Systems and methods for automated content curation using signature analysis
Systems and methods are described herein for curating content that follows a narrative structure. A narrative structure comprises narrative portions that have a defined order. Signature analysis of known content that follows the narrative structure is used to train machine learning models for the narrative structure and the narrative portions that make up the narrative structure. Signature analysis of candidate content segments, along with machine learning models for the narrative portions, are used to identify candidate content segments that match the respective narrative portions. A candidate playlist is generated of the identified candidate content segments in the defined order. In one embodiment, the machine learning model for the narrative structure is used to validate the generated playlist.
1. A method for curating content, the method comprising:
retrieving a plurality of content items, wherein each of the plurality of content items comprises a plurality of content segments that follow a narrative structure, wherein the narrative structure comprises a plurality of portions having a temporal order, and wherein the plurality of content segments for each of the plurality of content items, correspond to the plurality of portions;
generating, for each portion of the narrative structure, segment signature vectors based on the content segments of the plurality of content items corresponding to the portion;
training, for each portion of the narrative structure, a segment machine learning model based on the plurality of segment signature vectors corresponding to the portion to generate a plurality of segment machine learning models, wherein the plurality of segment machine learning models are usable to identify candidate content segments for generating a candidate that follows the narrative structure;
generating, for each content item of the plurality of content items, a content signature vector based on the content item to generate a plurality of content signature vectors;
training a content machine learning model based on the plurality of content signature vectors, wherein the content machine learning model is usable to validate the candidate playlist of identified candidate content segments; and
outputting the plurality of segment machine learning models and the content machine learning model for generating one or more playlists.
2. The method of claim 1 wherein outputting the plurality of segment machine learning models comprises making the plurality of segment machine learning models and the content machine learning model available for generating one or more playlists from a plurality of candidate segments.
3. The method of claim 1 further comprising validating each of the segment machine learning models using test content segments.
4. The method of claim 3 wherein the validating, for each of the segment machine learning models, comprises analyzing the test content segments using the segment machine learning model to identify a match.
5. The method of claim 1 further comprising validating the content machine learning model using a plurality of test content items.
6. The method of claim 5 wherein the validating comprises analyzing the test content items using the content machine learning model to identify a match.
7. The method of claim 1 wherein the segment machine learning model comprises at least one neural network or a Bayesian network.
8. The method of claim 1 wherein the content machine learning model comprises at least one neural network or a Bayesian network.
9. The method of claim 1 wherein generating, for each content item of the plurality of content items, the content signature vector comprises applying mathematical operations across all of the content item.
10. The method of claim 1 wherein the segment signature vectors are generated based on characteristics associated with one of audio of the content, video frames of the content or combinations thereof.
11. A system comprising:
memory; and
control circuitry configured to:
retrieve a plurality of content items, wherein each of the plurality of content items comprises a plurality of content segments that follow a narrative structure, wherein the narrative structure comprises a plurality of portions having a temporal order, and wherein the plurality of content segments for each of the plurality of content items, correspond to the plurality of portions;
generate, for each portion of the narrative structure, segment signature vectors based on the content segments of the plurality of content items corresponding;
train, for each portion of the narrative structure, a segment machine learning model based on the plurality of segment signature vectors corresponding to the portion to generate a plurality of segment machine learning models, wherein the plurality of segment machine learning models are usable to identify candidate content segments for generating a candidate playlist that follows the narrative structure;
generate, for each content item of the plurality of content items, a content signature vector based on the content item to generate a plurality of content signature vectors;
train a content machine learning model based on the plurality of content signature vectors, wherein the content machine learning model is usable to validate the candidate playlist of identified candidate content segments; and
output the plurality of segment machine learning models and the content machine learning model for generating one or more playlists.
12. The system of claim 11 wherein to output the plurality of segment machine learning models, the control circuitry is configured to:
make the plurality of segment machine learning models and the content machine learning model available for generating one or more playlists from a plurality of candidate segments.
13. The system of claim 11 wherein the control circuitry is configured to:
validate each of the segment machine learning models using test content segments.
14. The system of claim 13 wherein to validate each of the segment machine learning models, the control circuitry is configured to:
analyze the test content segments using the segment machine learning model to identify a match.
15. The system of claim 11 wherein the control circuitry is configured to:
validate the content machine learning model using a plurality of test content items.
16. The system of claim 15 wherein to validate the content machine model, the control circuitry is configured to:
analyze the test content items using the content machine learning model to identify a match.
17. The system of claim 11 wherein the segment machine learning model comprises at least one neural network or a Bayesian network.
18. The system of claim 11 wherein the content machine learning model comprises at least one neural network or a Bayesian network.
19. The system of claim 11 wherein to generate, for each content item of the plurality of content items, the content signature vector, the control circuitry is configured to:
apply mathematical operations across all of the content item.
20. The system of claim 11 wherein the segment signature vectors are generated based on characteristics associated with one of audio of the content, video frames of the content or combinations thereof.