Identifying multimedia asset similarity using blended semantic and latent feature
Methods and system for determining a similarity relationship between a plurality of digital assets and a target digital asset comprises creating a normalized semantic feature vector associated with a search query, discovering the target asset based on the normalized semantic feature vector, generating a normalized latent feature vector associated with the target asset, comparing the normalized semantic feature vector with semantic feature vectors for each of the digital assets to generate a semantic comparison value, comparing the normalized target latent feature vector with latent feature vectors for each of the digital assets to generate a latent comparison value, blending the semantic comparison vector value with the latent feature comparison vector value to create a target comparison value for each of the digital assets, and reporting the digital assets having the highest target comparison values to the user or group of users.
1. A method comprising:
receiving a search query comprising one or more terms;
generating, at a computing device, a plurality of target features based on the search query, the plurality of target features comprising:
target semantic features based on the one or more terms; and
target latent features based on user information associated with the one or more terms;
executing, based on the plurality of target features, the search query for a plurality of digital assets stored at a database;
retrieving, from the database based on the executed search query, identifiers of the plurality of digital assets;
for each digital asset of the plurality of digital assets:
generating, at the computing device, a semantic feature score based on a respective digital asset and the target semantic features;
generating, at the computing device, a latent feature score based on the respective digital asset and the target latent features;
generating, at the computing device, a first factor proportional to an amount of user information collected for the respective digital asset;
generating, at the computing device, a second factor that is modified as a function of the first factor, wherein an increase in the first factor corresponds to a decrease in the second factor, wherein an increase in the second factor corresponds to a decrease in the first factor, and wherein the first factor and the second factor add to one hundred percent; and
generating, at the computing device, a comparison score for the respective digital asset, wherein generating the comparison score comprises blending the latent feature score weighted by the first factor and the semantic feature score weighted by the second factor; and
causing to be displayed, at a display associated with the computing device, an ordered list of identifiers of the plurality of digital assets ordered at least partially based on the comparison scores of the plurality of digital assets.
2. The method of claim 1 , further comprising normalizing the comparison scores for the plurality of digital assets, and wherein the ordered list is ordered at least partially based on the normalized comparison scores of the plurality of digital assets.
3. The method of claim 1 , wherein the amount of user information comprises a number of user interactions and user activities collected for the respective digital asset of the plurality of digital assets.
4. The method of claim 1 , further comprising:
in response to determining that the amount of user information collected for the respective digital asset of the plurality of digital assets has increased:
increasing the first factor based on the increased amount of user information; and
wherein the second factor is reduced based on the increase of the first factor.
5. The method of claim 1 , wherein generating the semantic feature score comprises comparing the target semantic features to semantic features based on one or more terms of metadata describing contents of the respective digital asset of the plurality of digital assets.
6. The method of claim 1 , wherein generating the latent feature score comprises comparing the target latent features to latent features based on the user information collected for the respective digital asset of the plurality of digital assets.
7. The method of claim 1 , wherein the user information comprises user information associated with the one or more terms of the search query.
8. The method of claim 1 , wherein the plurality of target features comprises one or more of: title, creation date, director, producer, writer, production studio, actors, characters, dialog, subject matter, genre, objects, settings, locations, themes, or legal clearance to third party copyrighted material associated with the search query.
9. The method of claim 1 , further comprising, for the one or more terms, performing one or more of: tokenizing, stemming, determining synonyms, lower-casing, spell correcting, generating a searchable index, or generating a searchable inverted index.
10. The method of claim 1 , wherein the user information comprises one or more of: user ratings, user feedback, user recommendations, and user reviews.
11. A system comprising:
display circuitry configured to display one or more digital assets; and
a computing device comprising control circuitry configured to:
receive a search query comprising one or more terms;
generate, at the computing device, a plurality of target features based on the search query, the plurality of target features comprising:
target semantic features based on the one or more terms; and
target latent features based on user information associated with the one or more terms;
execute, based on the plurality of target features, the search query for a plurality of digital assets stored at a database;
retrieve, from the database based on the executed search query, identifiers of the plurality of digital assets;
for each digital asset of the plurality of digital assets:
generate, at the computing device, a semantic feature score based on a respective digital asset and the target semantic features;
generate, at the computing device, a latent feature score based on the respective digital asset and the target latent features;
generate, at the computing device, a first factor proportional to an amount of user information collected for the respective digital asset;
generate, at the computing device, a second factor that is modified as a function of the first factor, wherein an increase in the first factor corresponds to a decrease in the second factor, wherein an increase in the second factor corresponds to a decrease in the first factor, and wherein the first factor and the second factor add to one hundred percent; and
generate, at the computing device, a comparison score for the respective digital asset, wherein the control circuitry is configured to blend the latent feature score weighted by the first factor and the semantic feature score weighted by the second factor; and
cause to be displayed, via the display circuitry, an ordered list of identifiers of the plurality of digital assets ordered at least partially based on the comparison scores of the plurality of digital assets.
12. The system of claim 11 , wherein the control circuitry is further configured to normalize the comparison scores for the plurality of digital assets, and wherein the ordered list is ordered at least partially based on the normalized comparison scores of the plurality of digital assets.
13. The system of claim 11 , wherein the amount of user information comprises a number of user interactions and user activities collected for the respective digital asset of the plurality of digital assets.
14. The system of claim 11 , wherein the control circuitry is further configured to:
in response to determining that the amount of user information collected for the respective digital asset of the plurality of digital assets has increased:
increase the first factor based on the increased amount of user information; and
wherein the second factor is reduced based on the increase of the first factor.
15. The system of claim 11 , wherein the control circuitry, when generating the semantic feature score, is configured to compare the target semantic features to semantic features based on one or more terms of metadata describing contents of the respective digital asset of the plurality of digital assets.
16. The system of claim 11 , wherein the control circuitry, when generating the latent feature score, is configured to compare the target latent features to latent features based on the user information collected for the respective digital asset of the plurality of digital assets.
17. The system of claim 11 , wherein the user information comprises user information associated with the one or more terms of the search query.
18. The system of claim 11 , wherein the plurality of target features comprises one or more of: title, creation date, director, producer, writer, production studio, actors, characters, dialog, subject matter, genre, objects, settings, locations, themes, or legal clearance to third party copyrighted material associated with the search query.
19. The system of claim 11 , wherein the control circuitry is further configured to, for the one or more terms, perform one or more of: tokenizing, stemming, determining synonyms, lower-casing, spell correcting, generating a searchable index, or generating a searchable inverted index.
20. The system of claim 11 , wherein the user information comprises one or more of: user ratings, user feedback, user recommendations, and user reviews.