Automated and customized post-production release review of a model

1. A non-transitory computer readable medium comprising computer executable instructions that, when executed, cause at least one processor to:

train a model to perform at least one of a prediction operation, a diagnostic operation, or a classification operation based on a training dataset;

deploy the model in a production computer system to perform the at least one operation on field data;

monitor signal data associated with the model, the signal data including specific or derived signal data representing characteristics of an ecosystem in which the model is deployed and new observations in incoming field data;

monitor accuracy of the model by applying a statistical tool to a plurality of data points of the signal data;

apply a secondary machine learning predictive engine to the plurality of data points of the signal data to predict future data points of the signal data;

determine whether the signal data represents an unstable process by identifying future outlier data points from among the plurality of future data points of the signal data;

select a rule corresponding to the future outlier data points, the rule to suggest at least one of a cause of the unstable process or an effect of the unstable process on the signal data; and

generate an indication that a corrective action should be taken on the model based on a result of the determination, the indication to identify the rule and at least one of the cause of the unstable process or the effect of the unstable process on the signal data.

2. The non-transitory computer readable medium of claim 1 , wherein the instructions, when executed, cause the at least one processor to:

perform the corrective action on the model to retrain and redeploy the model or create a new model responsive to the indication; and

continuously monitor accuracy of the retrained and redeployed model or the created new model by applying the statistical tool to the signal data.

3. The non-transitory computer readable medium of claim 1 , wherein the secondary machine learning predictive engine is at least one of an autoregressive integrated moving average model or an autoregressive moving average mode model.

4. The non-transitory computer readable medium of claim 1 , wherein the signal data includes one or more of: file type data, internal process data, external process data, feature data, data decay data, mathematical verification data, model fitting data, machine learning periodicity data, network architecture data, design space boundary data, scale data, algorithm data, ground truth data, noise data, source of error data, contractual or liability data, model telemetry data, behavior metrics data, significant growth data, or base rate data.

5. The non-transitory computer readable medium of claim 1 , wherein the instructions, when executed, cause the at least one processor to identify the future outlier data points based on locations of the future data points of the signal data with respect to at least three sigma zones defined by corresponding standard deviations from the model, respectively.

6. The non-transitory computer readable medium of claim 1 , wherein the indication further identifies a probability corresponding to the rule.

7. The non-transitory computer readable medium of claim 1 , wherein the instructions, when executed, cause the at least one processor to apply statistical process control to the plurality of future data points of the signal data.

8. The non-transitory computer readable medium of claim 1 , wherein the instructions, when executed, cause the at least one processor to notify a user to at least one of retrain and redeploy the model, or create a new model.

9. A method for automated assessment of a model, the method comprising:

training, by executing an instruction with at least one processor, a model to perform at least one of a prediction operation, a diagnostic operation, or a classification operation based on a training dataset;

deploying, by executing an instruction with the at least one processor, the model in a production computer system to perform the at least one operation on field data;

monitoring, by executing an instruction with the at least one processor, signal data associated with the model, the signal data including specific or derived signal data representing characteristics of an ecosystem in which the model is deployed and new observations in incoming field data;

monitoring, by executing an instruction with the at least one processor, accuracy of the model by applying a statistical tool to a plurality of data points of the signal data;

applying, by executing an instruction with at least one processor, a secondary machine learning predictive engine to the plurality of data points of the signal data to predict future data points of the signal data;

determining, by executing an instruction with the at least one processor, whether the signal data represents an unstable process by identifying future outlier data points from among the plurality of future data points of the signal data;

selecting, by executing an instruction with the at least one processor, a rule corresponding to the future outlier data points, the rule to suggest at least one of a cause of the unstable process or an effect of the unstable process on the signal data; and

generating, by executing an instruction with the at least one processor, an indication that a corrective action should be taken on the model based on a result of the determination, the indication to identify the rule and at least one of the cause of the unstable process or the effect of the unstable process on the signal data; and

displaying, by executing an instruction with the at least one processor, the indication on a display.

10. The method of claim 9 , further including:

performing, by executing an instruction with the at least one processor, the corrective action on the model to retrain and redeploy the model or create a new model responsive to the indication; and

continuously monitoring, by executing an instruction with the at least one processor, accuracy of the retrained and redeployed model or the created new model by applying the statistical tool to the signal data.

11. The method of claim 9 , wherein the secondary machine learning predictive engine is at least one of an autoregressive integrated moving average model or an autoregressive moving average mode model.

12. The method of claim 9 , wherein the signal data includes one or more of: file type data, internal process data, external process data, feature data, data decay data, mathematical verification data, model fitting data, machine learning periodicity data, network architecture data, design space boundary data, scale data, algorithm data, ground truth data, noise data, source of error data, contractual or liability data, model telemetry data, behavior metrics data, significant growth data, or base rate data.

13. The method of claim 9 , further including identifying, by executing an instruction with the at least one processor, the future outlier data points based on locations of the future data points of the signal data with respect to at least three sigma zones defined by corresponding standard deviations from the model, respectively.

14. The method of claim 9 , wherein the indication further identifies a probability corresponding to the rule.

15. The method of claim 9 , further including applying, by executing an instruction with the at least one processor, statistical process control to the plurality of future data points of the signal data.

16. The method of claim 9 , further including notifying, by executing an instruction with the at least one processor, a user to at least one of retrain and redeploy the model, or create a new model.

17. A computer system configured for automated assessment of a model, comprising:

memory; and

at least one processor to execute machine readable instructions to:

train a model to perform at least one of a prediction operation, a diagnostic operation, or a classification operation based on a training dataset;

deploy the model in a production computer system to perform the at least one operation on field data;

monitor signal data associated with the model, the signal data including specific or derived signal data representing characteristics of an ecosystem in which the model is deployed and new observations in incoming field data;

monitor accuracy of the model by applying a statistical tool to a plurality of data points of the signal data;

apply a secondary machine learning predictive engine to the plurality of data points of the signal data to predict future data points of the signal data;

determine whether the signal data represents an unstable process by identifying future outlier data points from among the plurality of future data points of the signal data;

select a rule corresponding to the future outlier data points, the rule to suggest at least one of a cause of the unstable process or an effect of the unstable process on the signal data; and

generate an indication that a corrective action should be taken on the model based on a result of the determination, the indication to identify the rule and at least one of the cause of the unstable process or the effect of the unstable process on the signal data.

18. The computer system of claim 17 , wherein the at least one processor is to identify the future outlier data points based on locations of the future data points of the signal data with respect to at least three sigma zones defined by corresponding standard deviations from the model, respectively.

19. The computer system of claim 17 , wherein the secondary machine learning predictive engine is at least one of an autoregressive integrated moving average model or an autoregressive moving average mode model.

20. The computer system of claim 17 , wherein the at least one processor is to:

perform the corrective action on the model to retrain and redeploy the model or create a new model responsive to the indication; and

continuously monitor accuracy of the retrained and redeployed model or the created new model by applying the statistical tool to the signal data.