Machine learning templates in a machine learning framework

1. A method of creating a template associated with an industrial problem, the method comprising:

receiving, from a machine learning library, two or more non-platform specific kernels that are each specific to an industrial related problem including a manifold learning kernel and at least one of a pre-processing kernel, a variable transformation kernel, and a post-processing kernel, wherein the two or more kernels are daisy-chained to pass information between each other using a standardized set of variables;

aggregating, via a wrapper, the two or more kernels into a template; and

integrating the template into a user interface by writing the wrapper around the template including automatically inserting the template within a system specific software module adapted to interface with a template-executing system.

2. The method of claim 1 , wherein integrating the template into a user interface further comprises:

adding a container image for the template.

3. The method of claim 2 , wherein the container image comprises a complete file system that contains code, runtime files, system tools, and system libraries so that the template can be executed regardless of its environment.

4. The method of claim 2 , wherein integrating the template into a user interface further comprises:

adding an html file to present parameters for a user to review and/or modify.

5. The method of claim 4 , wherein integrating the template into a user interface further comprises:

adding JAVASCRIPT and Angular directives to plot results in a graphical format.

6. The method of claim 4 , wherein integrating the template into a user interface further comprises:

adding shell script for a framework to call when a template is chosen to be executed.

7. The method of claim 6 , further comprising:

executing the template wherein execution of the template causes the shell script to be called and the shell script executes the container image.

8. The method of claim 1 , wherein the template comprises at least two of data quality assessment algorithms, missing value imputation algorithms, outlier removal algorithms, feature engineering algorithms, classifiers, regressions, anomaly detection algorithms, and post-processing algorithms.

9. The method of claim 1 , wherein each of the kernels in the template is adapted to determine a solution to the specific industrial related problem of a specific domain.

10. A non-transitory computer-readable medium comprising instructions that when executed by a processor perform a method associated with an industrial problem, the method comprising:

receiving, from a machine learning library, two or more non-platform specific kernels that are each specific to an industrial related problem including a manifold learning kernel and at least one of a pre-processing kernel, a variable transformation kernel, and a post-processing kernel, wherein the two or more kernels are daisy-chained to pass information between each other using a standardized set of variables;

aggregating, via a wrapper, the two or more kernels into a template;

integrating the template into a user interface by writing the wrapper around the template including automatically inserting the template within a system specific software module adapted to interface with a template-executing system; and

executing the template.

11. The medium of claim 10 , wherein integrating the template into a user interface comprises:

adding a container image for the template.

12. The medium of claim 11 , wherein the container image comprises a complete file system that contains code, runtime files, system tools, and system libraries so that the template can be executed regardless of its environment.

13. The medium of claim 11 , wherein integrating the template into a user interface further comprises:

adding html file to present parameters for a user to review and/or modify; and

adding JAVASCRIPT and Angular directives to plot results in a graphical format.

14. The medium of claim 13 , wherein integrating the template into a user interface further comprises:

adding shell script for a framework to call when a template is chosen to be executed.

15. The medium of claim 14 , wherein execution of the template causes the shell script to be called and the shell script executes the container image.

16. A system for detection of problems associated with a machine, the system comprising:

a processor; and

a non-transitory computer-readable medium comprising instructions that when executed by the processor perform a method associated with an industrial problem, the method comprising:

receiving, from a machine learning library, two or more non-platform specific kernels that are each specific to an industrial related problem including a manifold learning kernel and at least one of a pre-processing kernel, a variable transformation kernel, and a post-processing kernel, wherein the two or more kernels are daisy-chained to pass information between each other using a standardized set of variables;

aggregating, via a wrapper, the two or more kernels into a template;

integrating the template into a user interface by writing the wrapper around the template including automatically inserting the template within a system specific software module adapted to interface with a template-executing system; and

executing the template.

17. The system of claim 16 , wherein integrating the template into a user interface comprises:

adding a container image for the template.

18. The system of claim 17 , wherein the container image comprises a complete file system that contains code, runtime files, system tools, and system libraries so that the template can be executed regardless of its environment.

19. The system of claim 17 , wherein integrating the template into a user interface further comprises:

adding shell script for a framework to call when a template is chosen to be executed.

20. The system of claim 19 , wherein execution of the template causes the shell script to be called and the shell script executes the container image.