Methods and system for managing predictive models

Patent metadata

Field	Value
Publication number	US-10528872-B2
Application number	US-201414500990-A
Country	US
Kind code	B2
Filing date	Sep 29, 2014
Priority date	May 30, 2014
Publication date	Jan 7, 2020
Grant date	Jan 7, 2020

Abstract

Disclosed herein is a technique for implementing a framework that enables application developers to enhance their applications with dynamic adjustment capabilities. Specifically, the framework, when utilized by an application on a mobile computing device that implements the framework, can enable the application to establish predictive models that can be used to identify meaningful behavioral patterns of an individual who uses the application. In turn, the predictive models can be used to preempt the individual's actions and provide an enhanced overall user experience. The framework is configured to interface with other software entities on the mobile computing device that conduct various analyses to identify appropriate times for the application to manage and update its predictive models. Such appropriate times can include, for example, identified periods of time where the individual is not operating the mobile computing device, as well as recognized conditions where power consumption is not a concern.

First claim

What is claimed is:

1. A method, comprising:
at a mobile computing device:
caching, by a predictive model framework provided on a mobile computing device, a predictive model into a memory of the mobile computing device, the predictive model having been generated by a server computing device using a first set of training data provided by one or more computing devices, and the predictive model being stored as a contiguous file within the memory to facilitate increasing performance of the mobile computing device when updating the predictive model;
establishing, by the predictive model framework provided on the mobile computing device, a second set of training data that is specific to the predictive model;
identifying, by the predictive model framework provided on the mobile computing device, a time at which to update the predictive model based on the second set of training data, wherein identifying the time is based at least on a power-related metric of the mobile computing device;
retrieving, by the predictive model framework provided on the mobile computing device, the predictive model from the memory;
updating, by the predictive model framework provided on the mobile computing device, the predictive model based on the second set of training data to produce an updated predictive model, wherein updating the predictive model occurs without mapping the predictive model to a database based on the predictive model being stored as the contiguous file; and
caching, by the predictive model framework provided on the mobile computing device, the updated predictive model as a second contiguous file into the memory.

2. The method of claim 1 , wherein the predictive model is provided by the server computing device to the mobile computing device, and the predictive model is:
configured according to capabilities of the mobile computing device, and
based on a raw predictive model managed by the server computing device.

3. The method of claim 2 , further comprising:
receiving, from the server computing device, an update to the predictive model;
determining that the update to the predictive model is different from the predictive model cached in the memory; and
in response to determining, and based on the update to the predictive model:
updating or replacing the predictive model cached in the memory.

4. The method of claim 1 , wherein storing the predictive model as the contiguous file within the memory enables a hierarchical structure of the predictive model to be retained without requiring a deconstruction of the hierarchical structure of the predictive model prior to storing the predictive model into the memory.

5. The method of claim 4 , wherein, when the predictive model is retrieved from the memory, the hierarchical structure of the predictive model is intact and does not require a reconstruction of the predictive model.

6. The method of claim 4 , wherein the predictive model is contiguously stored within the memory based on a hardware configuration and/or software configuration of the mobile computing device.

7. The method of claim 4 , wherein the predictive model is contiguously stored within the memory based on a format that is specific to an application executing on the mobile computing device.

8. The method of claim 1 , wherein the power-related metric includes at least one of: a current charge level of a battery included in the mobile computing device, an average rate at which a charge level of the battery declines, or an indication whether the mobile computing device is plugged in to a power outlet and is charging.

9. The method of claim 1 , wherein identifying the time is further based on tracking activities that occur within the mobile computing device to identify a period of time where a user is not utilizing the mobile computing device.

10. The method of claim 1 , wherein an application executing on the mobile computing device utilizes an API call to access the predictive model framework, and the second set of training data is provided, by the application via a respective API call, to a training algorithm included in the predictive model framework.

11. A non-transitory computer readable storage medium configured to store instructions that, when executed by a processor included in a mobile computing device, cause the mobile computing device to perform operations comprising:
at the mobile computing device:
caching, by a predictive model framework provided on a mobile computing device, a predictive model into a memory of the mobile computing device, wherein the predictive model having been generated by a server computing device using a first set of training data provided by one or more computing devices, and
the predictive model is stored as a contiguous file within the memory to facilitate increasing performance of the mobile computing device when updating the predictive model;

establishing, by the predictive model framework provided on the mobile computing device, a second set of training data that is specific to the predictive model;
identifying, by the predictive model framework provided on the mobile computing device, a time at which to update the predictive model based on the second set of training data, wherein identifying the time is based at least on a power-related metric of the mobile computing device;
retrieving, by the predictive model framework provided on the mobile computing device, the predictive model from the memory;
updating, by the predictive model framework provided on the mobile computing device, the predictive model based on the second set of training data to produce an updated predictive model, wherein updating the predictive model occurs without mapping the predictive model to a database based on the predictive model being stored the contiguous file; and
caching, by the predictive model framework provided on the mobile computing device, the updated predictive model as a second contiguous file into the memory.

12. The non-transitory computer readable storage medium of claim 11 , wherein the predictive model is provided by the server computing device to the mobile computing device, and the predictive model is:
configured according to capabilities of the mobile computing device, and
based on a raw predictive model managed by the server computing device.

13. The non-transitory computer readable storage medium of claim 12 ,
further comprising:
receiving, from the server computing device, an update to the predictive model;
determining that the update to the predictive model is different from the predictive model cached in the memory; and
in response to determining, and based on the update to the predictive model:
updating or replacing the predictive model cached in the memory.

14. The non-transitory computer readable storage medium of claim 11 , wherein storing the predictive model as the contiguous file within the memory enables a hierarchical structure of the predictive model to be retained without requiring a deconstruction of the hierarchical structure of the predictive model prior to storing the predictive model into the memory.

15. The non-transitory computer readable storage medium of claim 14 , wherein, when the predictive model is retrieved from the memory, the hierarchical structure of the predictive model is intact and does not require a reconstruction of the predictive model.

16. The non-transitory computer readable storage medium of claim 14 , wherein the predictive model is contiguously stored within the memory based on a hardware configuration and/or software configuration of the mobile computing device.

17. The n

Assignees

• Apple Inc

Inventors

• Mathew Binu K
• Wan Kit-Man
• Kapoor Gaurav

Classifications

• G06N5/043

  Distributed expert systems; Blackboards · CPC title

• H04L67/10

  in which an application is distributed across nodes in the network (software deployment G06F8/60; multiprogramming arrangements G06F9/46) · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• G06N5/04 Primary

  Inference or reasoning models · CPC title