US 12,730,813 B1
Systems and methods for dynamic semantic memory graph-based query augmentation for natural language processing techniques
Ahmed A Khan, Plano, TX (US); Mohammed Alsadi, Redmond, WA (US); Rangesh Sripathi, Chennai (IN); and Prasannakumar Palanisamy, Chennai (IN)
Assigned to Verizon Patent and Licensing Inc., Basking Ridge, NJ (US)
Filed by Verizon Patent and Licensing Inc., Basking Ridge, NJ (US)
Filed on May 14, 2025, as Appl. No. 19/208,380.
Int. Cl. G06F 16/00 (2019.01); G06F 16/242 (2019.01); G06F 16/2455 (2019.01)
CPC G06F 16/2455 (2019.01) [G06F 16/243 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A device, comprising:
one or more processors configured to:
identify a plurality of tokens or intents associated with a plurality of queries;
maintain a plurality of nodes that are each associated with a particular respective token or intent of the plurality of tokens or intents;
identify that a particular query, of the plurality of queries, includes a first token or intent and a second token or intent, wherein a first node of the plurality of nodes is associated with the first token or intent and wherein a second node of the plurality of nodes is associated with the second token or intent;
determine a particular measure of relatedness between the first and second nodes based on identifying that the first token or intent and the second token or intent are present in the same particular query;
determine that the particular measure of relatedness, between the first and second nodes, is greater than a threshold measure of relatedness;
associate the first and second nodes with a particular cluster, based on determining that the particular measure of relatedness between the two particular nodes is greater than the threshold measure of relatedness,
wherein the particular cluster further includes a third node that is associated with a third token or intent of the plurality of tokens or intents,
wherein the particular cluster is associated with particular context information that is based on the third node;
receive a query;
identify a particular set of tokens or intents associated with the received query;
identify, based on the particular set of tokens or intents associated with the received query, that the query is associated with the particular cluster;
generate an augmented query based on the received query and further based on the particular context information that is based on the third node associated with the particular cluster; and
output the augmented query to a Natural Language Processing (“NLP”) system that generates a response based on the augmented query.