Boosting AI identification learning

1. A machine-learning classification system comprising:

a first machine-learning classifier configured to:

receive a plurality of data items, each data item of the plurality of data items including one or more elements; and

classify each element of each data item to generate a plurality of classified data items;

a second machine-learning classifier coupled to the first machine-learning classifier and configured to:

identify misclassified elements of the plurality of classified data items; and

reclassify each of the identified misclassified elements to generate a first plurality of reclassified data items;

a third machine-learning classifier coupled to the first machine-learning classifier and configured to:

identify unclassified elements of the plurality of classified data items; and

classify each identified unclassified elements to generate a first plurality of newly-classified data items; and

a first ensemble classifier coupled to the first, second, and third machine-learning classifiers and configured to adjust the classifications of the elements of the plurality of classified data items in response to the first plurality of newly-classified data items and the first plurality of reclassified data items to generate the first ensemble classified data items.

2. The machine-learning classification system of claim 1 , wherein the second and third machine-learning classifiers are configured to execute in parallel.

3. The machine-learning classification system of claim 1 , wherein when the first and second classifiers generate different classifications for elements in a data item of the plurality of data items, the first ensemble classifier is configured to change the classification of the elements in the data item according to the classification generated by the second classifier to generate the first ensemble classified data items.

4. The machine-learning classification system of claim 1 , wherein when the third classifier generates a classification for elements in a data item that is not generated by the first classifier, the first ensemble classifier is configured to add the classified elements of the data item generated by the third classifier to the classified elements generated by the first classifier to generate the first ensemble classified data items.

5. The machine-learning classification system of claim 1 , wherein the data items are images including a plurality of pixels, and the elements are groups of pixels in the images.

6. The machine-learning classification system of claim 1 , wherein:

the first machine-learning classifier includes a first trained classifier that is trained using a first set of labeled data items; and

the second and third machine-learning classifiers include second and third trained classifiers that are trained on a second set of labeled data items different from the first set of labeled data items, based on classifications of the second set of labeled data items generated by the first trained classifier.

7. The machine-learning classification system of claim 1 , further comprising:

a fourth machine-learning classifier coupled to the ensemble classifier and configured to identify misclassified elements of the first ensemble classified data items and to reclassify the misclassified elements of the first ensemble classified elements to generate a second plurality of reclassified data items;

a fifth machine-learning classifier coupled to the ensemble classifier and configured to identify unclassified elements of the first ensemble classified data items and to classify the identified unclassified elements of the first ensemble classified data items to generate a second plurality of newly-classified data items; and

a second ensemble classifier coupled to the first ensemble classifier and to the fourth and fifth machine-learning classifiers and configured to change the classification of the elements of the plurality of first ensemble classified data items in response to the second plurality of reclassified data items and the second plurality of newly-classified data items to generate second ensemble classified data items.

8. The machine-learning classification system of claim 7 , wherein the fourth and fifth machine-learning classifiers are trained on a third set of labeled data items, different from the first and second sets of labeled data items, based on classifications of the third set of labeled data items generated by the first ensemble classifier.

9. The machine-learning classification system claim 1 , wherein each of the first and second machine-learning classifiers comprises a support vector classifier, a random forest classifier, a decision tree classifier, a neural network classifier, a genetic classifier, or a linear regression classifier.

10. A machine-learning classification method for classifying a plurality of data items, each data item of the plurality of data items including a plurality of elements, the method comprising:

classifying, by a first machine-learning classifier, each element of the plurality of data items to generate a plurality of classified data items;

identifying, by a second machine-learning classifier, unclassified elements of the plurality of classified data items to generate a first plurality of unclassified data items;

classifying, by the second machine-learning classifier, each unclassified element in the first plurality of unclassified data items into at least one of a plurality of categories to generate a first plurality of newly-classified data items;

identifying, by a third machine-learning classifier, misclassified elements of the plurality of classified data items to generate a first plurality of misclassified data items;

reclassifying, by the third machine-learning classifier, each of the misclassified elements in the first plurality of misclassified data items to generate a first plurality of reclassified data items; and

adjusting the classified data items using a first ensemble classifier in response to the first plurality of newly-classified data items and the first plurality of reclassified data items to generate the first ensemble classified data items.

11. The method of claim 10 , further comprising processing each of the first plurality of misclassified data items to change respective representations of the misclassified elements in the first plurality of misclassified data items before the reclassifying the misclassified elements in the first plurality of misclassified data items using the third machine-learning classifier.

12. The method of claim 11 , wherein the first plurality of unclassified data items and the first plurality of misclassified data items comprise a plurality of images, each including a plurality of pixels, and the unclassified elements and the misclassified elements include respective groups of pixels in the plurality of images.

13. The method of claim 10 , further comprising:

training the first machine-learning classifier using a first set of labeled data items; and

training the second and third machine-learning classifiers using a second set of labeled data items different from the first set of labeled data items, based on classifications of the second set of labeled data items generated by the first trained classifier.

14. The method of claim 10 , further comprising:

identifying, by a fourth machine-learning classifier, unclassified elements of the first ensemble classified data items;

classifying, by the fourth machine-learning classifier, the identified misclassified elements of the first ensemble classified elements to generate a second plurality of newly-classified data items;

identifying, by a fifth machine-learning classifier, misclassified elements of the first ensemble classified data items to generate a second plurality of misclassified data items;

reclassifying the misclassified elements of the second misclassified data items to generate a second plurality of reclassified data items; and

changing, by a second ensemble classifier, the classifications of the elements of the plurality of first ensemble classified data items in response to the second plurality of newly-classified data items and the second plurality of reclassified data items to generate second ensemble classified data items.

15. The method of claim 14 , comprising training the fourth and fifth machine-learning classifiers using a third set of labeled data items, different from the first and second sets of labeled data items, based on classifications of the third set of labeled data items generated by the first ensemble classifier.

16. A non-transitory computer-readable medium storing computer instructions for classifying a plurality of data items, each data item in the plurality of data items including a plurality of elements, the computer instructions, when executed by one or more processors, cause the one or more processors to:

implement a first machine-learning classifier that is configured to classify each element of the plurality of data items to generate a plurality of classified data items;

implement a second machine-learning classifier that is configured to identify unclassified elements of the plurality of classified data items and to classify each of the identified unclassified elements to generate a first plurality of newly-classified data items;

implement a third machine-learning classifier that is configured to classify misclassified elements of the plurality of classified data items and to reclassify each of the identified misclassified elements to generate a first plurality of reclassified data items; and

implement a first ensemble classifier that is configured to adjust the classifications of the elements of the plurality of classified data items in response to the first plurality of newly-classified data items and the first plurality of reclassified data items to generate the first ensemble classified data items.

17. The non-transitory computer-readable medium of claim 16 , wherein the computer instructions configure the one or more processors to operate the second and third machine-learning classifiers in parallel.