Generation of a metadata-driven artificial intelligence platform

1. A method, comprising:

generating, via a computational device, a set of metafiles that store at least metadata information and provenance information of an artificial intelligence (AI) process that is trained with a source data;

accessing, via the computational device, the set of metafiles via an application programming interface (API) to the set of metafiles; and

in response to accessing the set of metafiles, transferring, via the computational device, the source data in the set of metafiles to a cache for processing by the AI process, wherein the set of metafiles is comprised of an input metafile and a runtime metafile, and wherein the runtime metafile stores information related to a runtime environment of the AI process.

2. The method of claim 1 , wherein the runtime metafile stores the provenance information, and wherein the provenance information comprises information relating to a framework, an algorithm, and a model of the AI process.

3. The method of claim 1 , wherein the input metafile stores at least filenames and resolvable addresses of each file of the source data.

4. The method of claim 1 , wherein the API is comprised of:

a first interface to access the set of metafiles to retrieve metadata;

a second interface to generate the cache;

a third interface to trigger actions including moving or copying of data to the cache; and

a fourth interface to record the provenance information in the runtime metafile.

5. A system, comprising:

a memory; and

a processor coupled to the memory, wherein the processor executes operations, the operations comprising:

generating a set of metafiles that store at least metadata information and provenance information of an artificial intelligence (AI) process that is trained with a source data;

accessing the set of metafiles via an application programming interface (API) to the set of metafiles; and

in response to accessing the set of metafiles, transferring the source data in the set of metafiles to a cache for processing by the AI process, wherein the set of metafiles is comprised of an input metafile and a runtime metafile, and wherein the runtime metafile stores information related to a runtime environment of the AI process.

6. The system of claim 5 , wherein the runtime metafile stores the provenance information, and wherein the provenance information comprises information relating to a framework, an algorithm, and a model of the AI process.

7. The system of claim 5 , wherein the input metafile stores at least filenames and resolvable addresses of each file of the source data.

8. The system of claim 5 , wherein the API is comprised of:

a first interface to access the set of metafiles to retrieve metadata;

a second interface to generate the cache;

a third interface to trigger actions including moving or copying of data to the cache; and

a fourth interface to record the provenance information in the runtime metafile.

9. A computer program product comprising a computer readable storage medium having computer readable program code embodied therewith, the computer readable program code configured to perform operations, the operations comprising:

generating a set of metafiles that store at least metadata information and provenance information of an artificial intelligence (AI) process that is trained with a source data;

accessing the set of metafiles via an application programming interface (API) to the set of metafiles; and

in response to accessing the set of metafiles, transferring the source data in the set of metafiles to a cache for processing by the AI process, wherein the set of metafiles is comprised of an input metafile and a runtime metafile, and wherein the runtime metafile stores information related to a runtime environment of the AI process.

10. The computer program product of claim 9 , wherein the runtime metafile stores the provenance information, and wherein the provenance information comprises information relating to a framework, an algorithm, and a model of the AI process.

11. The computer program product of claim 9 , wherein the input metafile stores at least filenames and resolvable addresses of each file of the source data.

12. The computer program product of claim 9 , wherein the API is comprised of:

a first interface to access the set of metafiles to retrieve metadata;

a second interface to generate the cache;

a third interface to trigger actions including moving or copying of data to the cache; and

a fourth interface to record the provenance information in the runtime metafile.

13. An artificial intelligence platform for managing a set of artificial intelligence processes in a computational device, the artificial intelligence platform comprising:

a set of metafiles, wherein the set of metafiles store at least metadata information and provenance information of an artificial intelligence (AI) process that is trained with a source data;

an application programming interface (API) for accessing the set of metafiles; and

a cache, to which the source data in the set of metafiles is transferred for processing by the AI process, in response to accessing the set of metafiles, wherein the set of metafiles is comprised of an input metafile and a runtime metafile, and wherein the runtime metafile stores information related to a runtime environment of the AI process.

14. The artificial intelligence platform of claim 13 , wherein the runtime metafile stores the provenance information, and wherein the provenance information comprises information relating to a framework, an algorithm, and a model of the AI process.

15. The artificial intelligence platform of claim 13 , wherein the input metafile stores at least filenames and resolvable addresses of each file of the source data.

16. The artificial intelligence platform of claim 13 , wherein the API is comprised of:

a first interface to access the set of metafiles to retrieve metadata;

a second interface to generate the cache;

a third interface to trigger actions including moving or copying of data to the cache; and

a fourth interface to record the provenance information in the runtime metafile.

17. An application programming interface (API) for a metadata-driven artificial intelligence platform implemented in a computational device, wherein the API is comprised of:

a first interface to access a set of metafiles to retrieve metadata;

a second interface to generate a cache;

a third interface to trigger actions including moving or copying of data to the cache; and

a fourth interface to record a provenance information in a runtime metafile, wherein the set of metafiles is comprised of an input metafile and the runtime metafile, and wherein the runtime metafile stores information related to a runtime environment of an Artificial Intelligence (AI) process.

18. The application programing interface of claim 17 , wherein the metadata-driven artificial intelligence platform is comprised of:

the set of metafiles, wherein the set of metafiles store at least metadata information and the provenance information of the AI process wherein the AI process is trained with a source data;

the application programming interface for accessing the set of metafiles; and

the cache to which the source data in the set of metafiles is transferred for processing by the AI process, in response to accessing the set of metafiles.

19. The application programing interface of claim 18 , wherein the runtime metafile stores the provenance information, and wherein the provenance information comprises information relating to a framework, an algorithm, and a model of the AI process.

20. The application programing interface of claim 18 , wherein the input metafile stores at least filenames and resolvable addresses of each file of the source data.