Systems and methods for detecting required rule engine updated using artificial intelligence models
The systems and methods provide a model deployment criterion. The model deployment criterion indicates a difference in a value against which the proxy model may be measured to determine when, if ever, the proxy model should be deployed to replace the existing rule engine. The model deployment criterion may be keyed to the proxy model (e.g., based on a difference in its size, throughput speed, number of changes, etc.), the existing rule engine (e.g., based on a difference in its age, update occurrences to its rule base, etc.), and/or comparisons between models (e.g., based on differences in results, throughput speed, efficiency, etc.).
1. A system for detecting required rule engine updated using artificial intelligence models that do not require training specific to model components and objectives, the system comprising:
one or more processors; and
one or more non-transitory, computer-readable media comprising instructions recorded thereon that when executed by the one or more processors cause operations comprising:
identifying a first ensemble component for a first artificial intelligence system, wherein the first artificial intelligence system comprises a plurality of ensemble components, and wherein the first artificial intelligence system comprises a first model architecture;
determining a first ensemble function for the first ensemble component in the first artificial intelligence system based on an output of the first ensemble component;
determining a first setting for the first ensemble component;
determining a first configuration location for the first ensemble component in the first model architecture based on the first setting;
selecting, using a second artificial intelligence system model, a second ensemble component from a plurality of available ensemble components based on the first ensemble function and the first configuration location, wherein the second artificial intelligence system is trained to select substitute ensemble components based on matching characteristics of known ensemble components and known artificial intelligence systems; and
modifying the first artificial intelligence system by replacing the first ensemble component with the second ensemble component.
2. A method for modifying architectures of artificial intelligence models without requiring training data that is specific to model components and objectives, the method comprising:
identifying a first component for a first artificial intelligence system, wherein the first artificial intelligence system comprises a plurality of components, and wherein the first artificial intelligence system comprises a first model architecture;
determining a first function for the first component in the first artificial intelligence system;
determining a first setting for the first component;
determining a first configuration location for the first component in the first model architecture based on the first setting;
selecting, using a second artificial intelligence system model, a second component from a plurality of available components based on the first function and the first configuration location, wherein the second artificial intelligence system is trained to select substitute components based on matching characteristics of known components and known artificial intelligence systems; and
modifying the first artificial intelligence system by replacing the first component with the second component.
3. The method of claim 2 , wherein the first component comprises an architecture component, and wherein identifying the first function for the first ensemble component in the first artificial intelligence system further comprises:
determining a model layer corresponding to the architecture component; and
identifying the first function based on the model layer.
4. The method of claim 3 , wherein determining the model layer corresponding to the architecture component further comprises:
determining an activation function for the architecture component; and
determining the model layer based on the activation function.
5. The method of claim 2 , wherein the first component comprises an ensemble component, and wherein identifying the first function for the first ensemble component in the first artificial intelligence system further comprises:
determining a plurality of ensemble models in the first artificial intelligence system;
determining an ensemble function for a first ensemble model in the plurality of ensemble models; and
identifying the first function based on the ensemble function.
6. The method of claim 5 , wherein determine the ensemble function for the first ensemble model in the plurality of ensemble models further comprises:
determining an ensemble output from the first ensemble model; and
determining the first ensemble function based on the ensemble output.
7. The method of claim 2 , wherein determining the first setting for the first component further comprises:
determining a learning rate for the first component; and
determining the first setting based on the learning rate.
8. The method of claim 2 , wherein determining the first setting for the first component further comprises:
determining a first hyperparameter for the first component; and
determining the first setting based on the first hyperparameter.
9. The method of claim 2 , wherein determining the first setting for the first component further comprises:
determining a first filter size for the first component; and
determining the first setting based on the first filter size.
10. The method of claim 2 , wherein determining the first configuration location for the first component in the first model architecture further comprises:
determining a first placement for the first component; and
determining the first configuration location based on the first placement.
11. The method of claim 2 , wherein modifying the first artificial intelligence system by replacing the first component with the second component further comprises:
determining an input interface for the first component;
determining an output interface for the first component;
preserving the input interface and the output interface; and
disconnecting the first component from the first artificial intelligence system.
12. The method of claim 2 , wherein modifying the first artificial intelligence system by replacing the first component with the second component further comprises:
determining a first data format for the first component;
determining a second data format for the second component;
determining a format difference between the first data format and the second data format; and
determining a compatibility of the second data format based on determining the format difference.
13. The method of claim 2 , wherein modifying the first artificial intelligence system by replacing the first component with the second component further comprises:
determining a first data dimension for the first component;
determining a second data dimension for the second component; and
converting the first data dimension to the second data dimension.
14. The method of claim 2 , wherein modifying the first artificial intelligence system by replacing the first component with the second component further comprises:
determining a first output specification for the first component;
determining a second output specification for the second component; and
mapping the second output specification to match the first output specification.
15. One or more non-transitory, computer-readable media comprising instructions recorded thereon that when executed by one or more processors cause operations comprising:
identifying a first component for a first artificial intelligence system, wherein the first artificial intelligence system comprises a plurality of components, and wherein the first artificial intelligence system comprises a first model architecture;
determining a first function for the first component in the first artificial intelligence system;
determining a first setting for the first component;
determining a first configuration location for the first component in the first model architecture based on the first setting;
selecting, using a second artificial intelligence system model, a second component from a plurality of available components based on the first function and the first configuration location, wherein the second artificial intelligence system is trained to select substitute components based on matching characteristics of known components and known artificial intelligence systems; and
modifying the first artificial intelligence system by replacing the first component with the second component.
16. The one or more non-transitory, computer-readable media of claim 15 , wherein the first component comprises an architecture component, and wherein identifying the first function for the first ensemble component in the first artificial intelligence system further comprises:
determining a model layer corresponding to the architecture component; and
identifying the first function based on the model layer.
17. The one or more non-transitory, computer-readable media of claim 16 , wherein determining the model layer corresponding to the architecture component further comprises:
determining an activation function for the architecture component; and
determining the model layer based on the activation function.
18. The one or more non-transitory, computer-readable media of claim 15 , wherein the first component comprises an ensemble component, and wherein identifying the first function for the first ensemble component in the first artificial intelligence system further comprises:
determining a plurality of ensemble models in the first artificial intelligence system;
determining an ensemble function for a first ensemble model in the plurality of ensemble models; and
identifying the first function based on the ensemble function.
19. The one or more non-transitory, computer-readable media of claim 18 , wherein determine the ensemble function for the first ensemble model in the plurality of ensemble models further comprises:
determining an ensemble output from the first ensemble model; and
determining the first ensemble function based on the ensemble output.