Predictive model importation

• US 9,070,089 B1
• Filed: 10/07/2013
• Issued: 06/30/2015
• Est. Priority Date: 05/09/2011
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Lin, Wei-Hao, Green, Travis H. K., Kaplow, Robert, Fu, Gang, Mann, Gideon S.
• Primary Examiner(s): Holmes, Michael B
• Application Number:
  US14/047,576
• Time in Patent Office: 631 Days
• Field of Search: 706/21
• US Class Current: 1/1
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for obtaining a plurality of model representations of predictive models, each model representation associated with a respective user and expresses a respective predictive model, and selecting a model implementation for each of the model representations based on one or more system usage properties associated with the user associated with the corresponding model representation.

Claims

1. A computer-implemented method comprising:

obtaining a model representation of a predictive model, wherein the model representation is associated with a respective user and expresses a respective predictive model; and

selecting a model implementation for the model representation, the model implementation comprising one or more computer programs operable to be executed on one or more computing devices, and the model implementation being selected for the corresponding model representation based on one or more system usage properties associated with the user associated with the corresponding model representation; and

causing execution of the selected model implementation.

2. The method of claim 1, further comprising:

obtaining a second model representation of a second predictive model;

selecting a second model implementation for the second model representation; and

associating each of the selected model implementations with a node in a directed graph, wherein for one or more ordered pairs of nodes in the graph the prediction output of a model implementation associated with a tail node in the pair serves as input to a model implementation associated with a head node in the pair.

3. The method of claim 2, wherein causing execution of the selected model implementation comprises executing each model implementation in an order prescribed by the directed graph.

4. The method of claim 1, further comprising:

assigning a system resources quota to the user;

determining an amount of system resources that the user has consumed over a given period of time; and

selecting the model implementation based on an amount of unconsumed system resources according to the system resource quota, wherein an amount of system resources that can be consumed by the selected model implementation does not exceed the amount of unconsumed system resources.

5. The method of claim 1, further comprising:

selecting a computationally-intensive model implementation for the model representation based on a determination that system resources were previously used during an off-peak time of day.

6. The method of claim 1, further comprising:

determining an account balance for the user based on an amount of money that was previously paid by the user for system resources; and

selecting the model implementation based on the account balance, wherein an estimated amount of money that the user can be charged for use of the selected model implementation does not result in a negative account balance.

7. A storage medium having instructions stored thereon that, when executed by one or more computers, cause the one or more computers to perform a method comprising:

obtaining a model representation of a predictive model, wherein the model representation is associated with a respective user and expresses a respective predictive model; and

selecting a model implementation for the model representation, the model implementation comprising one or more computer programs operable to be executed on

one or more computing devices, and the model implementation being selected for the corresponding model representation based on one or more system usage properties associated with the user associated with the corresponding model representation; and

causing execution of the selected model implementation.

8. The storage medium of claim 7, wherein the method further comprises:

obtaining a second model representation of a second predictive model;

selecting a second model implementation for the second model representation; and

associating each of the selected model implementations with a node in a directed graph, wherein for one or more ordered pairs of nodes in the graph the prediction output of a model implementation associated with a tail node in the pair serves as input to a model implementation associated with a head node in the pair.

9. The storage medium of claim 8, wherein causing execution of the selected model implementation comprises:

executing each model implementation in an order prescribed by the directed graph.

10. The storage medium of claim 7, wherein the method further comprises:

assigning a system resources quota to the user;

determining an amount of system resources that the user has consumed over a given period of time; and

selecting the model implementation based on an amount of unconsumed system resources according to the system resource quota, wherein an amount of system resources that can be consumed by the selected model implementation does not exceed the amount of unconsumed system resources.

11. The storage medium of claim 7, wherein the method further comprises:

selecting a computationally-intensive model implementation for the model representation based on a determination that system resources were previously used during an off-peak time of day.

12. The storage medium of claim 7, wherein the method further comprises:

determining an account balance for the user based on an amount of money that was previously paid by the user for system resources; and

selecting the model implementation based on the account balance, wherein an estimated amount of money that the user can be charged for use of the selected model implementation does not result in a negative account balance.

13. A system comprising:

one or more computers configured to perform a method comprising;

obtaining a model representation of a predictive model, wherein the model representation is associated with a respective user and expresses a respective predictive model; and

selecting a model implementation for the model representation, the model implementation comprising one or more computer programs operable to be executed on one or more computing devices, and the model implementation being selected for the corresponding model representation based on one or more system usage properties associated with the user associated with the corresponding model representation; and

causing execution of the selected model implementation.

14. The system of claim 13, wherein the method further comprises:

obtaining a second model representation of a second predictive model;

selecting a second model implementation for the second model representation; and

associating each of the selected model implementations with a node in a directed graph, wherein for one or more ordered pairs of nodes in the graph the prediction output of a model implementation associated with a tail node in the pair serves as input to a model implementation associated with a head node in the pair.

15. The system of claim 14, wherein causing execution of the selected model implementation comprises:

executing each model implementation in an order prescribed by the directed graph.

16. The system of claim 13, wherein the method further comprises:

assigning a system resources quota to the user;

determining an amount of system resources that the user has consumed over a given period of time; and

selecting the model implementation based on an amount of unconsumed system resources according to the system resource quota, wherein an amount of system resources that can be consumed by the selected model implementation does not exceed the amount of unconsumed system resources.

17. The system of claim 13, wherein the method further comprises:

selecting a computationally-intensive model implementation for the model representation based on a determination that system resources were previously used during an off-peak time of day.

18. The system of claim 13, wherein the method further comprises:

determining an account balance for the user based on an amount of money that was previously paid by the user for system resources; and

selecting the model implementation based on the account balance, wherein an estimated amount of money that the user can be charged for use of the selected model implementation does not result in a negative account balance.

19. The system of claim 13, wherein causing execution of the selected model implementation comprises:

causing execution of the selected model implementation by one or more computers that are external to the one or more computers of the system.