Efficient processing of neighborhood data

A computer-readable medium bearing computer-executable instructions which, when executed on a computing system comprising at least a processor retrieving instructions from the medium, carry out a method for determining an aggregated embedding vector for a target node of a corpus graph, the method comprising:

• accessing a target node of a corpus graph;
• determining an aggregated embedding vector relating to the target node, comprising:
• determining a relevant neighborhood of the target node within the corpus graph through execution by a central processing unit (CPU);
• and
• generating an aggregated embedding vector for the target node in view of the relevant neighborhood of the target node through execution by a graphic processing unit (GPU);
• wherein the aggregated embedding vector comprises an embedding vector of the target node concatenated with neighborhood embedding information for the target node determined from the relevant neighborhood of the target node;
• and
• associating the aggregated embedding vector with the target node.

A computer system configured as a recommender system for providing recommended content in response to a user query, the system comprising a processor and a memory, wherein the processor executes instructions stored in the memory as part of or in conjunction with additional components to respond to the user's query, the additional components comprising:

• a content recommender component that, in execution on the computer system:
• receives the query from the user over a network, the query comprising an item of content;
• identifies a corresponding item of content to the query in a corpus graph maintained by the recommender system;
• determines a set of potential content recommendation for the user according to an aggregated embedding vector associated with the corresponding item of content in the corpus graph;
• selects at least one content recommendation from the set of potential content recommendations;
• and
• provides the at least one content recommendation to the user;
• and
• an embedding vector generator that, in execution:
• accesses a target node of a corpus graph;
• and
• determines an aggregated embedding vector relating to the target node, comprising:
• determining a relevant neighborhood of the target node within the corpus graph through execution by a central processing unit (CPU);
• and
• generating an aggregated embedding vector for the target node in view of the relevant neighborhood of the target node through execution by a graphic processing unit (GPU);
• wherein the aggregated embedding vector comprises an embedding vector of the target node concatenated with neighborhood embedding information for the target node determined from the relevant neighborhood of the target node.