Contextualized Novelty for Personalized Discovery

A method implemented in a data processing system for conducting a contextualized search for relevant and novel content, the method comprising:

• obtaining query parameters for a query for content items in a datastore, the query parameters including attributes of content items for which a search is to be conducted;
• obtaining a first set of content items from a content datastore based on the query parameters;
• analyzing the first set of content items using a first machine learning model trained to generate relevant content information that identifies a plurality of relevant content items included in the first set of content items, the first machine learning model is a graph builder model configured to output a weighted graph in which nodes of the weighted graph represent content items or attributes of content items and edges of the weighted graph represent relationships between the nodes, and identifying the plurality of relevant content items further comprises:
• generating a subgraph comprising a plurality of nodes and edges from the weighted graph by filtering the weighted graph;
• traversing the subgraph to estimate for a respective node a likelihood of observing the respective node given a specified set of nodes of the subgraph;
• ranking the plurality of nodes based the likelihood associated with each of nodes;
• and
• selecting a subset of the nodes of the subgraph based on a rank associated with each of the nodes;
• and
• analyzing the plurality of relevant content items using a second machine learning model configured to output novel content information, the novel content information including a plurality of content items predicted to be relevant and novel, the novel content information ranking the plurality of content items predicted to be relevant and novel based on a novelty score associated with each respective content item.

A machine-readable medium on which are stored instructions that, when executed, cause a processor of a programmable device to perform operations of:

• obtaining query parameters for a query for content items in a datastore, the query parameters including attributes of content items for which a search is to be conducted;
• obtaining a first set of content items from a content datastore based on the query parameters;
• analyzing the first set of content items using a first machine learning model trained to generate relevant content information that identifies a plurality of relevant content items included in the first set of content items, the first machine learning model is a graph builder model configured to output a weighted graph in which nodes of the weighted graph represent content items or attributes of content items and edges of the weighted graph represent relationships between the nodes, and identifying the plurality of relevant content items further comprises:
• generating a subgraph comprising a plurality of nodes and edges from the weighted graph by filtering the weighted graph;
• traversing the subgraph to estimate for a respective node a likelihood of observing the respective node given a specified set of nodes of the subgraph;
• ranking the plurality of nodes based the likelihood associated with each of nodes;
• and
• selecting a subset of the nodes of the subgraph based on a rank associated with each of the nodes;
• and
• analyzing the plurality of relevant content items using a second machine learning model configured to output novel content information, the novel content information including a plurality of content items predicted to be relevant and novel, the novel content information ranking the plurality of content items predicted to be relevant and novel based on a novelty score associated with each respective content item.