Intent recognition model creation from randomized intent vector proximities

1. A computer-implemented method, comprising:

by a processor:

generating, from an input intent vector, a plurality of candidate intent vectors;

performing a validation of the plurality of candidate intent vectors that selects as valid intent vectors any of the plurality of candidate intent vectors that are semantically similar to the input intent vector; and

iteratively processing the input intent vector in parallel with a noise vector such that each candidate intent vector of the plurality of candidate intent vectors is randomly distributed relative to the input intent vector.

2. The computer-implemented method of claim 1 , where the processor generating, from the input intent vector, the plurality of candidate intent vectors comprises the processor, for each candidate intent vector of the plurality of candidate intent vectors:

processing the input intent vector in parallel with the noise vector within a long short-term memory (LSTM) model.

3. The computer-implemented method of claim 1 , where the processor performing the validation of the plurality of candidate intent vectors that selects as the valid intent vectors any of the plurality of candidate intent vectors that are semantically similar to the input intent vector comprises the processor, for each candidate intent vector of the plurality of candidate intent vectors:

determining a multi-dimensional intent vector distance between the input intent vector and a respective candidate intent vector; and

selecting candidate intent vectors that are within a configured multi-dimensional intent vector distance of the input intent vector as the valid intent vectors that are semantically similar to the input intent vector.

4. The computer-implemented method of claim 1 , further comprising the processor:

in response to determining during the validation that a generated candidate intent vector is not semantically similar to the input intent vector, performing candidate intent vector generation feedback from the validation to the generating that adjusts vector generation parameters used to generate candidate intent vector; and

iteratively adjusting the vector generation parameters by use of the candidate intent vector generation feedback until a resulting generated candidate intent vector is semantically similar to the input intent vector.

5. The computer-implemented method of claim 1 , further comprising the processor creating an intent recognition model from the valid intent vectors that are semantically similar to the input intent vector.

6. The computer-implemented method of claim 1 , where at least one of the processor generating and the processor performing the validation is provided as a service in a cloud environment.

7. A system, comprising:

a memory; and

at least one processor(s) set programmed to:

generate within the memory, from an input intent vector, a plurality of candidate intent vectors;

perform a validation of the plurality of candidate intent vectors that selects as valid intent vectors any of the plurality of candidate intent vectors that are semantically similar to the input intent vector; and

iteratively process the input intent vector in parallel with a noise vector such that each candidate intent vector of the plurality of candidate intent vectors is randomly distributed relative to the input intent vector.

8. The system of claim 7 , where, in being programmed to generate within the memory , from the input intent vector, the plurality of candidate intent vectors, the at least one processor(s) set is programmed to, for each candidate intent vector of the plurality of candidate intent vectors:

process the input intent vector in parallel with the noise vector within a long short-term memory (LSTM) model.

9. The system of claim 7 , where, in being programmed to perform the validation of the plurality of candidate intent vectors that selects as the valid intent vectors any of the plurality of candidate intent vectors that are semantically similar to the input intent vector, the at least one processor(s) set is programmed to, for each candidate intent vector of the plurality of candidate intent vectors:

determine a multi-dimensional intent vector distance between the input intent vector and a respective candidate intent vector; and

select candidate intent vectors that are within a configured multi-dimensional intent vector distance of the input intent vector as the valid intent vectors that are semantically similar to the input intent vector.

10. The system of claim 7 , where the at least one processor(s) set is further programmed to:

in response to determining during the validation that a generated candidate intent vector is not semantically similar to the input intent vector, perform candidate intent vector generation feedback from the validation to the generating that adjusts vector generation parameters used to generate candidate intent vector; and

iteratively adjust the vector generation parameters by use of the candidate intent vector generation feedback until a resulting generated candidate intent vector is semantically similar to the input intent vector.

11. The system of claim 7 , where the at least one processor(s) set is further programmed to create an intent recognition model from the valid intent vectors that are semantically similar to the input intent vector.

12. A computer program product, comprising:

a computer readable storage medium having computer readable program code embodied therewith, where the computer readable storage medium is not a transitory signal per se and where the computer readable program code when executed on a computer causes the computer to:

generate, from an input intent vector, a plurality of candidate intent vectors;

perform a validation of the plurality of candidate intent vectors that selects as valid intent vectors any of the plurality of candidate intent vectors that are semantically similar to the input intent vector; and

iteratively process the input intent vector in parallel with a noise vector such that each candidate intent vector of the plurality of candidate intent vectors is randomly distributed relative to the input intent vector.

13. The computer program product of claim 12 , where, in causing the computer to generate, from the input intent vector, the plurality of candidate intent vectors, the computer readable program code when executed on the computer causes the computer to, for each candidate intent vector of the plurality of candidate intent vectors:

process the input intent vector in parallel with the noise vector within a long short-term memory (LSTM) model.

14. The computer program product of claim 12 , where, in causing the computer to perform the validation of the plurality of candidate intent vectors that selects as the valid intent vectors any of the plurality of candidate intent vectors that are semantically similar to the input intent vector, the computer readable program code when executed on the computer causes the computer to, for each candidate intent vector of the plurality of candidate intent vectors:

determine a multi-dimensional intent vector distance between the input intent vector and a respective candidate intent vector; and

select candidate intent vectors that are within a configured multi-dimensional intent vector distance of the input intent vector as the valid intent vectors that are semantically similar to the input intent vector.

15. The computer program product of claim 12 , where the computer readable program code when executed on the computer further causes the computer to:

in response to determining during the validation that a generated candidate intent vector is not semantically similar to the input intent vector, perform candidate intent vector generation feedback from the validation to the generating that adjusts vector generation parameters used to generate candidate intent vector; and

iteratively adjust the vector generation parameters by use of the candidate intent vector generation feedback until a resulting generated candidate intent vector is semantically similar to the input intent vector.

16. The computer program product of claim 12 , where the computer readable program code when executed on the computer further causes the computer to create an intent recognition model from the valid intent vectors that are semantically similar to the input intent vector.

17. The computer program product of claim 12 , where at least one of the computer generating and the computer performing the validation is provided as a service in a cloud environment.