Review machine learning system

1. A processing device, comprising:

a non-transitory memory storing instructions; and

one or more processors in communication with the non-transitory memory, wherein the one or more processors execute the instructions to:

receive first data;

generate first features based on the first data;

identify a first set of labels for the first data;

train a first machine learning model, using the first features and the first set of labels;

determine whether a trigger has occurred; and

review the first machine learning model to generate a second machine learning model in response to determining that the trigger has occurred, the determining whether a trigger has occurred including determining whether a count of the labels in the second set of labels has exceeded a threshold, the reviewing comprising:

receiving a second set of labels for the first data, and

training the second machine learning model using the second set of labels and reusing the first features generated based on the first data.

2. The processing device of claim 1 , wherein the first data is received during a first time period, and second data is received during a second time period after the first time period.

3. The processing device of claim 1 , wherein the determination whether the trigger has occurred includes determining whether a timer has expired.

4. The processing device of claim 1 , wherein the determination whether the trigger has occurred includes determining whether an error value in connection with a performance of the first machine learning model has exceeded a threshold.

5. The processing device of claim 1 , wherein a generation of second features based on the second set of labels is avoided by reusing at least a portion of the first features in connection with the generation of the second machine learning model.

6. The processing device of claim 1 , wherein the second machine learning model is generated utilizing second features different from the first features generated based on the second set of labels in addition to at least a portion of the first features.

7. The processing device of claim 1 , wherein the second set of labels replaces the first set of labels.

8. The processing device of claim 1 , wherein the one or more processors execute the instructions to maintain model metadata by recording properties for the first machine learning model and the second machine learning model, the properties including a name, a portion of the first features selected for training, an algorithm, or a set of labels.

9. The processing device of claim 1 , wherein the one or more processors execute the instructions to maintain review metadata by recording properties for the first machine learning model and the second machine learning model, the properties including at least one of accuracy, a trigger event for each new machine learning model generation, a label, a time stamp, or a model name.

10. The processing device of claim 1 , wherein the one or more processors execute the instructions to store a feature and label table in a memory, the feature and label table including the first features and the first set of labels for the first data.

11. The processing device of claim 10 , wherein the one or more processors execute the instructions to update the feature and the label table in the memory to include the second set of labels for the first data.

12. The processing device of claim 1 , wherein a first portion of the first features is used to train the first machine learning model and a second portion of the first features that comprises a subset of the first portion is used to train the second machine learning model.

13. The processing device of claim 1 , wherein a first portion of the first features is used to train the first machine learning model and the first portion of the first features is used to train the second machine learning model.

14. A computer-implemented method comprising:

receiving first data;

generating first features based on the first data;

identifying a first set of labels for the first data;

training a first machine learning model, using the first features and the first set of labels;

determining whether a trigger has occurred; and

reviewing the first machine learning model to generate a second machine learning model in response to determining that the trigger has occurred, the determining whether a trigger occurred including determining whether a count of the labels in the second set of labels has exceeded a threshold, the reviewing comprising:

receiving a second set of labels for the first data, and

training the second machine learning model using the second set of labels and reusing the first features generated based on the first data.

15. The method of claim 14 , wherein a generation of second features based on the second set of labels is avoided by reusing at least a portion of the first features in connection with the generation of the second machine learning model.

16. The method of claim 14 , further comprising:

determining whether a trigger has occurred; and

generating the second machine learning model in response to determining that the trigger has occurred.

17. The method of claim 14 , wherein the second set of labels replaces the first set of labels.

18. A non-transitory computer-readable media storing computer instructions, that when executed by one or more processors, cause the one or more processors to perform the steps of:

receiving first data;

generating first features based on the first data;

identifying a first set of labels for the first data;

training a first machine learning model, using the first features and the first set of labels;

determining whether a trigger has occurred; and

reviewing the first machine learning model to generate a second machine learning model in response to determining that the trigger has occurred, the determining whether a trigger occurred including determining whether a count of the labels in the second set of labels has exceeded a threshold, the reviewing comprising:

receiving a second set of labels for the first data, and

training the second machine learning model using the second set of labels and reusing the first features generated based on the first data.