Systems and methods for classifying data sets using corresponding neural networks

1. A system for classifying data sets by corresponding functions, comprising:

at least one processor; and

at least one non-transitory memory storing instructions that, when executed by the at least one processor cause the system to perform operations comprising:

obtaining a first data set;

obtaining a neural network associated with a characteristic of the first data set, the neural network being trained to generate synthetic data sets having at least the characteristic of the first data set;

selecting a test set of inputs to the neural network;

obtaining a corresponding set of outputs, from the neural network, by applying the neural network to the test set of inputs, the neural network outputs comprising a second data set, the second data set comprising synthetic data;

estimating one or more corresponding functions that model the test set of inputs to the corresponding set of neural network outputs, the one or more functions comprising local approximations of a global function representing the neural network;

indexing the estimated one or more functions to the first data set; and

using the one or more functions to generate a third data set, the third data set comprising synthetic data and having the characteristic of the first data set.

2. The system of claim 1 , wherein the neural network comprises at least one of a convolutional neural network, a recurrent neural network, an auto-encoder, a variational auto-encoder, or a generative adversarial network.

3. The system of claim 1 , wherein the operations further comprise generating the test set of inputs using a stochastic algorithm.

4. The system of claim 3 , wherein the stochastic algorithm comprises a Monte Carlo algorithm.

5. The system of claim 1 , wherein the one or more functions comprise polynomial approximations.

6. The system of claim 1 , wherein the one or more functions comprise harmonic approximations.

7. The system of claim 1 , wherein the first data set comprises at least one of text files or image files.

8. The system of claim 1 , wherein the synthetic data sets are grouped within a same category as the first data set.

9. The system of claim 8 , wherein the category comprises an application generating the first data set.

10. The system of claim 8 , wherein the category comprises an output from a classifier applied to the first data set.

11. A system for classifying data sets by corresponding functions, comprising:

at least one processor; and

at least one non-transitory memory storing instructions that, when executed by the at least one processor cause the system to perform operations comprising:

obtaining a first data set;

training, using the first data set, a neural network to generate synthetic data sets having a characteristic associated with the first data set;

selecting a test set of inputs to the neural network;

obtaining a corresponding set of outputs, from the neural network, by applying the neural network to the test set of inputs, the neural network outputs comprising a second data set, the second data comprising synthetic data;

estimating one or more functions that model the test set of inputs to the corresponding set of neural network outputs, the one or more functions comprising local approximations of a global function representing the neural network;

indexing the estimated one or more functions to the first data set; and

using the one or more functions to generate a third data set, the third data set comprising synthetic data and having the characteristic of the first data set.

12. The system of claim 11 , wherein the first data set comprises at least one of text files or image files.

13. The system of claim 11 , wherein training the neural network further uses a plurality of test sets having an associated category matching a category of the first data set.

14. The system of claim 13 , wherein the category comprises an application generating the first data set and the test sets.

15. The system of claim 13 , wherein the category is determined by one or more classification models.

16. The system of claim 11 , wherein the operations further comprise generating the test set of inputs using a stochastic algorithm.

17. The system of claim 16 , wherein the stochastic algorithm comprises a Monte Carlo algorithm.

18. A system for classifying data sets by corresponding functions, comprising:

at least one processor; and

at least one non-transitory memory storing instructions that, when executed by the at least one processor cause the system to perform operations comprising:

obtaining a first data set;

generating a neural network with a structure comprising a plurality of nodes across a plurality of layers;

using the first data set, training the neural network to generate synthetic data sets having a characteristic of the first data set;

selecting a test set of inputs to the neural network;

obtaining a corresponding set of outputs, from the neural network, by applying the neural network to the test set of inputs, the neural network outputs comprising a second data set, the second data set being a synthetic data set;

estimating one or more functions that model the test set of inputs to the corresponding set of neural network outputs, the one or more corresponding functions comprising local approximations of a global function representing the neural network;

using the one or more functions, determine branches of the structure that are correlated with parts of the test set of inputs;

indexing the estimated one or more functions to the determined branches; and

using the one or more functions to generate a third data set, the third data set comprising synthetic data and having the characteristic of the first data set.