Automated decision making for neural architecture search

1. A method for automating decision making for a neural architecture search in a computing environment by one or more processors comprising:

selecting one or more specifications for a dataset, tasks, and one or more constraints for a neural architecture search;

performing the neural architecture search based on the one or more specifications;

suggesting a deep learning model for the dataset, the tasks, and the one or more constraints based on the neural architecture search;

receiving an additional one or more specifications for the dataset, tasks, and one or more additional constraints for the neural architecture search; and

automatically outputting a prediction of a neural architecture search optimizer and a search space for performing a subsequent iteration of the neural architecture search to obtain an additional deep learning model to solve the tasks for the dataset based on information obtained from the suggestion of the deep learning model and previous neural architecture search.

2. The method of claim 1 , further including learning the one or more specifications from each previous neural architecture search.

3. The method of claim 1 , further including receiving the one or more specifications for the dataset, the tasks, and the one or more constraints, wherein the one or more constraints include at least an allowed neural architecture search time and a permissible number of parameters in a deep learning model.

4. The method of claim 1 , further including automatically selecting the search space and a selected machine learning model by the one or more constraints for the neural architecture search.

5. The method of claim 1 , further including detecting a change to the one or more specifications, wherein the one or more specifications include a dataset dimension, dataset type, data distribution data, key performance indicators (“KPIs”) and metrics, computational resources, a search space for the neural architecture search, or a combination thereof.

6. The method of claim 1 , further including recommending a modification to a previously identified deep learning model for the neural architecture search.

7. The method of claim 1 , further including initiating a machine learning models to:

searching the search space to identify the deep learning model that maximizes an objective function of each task for a dataset; or

learning one or more decisions and settings relating to previous neural architecture searches for performing the neural architecture search.

8. A system for automating decision making for a neural architecture search in a computing environment, comprising:

one or more computers with executable instructions that when executed cause the system to:

select one or more specifications for a dataset, tasks, and one or more constraints for a neural architecture search;

perform the neural architecture search based on the one or more specifications;

suggest a deep learning model for the dataset, the tasks, and the one or more constraints based on the neural architecture search;

receive an additional one or more specifications for the dataset, tasks, and one or more additional constraints for the neural architecture search; and

automatically output a prediction of a neural architecture search optimizer and a search space for performing a subsequent iteration of the neural architecture search to obtain an additional deep learning model to solve the tasks for the dataset based on information obtained from the suggestion of the deep learning model and previous neural architecture search.

9. The system of claim 8 , wherein the executable instructions when executed cause the system to learn the one or more specifications from each previous neural architecture search.

10. The system of claim 8 , wherein the executable instructions when executed cause the system to receive the one or more specifications for the dataset, the tasks, and the one or more constraints, wherein the one or more constraints include at least an allowed neural architecture search time and a permissible number of parameters in a deep learning model.

11. The system of claim 8 , wherein the executable instructions when executed cause the system to automatically select the search space and a selected machine learning model by the one or more constraints for the neural architecture search.

12. The system of claim 8 , wherein the executable instructions when executed cause the system to detect a change to the one or more specifications, wherein the one or more specifications include a dataset dimension, dataset type, data distribution data, key performance indicators (“KPIs”) and metrics, computational resources, a search space for the neural architecture search, or a combination thereof.

13. The system of claim 8 , wherein the executable instructions when executed cause the system to recommend a modification to a previously identified deep learning model for the neural architecture search.

14. The system of claim 8 , wherein the executable instructions when executed cause the system to initiate a machine learning models to:

search the search space to identify the deep learning model that maximizes an objective function of each task for a dataset; or

learn one or more decisions and settings relating to previous neural architecture searches for performing the neural architecture search.

15. A computer program product for, by a processor, automating decision making for a neural architecture search in a computing environment, the computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code portions stored therein, the computer-readable program code portions comprising:

an executable portion that selects one or more specifications for a dataset, tasks, and one or more constraints for a neural architecture search;

an executable portion that performs the neural architecture search based on the one or more specifications;

an executable portion that suggests a deep learning model for the dataset, the tasks, and the one or more constraints based on the neural architecture search;

an executable portion that receives an additional one or more specifications for the dataset, tasks, and one or more additional constraints for the neural architecture search; and

an executable portion that automatically outputs a prediction of a neural architecture search optimizer and a search space for performing a subsequent iteration of the neural architecture search to obtain an additional deep learning model to solve the tasks for the dataset based on information obtained from the suggestion of the deep learning model and previous neural architecture search.

16. The computer program product of claim 15 , further including an executable portion that learns the one or more specifications from each previous neural architecture search.

17. The computer program product of claim 15 , further including an executable portion that:

receives the one or more specifications for the dataset, the tasks, and the one or more constraints, wherein the one or more constraints include at least an allowed neural architecture search time and a permissible number of parameters in a deep learning model; and

automatically selects the search space and a selected machine learning model by the one or more constraints for the neural architecture search.

18. The computer program product of claim 15 , further including an executable portion that detects a change to the one or more specifications, wherein the one or more specifications include a dataset dimension, dataset type, data distribution data, key performance indicators (“KPIs”) and metrics, computational resources, a search space for the neural architecture search, or a combination thereof.

19. The computer program product of claim 15 , further including an executable portion that recommends a modification to a previously identified deep learning model for the neural architecture search.

20. The computer program product of claim 15 , further including an executable portion that initiates a machine learning models to:

search the search space to identify the deep learning model that maximizes an objective function of each task for a dataset; or

learn one or more decisions and settings relating to previous neural architecture searches for performing the neural architecture search.