Conversational reasoning with knowledge graph paths for assistant systems

One or more computer-readable non-transitory storage media embodying software that is operable when executed to:

• receive, from a client system associated with a user, a query from the user;
• access a knowledge graph comprising a plurality of nodes and a plurality of edges connecting the nodes, wherein each node corresponds to an entity, and wherein each edge corresponds to a relationship between the entities corresponding to the connected nodes;
• determine, based on the query, one or more initial entities associated with the query;
• determine, by a conversational reasoning model, a path connecting nodes corresponding to the one or more initial entities and nodes corresponding to candidate entities based on one or more entity paths and one or more relation paths, wherein the one or more entity paths are determined based on optimizing a loss function formulated based on a context vector associated with the query and context vectors associated with the candidate entities, and wherein the one or more relation paths are determined based on relationships between the one or more initial entities and the candidate entities;
• select, based on the path by the conversational reasoning model, one or more candidate nodes from the knowledge graph corresponding to one or more candidate entities, respectively;
• generate a response based on the one or more initial entities and the one or more candidate entities;
• and
• send, to the client system in response to the query, instructions for presenting the response.

A system comprising:

• one or more processors;
• and a non-transitory memory coupled to the processors comprising instructions executable by the processors, the processors operable when executing the instructions to:
• receive, from a client system associated with a user, a query from the user;
• access a knowledge graph comprising a plurality of nodes and a plurality of edges connecting the nodes, wherein each node corresponds to an entity, and wherein each edge corresponds to a relationship between the entities corresponding to the connected nodes;
• determine, based on the query, one or more initial entities associated with the query;
• determine, by a conversational reasoning model, a path connecting nodes corresponding to the one or more initial entities and nodes corresponding to candidate entities based on one or more entity paths and one or more relation paths, wherein the one or more entity paths are determined based on optimizing a loss function formulated based on a context vector associated with the query and context vectors associated with the candidate entities, and wherein the one or more relation paths are determined based on relationships between the one or more initial entities and the candidate entities;
• select, based on the path by the conversational reasoning model, one or more candidate nodes from the knowledge graph corresponding to one or more candidate entities, respectively;
• generate a response based on the one or more initial entities and the one or more candidate entities;
• and
• send, to the client system in response to the query, instructions for presenting the response.