Generative graph-enhanced information retrieval

A system comprising:

• a processor;
• and
• a memory coupled to the processor, wherein the memory comprises an instruction that, when executed by the processor, causes the processor to:
• determine a first query portion and a second query portion of a query;
• match an embedding of the second query portion with an embedding of a portion of a digital content item;
• map the portion of the digital content item to a first node of a graph;
• by a generative machine learning model, construct a graph query based on at least the first node;
• via execution of the graph query on the graph, identify a second node of the graph that maps to the first query portion;
• extract a path from the graph, wherein the path comprises the first node, the second node, and an edge connecting the first node with the second node;
• and
• formulate the path for output at a device.

A non-transitory machine-readable storage medium comprising an instruction that, when executed by a processor, causes the processor to:

• determine a first query portion and a second query portion of a query;
• match an embedding of the second query portion with an embedding of a portion of a digital content item;
• map the portion of the digital content item to a first node of a graph;
• by a generative machine learning model, construct a graph query based on at least the first node;
• via the graph query, identify a second node of the graph that maps to the first query portion;
• and
• formulate a path for output at a device, wherein the path comprises the first node, the second node, and an edge connecting the first node with the second node.