US 12,393,878 B2
Systems and methods for managing, distributing and deploying a recursive decisioning system based on continuously updating machine learning models
Alex Muller, Kentfield, CA (US)
Assigned to SAVVI AI INC., Kentfield, CA (US)
Filed by SAVVI AI INC., Kentfield, CA (US)
Filed on May 23, 2024, as Appl. No. 18/672,889.
Application 18/672,889 is a continuation of application No. 17/930,511, filed on Sep. 8, 2022, granted, now 12,039,424.
Application 17/930,511 is a continuation of application No. 17/590,181, filed on Feb. 1, 2022, granted, now 11,501,214, issued on Nov. 15, 2022.
Claims priority of provisional application 63/146,484, filed on Feb. 5, 2021.
Prior Publication US 2024/0362541 A1, Oct. 31, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 20/00 (2019.01); G06F 18/21 (2023.01); G06F 18/2113 (2023.01)
CPC G06N 20/00 (2019.01) [G06F 18/2113 (2023.01); G06F 18/2193 (2023.01)] 21 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
dynamically updating a dataset, wherein the dataset is dynamically updated according to client requests processed using a machine learning-enabled decision engine (MLDE), and wherein the dataset is dynamically updated by adding new data points to the dataset according to the client requests;
detecting that a pre-established triggering threshold for generating a set of new predictive models is met, wherein the pre-established triggering threshold is met based on the new data points being added to the dataset;
generating a set of sub-datasets from the dataset, wherein the set of sub-datasets is generated according to an identified number of new predictive models being generated for the set of new predictive models;
automatically engineering a set of machine learning features, wherein the set of machine learning features are engineered based on a set of influencing data inputs and decision data inputs associated with the MLDE;
setting a set of predictive targets for the set of new predictive models;
generating the set of new predictive models, wherein the set of new predictive models is generated using the set of sub-datasets and according to the set of machine learning features and the set of predictive targets;
selecting one or more new predictive models from the set of new predictive models, wherein the one or more new predictive models are selected by evaluating the set of new predictive models according to a model statistic;
updating the MLDE to implement the one or more new predictive models; and
converting the updated MLDE into a microservice, wherein the microservice generates ranked lists of decision options (RLDOs) for different use cases, and wherein the one or more new predictive models associated with the updated MLDE are configured to select and present a decision option from an RLDO according to a provided set of influencing data inputs associated with a user and according to a use case.