Real-time predictions based on machine learning models

A non-transitory computer readable storage medium for storing instructions that when executed by a computer processor cause the computer processor to perform steps for performing predictions, the steps comprising:

• receiving, by the client device from an online system, a regression based machine learning model del;
• receiving a request for a first task;
• responsive to determining that a level of expected accuracy for the first task is below the measure of accuracy of the regression based machine learning model:
• extracting features of the first task;
• executing the regression based machine learning model with the extracted features of the first task as input;
• and
• performing the first task using the result of execution of the regression based machine learning model;
• receiving a request for a second task;
• and
• responsive to determining that a level of expected accuracy for the second task is above the measure of accuracy of the regression based machine learning model:
• extracting features of the second task;
• serializing the extracted features of the second task;
• transmitting, to the online system, the serialized features for executing a second machine learning based model;
• receiving, from the online system, the output of the second machine learning based model from the online system;
• and
• performing the second task using the received output.

A computer system comprising:

• a computer processor;
• and
• a non-transitory computer readable storage medium for storing instructions that when executed by a computer processor cause the computer processor to perform steps for performing predictions, the steps comprising:
• receiving, by the client device from an online system, a regression based machine learning model and a measure of accuracy of the regression based machine learning model;
• receiving a request for a first task;
• responsive to determining that a level of expected accuracy for the first task is below the measure of accuracy of the regression based machine learning model:
• extracting features of the first task;
• executing the regression based machine learning model with the extracted features of the first task as input;
• and
• performing the first task using the result of execution of the regression based machine learning model;
• receiving a request for a second task;
• and
• responsive to determining that a level of expected accuracy for the second task is above the measure of accuracy of the regression based machine learning model:
• extracting features of the second task;
• serializing the extracted features of the second task;
• transmitting, to the online system, the serialized features for executing a second machine learning based model;
• receiving, from the online system, the output of the second machine learning based model from the online system;
• and
• performing the second task using the received output.