Method and system for supporting inductive reasoning queries over multi-modal data from relational databases

1. A method of performing queries, comprising:

generating text representations of features of various types of data;

building a multi-modal word embedding model to capture relationships between the various types of data; and

based on the multi-modal word embedding model, performing an inductive reasoning query,

wherein the building of the multi-modal word embedding model further includes generating a vector.

2. The method according to claim 1 , wherein based on the multi-modal word embedding model, performing the inductive reasoning query via a SQL (Structured Query Language) runtime, and

wherein the vector relates to at least a query from among the queries.

3. The method according to claim 1 , further comprising of supporting inductive reasoning queries over the various types of data being multi-modal data from relational databases.

4. The method according to claim 1 , wherein building the text representations includes generating a relational table with different SQL types as input and returns an unstructured text corpus consisting of a set of sentences.

5. The method according to claim 1 , wherein an externally pre-trained model is inputted into multi-modal word embedding model as a starting point for training.

6. The method according to claim 1 , wherein the building a multi-modal word embedding model further includes generating the vector for each feature being queried.

7. The method according to claim 1 , wherein performing the inductive reasoning query includes user-defined functions to take relational values as input and compute semantic relationships between them using uniformly untyped meaning vectors.

8. The method according to claim 1 being cloud implemented,.

wherein the building of the multi-modal word embedding model further includes generating the vector for each feature being queried, and

wherein performing the inductive reasoning query includes user-defined functions to take relational values as input and compute semantic relationships between them using uniformly untyped meaning vectors.

9. A system for performing queries, comprising:

a memory storing computer instructions; and

a processor configured to execute the computer instructions to:

generate text representations of features of various types of data; and

build a multi-modal word embedding model to capture relationships between the various types of data; and

based on the multi-modal word embedding model, perform an inductive reasoning query,

wherein the building of the multi-modal word embedding model further includes generating a vector.

10. The system according to claim 9 , wherein based on the multi-modal word embedding model, performing the inductive reasoning query via a SQL (Structured Query Language) runtime.

11. The system according to claim 9 , further comprising the processor configured to perform inductive reasoning queries over the various types of data being multi-modal data from relational databases.

12. The system according to claim 9 , wherein building the text representations includes generating a relational table with different SQL types as input and returns an unstructured text corpus consisting of a set of sentences.

13. The system according to claim 9 , wherein an externally pre-trained model is inputted into multi-modal word embedding model as a starting point for training.

14. The system according to claim 9 , wherein the building a multi-modal word embedding model further includes generating the vector for each feature being queried.

15. The system according to claim 9 , wherein performing the inductive reasoning query includes user-defined functions to take relational values as input and compute semantic relationships between them using uniformly untyped meaning vectors.

16. A computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions readable and executable by a computer to cause the computer to perform a method, comprising:

generating text representations of features of various types of data; and

building a multi-modal word embedding model to capture relationships between the various types of data; and

based on the multi-modal word embedding model, performing an inductive reasoning query.

17. The computer program product according to claim 16 , wherein based on the multi-modal word embedding model, performing the inductive reasoning query via a SQL (Structured Query Language) runtime.

18. The computer program product according to claim 16 , further comprising of supporting inductive reasoning queries over the various types of data being multi-modal data from relational databases.

19. The computer program product according to claim 16 , wherein building the text representations includes generating a relational table with different SQL types as input and returns an unstructured text corpus consisting of a set of sentences.

20. The computer program product according to claim 16 , wherein an externally pre-trained model is inputted into multi-modal word embedding model as a starting point for training.