US 2026/0310657 A1Published application

Systems and methods for context-aware query interpretation using generative artificial intelligence models

AssigneeAccenture Global Solutions Limited

InventorsChellappan MURUGAPPAN, Papiya BATABYAL, Deepak Kumar MALVIYA, Padmini PATEL, Chitra Ashok PHADKE, Ricardo Héctor DI PASQUALE, Marko KREMA, Bhagwat SINGH, María Paz GELPI, Balasubramaniam RAMACHANDRAN, José Ramón CABREJAS, Suchit MALHOTRA, Ragav DEVANATHAN, Rajendra PRASAD TANNIRU

Life of this patent

Estimating the term…

patentgrep

Loading patent…

Abstract

Systems and methods for context-aware query interpretation using Generative Artificial Intelligence (GenAI) models is disclosed. The system receives a request for generating a response corresponding to a domain from user. The system identifies entities and an intent of the request received. Also, the system determines candidate tables, fields, metrics, and/or dimensions. The system generates candidate executable database query followed by validating the candidate executable database query. The system generates output data by executing the validated executable database query. Moreover, the system computes a confidence score for the validated executable database query. The system generates a response based on the computed confidence score and the generated output data. Consequently, the system outputs the response on a user interface of a user device.

Claims20

Dependent claims

A system comprising:

a processor;
and
a memory communicably coupled to the processor, wherein the memory comprises processor-executable instructions which, when executed by the processor, cause the processor to:
receive a request for generating a response corresponding to a domain from at least one user, wherein the request comprises a natural language query;
identify a plurality of entities and an intent of the received request by parsing the natural language query using a hybrid vector and graph model;
determine at least one of candidate tables, fields, metrics, and dimensions relevant to the natural language query by querying a knowledge graph based on the identified plurality of entities and the intent, wherein the knowledge graph comprises at least one of table metadata nodes, field nodes, dimension nodes, organization term nodes, organization metric nodes, synonym nodes, acronym nodes, user-query nodes, tip nodes, table-behavior nodes, and weighted relationships between each type of node;
generate at least one candidate executable database query based on the determined candidate tables, the fields, the metrics, and the dimensions relevant to the natural language query using at least one Generative Artificial Intelligence (GenAI) model;
validate the generated at least one candidate executable database query using a syntactic analysis model and iteratively refine the at least one candidate executable database query to generate a validated executable database query;
generate output data corresponding to the natural language query by executing the validated executable database query on at least one structured data source;
compute a confidence score for the validated executable database query based on at least one of quantitative metrics and qualitative metrics using a classifier model;
generate a response to the natural language query based on the computed confidence score and the generated output data, wherein the response comprises the computed confidence score, and an explanation of reasoning performed in generating and validating the executable database query;
and
output the response on a user interface of a user device.

10 more elements

A method comprising:

receiving, by a processor, a request for generating a response corresponding to a domain from at least one user, wherein the request comprises a natural language query;
identifying, by the processor, a plurality of entities and an intent of the received request by parsing the natural language query using a hybrid vector and graph model;
determining, by the processor, at least one of candidate tables, fields, metrics, and dimensions relevant to the natural language query by querying a knowledge graph based on the identified plurality of entities and the intent, wherein the knowledge graph comprises at least one of table metadata nodes, field nodes, dimension nodes, organization term nodes, organization metric nodes, synonym nodes, acronym nodes, user-query nodes, tip nodes, table-behavior nodes, and weighted relationships between each type of node;
generating, by the processor, at least one candidate executable database query based on the determined candidate tables, the fields, the metrics, and the dimensions relevant to the natural language query using at least one Generative Artificial Intelligence (GenAI) model;
validating, by the processor, the generated at least one candidate executable database query using a syntactic analysis model, and iteratively refine the at least one candidate executable database query to generate a validated executable database query;
generating, by the processor, output data corresponding to the natural language query by executing the validated executable database query on at least one structured data source;
computing, by the processor, a confidence score for the validated executable database query based on at least one of quantitative metrics and qualitative metrics using a classifier model;
generating, by the processor, a response to the natural language query based on the computed confidence score and the generated output data, wherein the response comprises the computed confidence score, and an explanation of reasoning performed in generating and validating the executable database query;
and
outputting, by the processor, the response on a user interface of a user device.

7 more elements

A non-transitory computer readable medium comprising a processor-executable instructions that cause a processor to:

receive a request for generating a response corresponding to a domain from at least one user, wherein the request comprises a natural language query;
identify a plurality of entities and an intent of the received request by parsing the natural language query using a hybrid vector and graph model;
determine at least one of candidate tables, fields, metrics, and dimensions relevant to the natural language query by querying a knowledge graph based on the identified plurality of entities and the intent, wherein the knowledge graph comprises at least one of table metadata nodes, field nodes, dimension nodes, organization term nodes, organization metric nodes, synonym nodes, acronym nodes, user-query nodes, tip nodes, table-behavior nodes, and weighted relationships between each type of node;
generate at least one candidate executable database query based on the determined candidate tables, the fields, the metrics, and the dimensions relevant to the natural language query using at least one Generative Artificial Intelligence (GenAI) model;
validate the generated at least one candidate executable database query using a syntactic analysis model, and iteratively refine the at least one candidate executable database query to generate a validated executable database query;
generate output data corresponding to the natural language query by executing the validated executable database query on at least one structured data source;
compute a confidence score for the validated executable database query based on at least one of quantitative metrics and qualitative metrics using a classifier model;
generate a response to the natural language query based on the computed confidence score and the generated output data, wherein the response comprises the computed confidence score, and an explanation of reasoning performed in generating and validating the executable database query;
and
output the response on a user interface of a user device.

7 more elements

Dependencies are inferred from the claim text.