CPC G06F 8/60 (2013.01) [G06F 8/71 (2013.01); G06F 9/54 (2013.01); G06F 11/3409 (2013.01); G06N 20/00 (2019.01)] | 20 Claims |
1. A method, comprising:
receiving, by a particular computer system included in an enterprise computer system, deployment instructions, from a user, for a particular version of a machine-learning model, wherein one or more versions of the machine-learning model are stored in a database;
selecting, by the particular computer system and based on the deployment instructions, a destination within the enterprise computer system for deploying the particular version, wherein the selected destination provides access to a particular data set, and wherein the particular data set includes personal information associated with one or more users of the enterprise computer system;
scheduling, by the particular computer system, a deployment of the particular version from the database to the selected destination, wherein the deployed version of the machine-learning model operates on the particular data set, and wherein operations performed on the particular data set include determining whether storage conditions of data items of the particular data set are in accordance with a set of data management rules for storing data items in the selected destination, and wherein a stop time is scheduled as part of the deployment;
collecting performance data associated with operation of the deployed version of the machine-learning model;
at the scheduled stop time, determining, by the particular computer system using the performance data, whether to roll back the deployed version of the machine-learning model; and
based on the determining, rolling back, by the particular computer system, the deployed version of the machine-learning model to a different version stored in the database.
|