance score for each target content item indicative of a relevance of the target content item to the user; and
providing, for presentation on a user device of the user, at least a portion of the target content items based on the relevance scores.
16. The computer system of claim 15 , wherein the plurality of content items are identified based at least in part on a category associated with each of the content items.
17. The computer system of claim 15 , wherein training the machine learning model using the content items and the time stamp for each content item comprises:
generating a content item vector representation for each content item; and
training the machine learning model using the content item vector representations.
18. The computer system of claim 17 , wherein the content item vector representation for a first content item is generated using text of the first content item.
19. The computer system of claim 17 , wherein training the machine learning model using the content items and the time stamp for each content item further comprises:
generating a time vector representation for the time stamp for each content item; and
training the machine learning model using the time vector representations.
20. The computer system of claim 15 , wherein generating the relevance score for a first target content item from the plurality of target content items comprises:
generating a vector representation of the first target content item; and
using the trained machine learning model to predict the relevance score for the first target content item using the vector representation of the first target content item.