US 12,298,975 B2
Dynamic query planning and execution
Manish R. Baldua, San Jose, CA (US); Daniel K. Hewlett, Clarksville, MD (US); Gregory E. Pounds, San Jose, CA (US); Xie Lu, Sunnyvale, CA (US); Jonathan Pohl, Concord, MA (US); and Peter Rigano, San Francisco, CA (US)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Jan. 31, 2024, as Appl. No. 18/429,219.
Claims priority of provisional application 63/587,393, filed on Oct. 2, 2023.
Prior Publication US 2025/0110957 A1, Apr. 3, 2025
Int. Cl. G06F 16/2453 (2019.01); G06F 16/242 (2019.01)
CPC G06F 16/24542 (2019.01) [G06F 16/2425 (2019.01); G06F 16/243 (2019.01)] 19 Claims
OG exemplary drawing
 
1. A method comprising:
receiving, via a user interface of an application, a first query comprising a user request for information retrievable using a first set of data resources, the first query comprising at least one first query term;
configuring, by a processor, at least one prompt to cause a generative machine learning model to (i) translate the at least one first query term into a set of functions that can be executed to obtain at least one second query term using a second set of data resources, the at least one second query term related to the at least one first query term and different from the at least one first query term, and (ii) generate and output a plan that is executable to create a modified version of the first query based on the at least one second query term, the plan comprising the set of functions configured by the generative machine learning model to obtain the at least one second query term using the second set of data resources that is different from the first set of data resources;
applying, by the processor, the generative machine learning model to the configured at least one prompt;
executing, by the processor, the plan generated and outputted by the generative machine learning model to (i) determine the at least one second query term, and (ii) create the modified version of the first query based on the at least one second query term;
executing the modified version of the first query based on the at least one second query term to provide, via the user interface, a response to the first query; and
responsive to determining that a value of at least one performance metric for the generative machine learning model meets or exceeds a maximum threshold value or is less than or equal to a minimum threshold value, configuring the at least one prompt to reduce at least one of: a number of prompts contained in the at least one prompt, or a number of communications with the generative machine learning model to obtain the plan, or a number of instructions included in the at least one prompt.