System and method for generating descriptive metadata
A computer-implemented method and system for generating a synopsis for a content item comprising: obtaining user data for at least one user and/or associated metadata, wherein the user data and/or associated metadata is based on user activity for the at least one user; receiving content metadata and/or other content information for the content item; generating or otherwise obtaining a synopsis for the content item based on the obtained content information and based on at least some of said obtained user data.
1 . A computer-implemented method for generating a synopsis for a content item comprising:
obtaining user data for at least one user, wherein the user data is based on user activity for the at least one user and represents user engagement with content available through a content distribution system;
receiving a request for one or more content recommendations;
generating a content recommendation for a set of content items based on at least content information for the set of content items and the obtained user data, wherein the set of content items comprises content recommendation candidates;
receiving at least one of content metadata and other content information for a content item of the set of content items;
providing a request for a synopsis to a generative text or language model based on the received at least one of content metadata and other content information and the obtained user data;
transmitting the request to a model server hosting the generative text or language model via a communication interface; and
generating by the generative text or language model on the model server a synopsis text for the content item based on the request.
2 . The method of claim 1 , wherein the generated synopsis comprises a customized synopsis for said at least one user.
3 . The method of claim 1 , wherein the content metadata comprises enriched content metadata.
4 . The method of claim 1 further comprises:
receiving a call request, for example, in accordance with application programming interface protocol for a synopsis based on the content metadata and user data;
transmitting the API call to a server or further networked processor;
generating the synopsis based on the API call.
5 . The method of claim 1 , wherein the content information comprises an initial synopsis for the content item and wherein generating the synopsis comprises at least one of modifying and re-generating the initial synopsis based on at least one of the obtained user data and the received content metadata.
6 . The method of claim 1 , wherein the generating of the synopsis text is performed for each content item of the set of content items.
7 . The method of claim 1 , wherein the synopsis is generated and stored at a first time during or in response to a content recommendation request and retrieved from storage at a second time in response to a user interaction with a content selection interface or other user interface.
8 . The method of claim 1 wherein generating the synopsis comprises applying a machine learning derived model to at least part of the content metadata, and at least part of the user data wherein the machine learning model is configured to output the synopsis.
9 . The method of claim 8 , wherein the machine learning derived model is applied to an initial synopsis.
10 . The method of claim 8 , further comprising selecting one or more parameters for the model, wherein the one or more parameters comprises at least one of: requested language, length of desired synopsis.
11 . The method of claim 1 further comprising obtaining user device information information relating to the size of a display or a display window, and generating the synopsis based on said user device information.
12 . The method of claim 1 wherein the at least one of user data and associated metadata is represented as a feature vector or other data structure and wherein the method comprises generating a prompt or other input for a model based on said feature vector or other data structure and said content metadata.
13 . The method of claim 12 , wherein the prompt is based on an initial synopsis.
14 . The method of claim 1 wherein generating the synopsis comprises packaging at least part of said at least one user data and associated content metadata, said content metadata and one or more selected parameters for a synopsis generator.
15 . The method of claim 1 wherein the method comprises performing at least one of a filtering and selection process on the at least one of user data and associated metadata and wherein the generating of the synopsis is based on the at least one of filtered and selected at least one of user data and associated metadata.
16 . The method of claim 1 wherein at least one of a), b): a) the content metadata is obtained from a first data source without the use of at least one of user data and associated metadata; b) wherein the synopsis is generated based on an initial synopsis that is independent of user activity.
17 . The method of claim 1 wherein generating or otherwise obtaining the synopsis comprises evaluating a set of pre-generated synopses for a suitable or closest match based on said at least one of user data and associated metadata.
18 . The method of claim 17 , wherein the generating or obtaining the synopsis comprises:
at least one of a), b): a) retrieving said suitable or closest match based on said evaluation; b) generating a synopsis in the absence of a suitable or closest match.
19 . The method of claim 1 , wherein the method comprises at least one of a), b):
a) identifying a group of users for at least one user from a plurality of groups based on said user data and wherein the generation of the synopsis comprises is based on said identification of group;
b) obtaining first party metadata from a first source and obtaining at least one of user data and associated content metadata from a second source.
20 . The method of claim 1 , wherein the method further comprises displaying the generated synopsis as part of a content selection interface or other user interface.
21 . A system comprising processing circuitry configured to:
obtain user data for at least one user, wherein the user data is based on user activity for the at least one user and represents user engagement with content available through a content distribution system;
receive a request for one or more content recommendations;
generate a content recommendation for a set of content items based on at least content information for the set of content items and the obtained user data, wherein the set of content items comprises content recommendation candidates;
receive at least one of content metadata and other content information for a content item of the set of content items;
provide a request for a synopsis to a generative text or language model based on the received at least one of content metadata and other content information and the obtained user data;
transmit the request to a model server hosting the generative text or language model via a communication interface; and
generate by the generative text or language model on the model server a synopsis text for the content item based on the request.
22 . A non-transitory computer-readable medium that comprises computer-readable instructions that are executable to:
obtain user data for at least one user, wherein the user data is based on user activity for the at least one user and represents user engagement with content available through a content distribution system;
receive a request for one or more content recommendations;
generate a content recommendation for a set of content items based on at least content information for the set of content items and the obtained user data, wherein the set of content items comprises content recommendation candidates;
receive at least one of content metadata and other content information for a content item of the set of content items;
provide a request for a synopsis to a generative text or language model based on the received at least one of content metadata and other content information and the obtained user data;
transmit the request to a model server hosting the generative text or language model via a communication interface; and
generate by the generative text or language model on the model server a synopsis text for the content item based on the request.