US 12,443,527 B1
Caching strategy based on model execution time, frequency and input order with configurable priority
Aneesh Dahiya, Zurich (CH); and Renata Khasanova, Zurich (CH)
Assigned to Oracle International Corporation, Redwood Shores, CA (US)
Filed by Oracle International Corporation, Redwood Shores, CA (US)
Filed on Apr. 11, 2024, as Appl. No. 18/632,946.
Int. Cl. G06F 12/0802 (2016.01); G06F 12/123 (2016.01)
CPC G06F 12/0802 (2013.01) [G06F 12/123 (2013.01); G06F 2212/60 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
receiving an input for a deterministic model;
determining whether the input and an output from the model corresponding to the input are stored in a cache;
in accordance with the input and the output not being stored in the cache:
computing the output from the model based on the input,
storing the input and the output in the cache as an element of the cache, and
storing a first raw score indicating a usage frequency of the element, a second raw score indicating an order of usage of the element, and a third raw score indicating a computation time for computing the output based on the input;
in accordance with the input and the output being stored in the cache, updating the first raw score and the second raw score; and
calculating a caching score associated with the element, wherein the caching score comprises a sum of a first normalized score corresponding to a product of the first raw score and a first scaling factor, a second normalized score corresponding to a product of the second raw score and a second scaling factor, and a third normalized score corresponding to a product of the third raw score and a third scaling factor.