Combined wide and deep machine learning models for automated database element processing systems, methods and apparatuses
A method of automated database element processing includes training a wide machine learning model with historical feature vector inputs to generate a wide ranked element output. The method includes training a deep machine learning model with the historical feature vector inputs to generate a deep ranked element output. The method includes generating a set of inputs specific to an individual entity, obtaining a set of current article database elements, and creating a feature vector input according to the set of inputs and the set of current article database elements. The method includes processing the feature vector input with the wide machine learning model to generate a wide ranked element list, processing the feature vector input with the deep machine learning model to generate a deep ranked element list, and merging database elements of the wide and deep ranked element lists to generate a ranked element recommendation output.
1. A computer-based roadblock article recommendation system comprising:
at least one computer readable non-transitory memory storing computer-executable instructions; and
at least one processor coupled with the at least one computer readable non-transitory memory and that executes the following operations upon execution of the computer-executable instructions:
generating a set of inputs specific to a subscriber, the set of inputs derived from article access records associated with the subscriber;
obtaining a set of candidate roadblock articles;
creating a feature vector input according to the set of inputs specific to the subscriber and the set of candidate roadblock articles;
processing the feature vector input using at least one trained machine learning model to generate a wide and deep merged ranked list of roadblock articles;
selecting a roadblock article from the wide and deep merged ranked list of roadblock articles;
generating a roadblock link for the selected roadblock article; and
inserting the roadblock link within a target article.
2. The computer-based roadblock article recommendation system of claim 1 , wherein the at least one trained machine learning model comprises a wide machine learning model or a deep machine learning model.
3. The computer-based roadblock article recommendation system of claim 2 , wherein the wide machine learning model comprises a term frequency-inverse document frequency (TFIDF) model or a best matching (BM25) model.
4. The computer-based roadblock article recommendation system of claim 2 , wherein the deep machine learning model comprises a Doc2Vec model or a deep contrastive learning (DECLUTR) model.
5. The computer-based roadblock article recommendation system of claim 1 , wherein the operations further include determining a location within the target article to insert the roadblock link.
6. The computer-based roadblock article recommendation system of claim 5 , wherein determining the location is based on a ranking of areas where a reader is more likely to look while reading the target article.
7. The computer-based roadblock article recommendation system of claim 1 , wherein the operations further include automatically generating content associated with the selected roadblock article or the roadblock link.
8. The computer-based roadblock article recommendation system of claim 7 , wherein automatically generating the content comprises using a generative pre-trained transformer (GPT) model.
9. The computer-based roadblock article recommendation system of claim 1 , wherein the operations further including automatically generating a description for the roadblock link.
10. The computer-based roadblock article recommendation system of claim 1 , wherein the set of inputs specific to the subscriber includes at least one of a geolocation, a device type, a sex, an age, a time, or a day.
11. The computer-based roadblock article recommendation system of claim 1 , wherein the article access records include at least one of a number of clicks on each article or an amount of time spent viewing each article.
12. The computer-based roadblock article recommendation system of claim 1 , wherein the operations further including retraining the at least one trained machine learning model periodically.
13. The computer-based roadblock article recommendation system of claim 12 , wherein the at least one trained machine learning model is retrained at least hourly.
14. The computer-based roadblock article recommendation system of claim 1 , wherein the set of candidate roadblock articles is obtained from articles published within a specified time period.
15. The computer-based roadblock article recommendation system of claim 1 , wherein the operations further includes inserting the roadblock link based on a template.
16. The computer-based roadblock article recommendation system of claim 1 , wherein the operations further includes generating multiple roadblock links for insertion within the target article.
17. The computer-based roadblock article recommendation system of claim 1 , wherein the roadblock link comprises a promotional article link or an advertisement.
18. The computer-based roadblock article recommendation system of claim 1 , wherein the at least one trained machine learning model is trained on historical article items.
19. The computer-based roadblock article recommendation system of claim 18 , wherein the historical article items are historical article items related to the subscriber.
20. The computer-based roadblock article recommendation system of claim 1 , wherein the operations further include generating roadblock link recommendations that include content opposing what the subscriber normally views thereby combating self-imposed filter bubbles.
21. A computer-based method comprising:
generating a set of inputs specific to a subscriber, the set of inputs derived from article access records associated with the subscriber;
obtaining a set of candidate roadblock articles;
creating a feature vector input according to the set of inputs specific to the subscriber and the set of candidate roadblock articles;
processing the feature vector input using at least one trained machine learning model to generate a wide and deep merged ranked list of roadblock articles;
selecting a roadblock article from the wide and deep merged ranked list of roadblock articles;
generating a roadblock link for the selected roadblock article; and
inserting the roadblock link within a target article.
22. A non-transitory computer-readable media comprising instructions which, when executed by at least one processor, cause the at least one processor to execute the following operations:
generating a set of inputs specific to a subscriber, the set of inputs derived from article access records associated with the subscriber;
obtaining a set of candidate roadblock articles;
creating a feature vector input according to the set of inputs specific to the subscriber and the set of candidate roadblock articles;
processing the feature vector input using at least one trained machine learning model to generate a wide and deep merged ranked list of roadblock articles;
selecting a roadblock article from the wide and deep merged ranked list of roadblock articles;
generating a roadblock link for the selected roadblock article; and
inserting the roadblock link within a target article.