Methods and systems for training a machine learning system using a reduced data set

1. A method for training a model using a reduced dataset, the method comprising:

receiving a first sequence of characters;

associating a first subset of the first sequence of characters with a first symbol, wherein the first subset corresponds to a first type of character;

associating a second subset of the first sequence of characters with a second symbol, wherein the second subset corresponds to a second type of character;

determining a first arrangement of the first and second symbols based on the first and second subsets within the first sequence of characters;

receiving user input specifying a first classification for the first symbol and a second classification for the second symbol;

automatically associating the first and second classifications with a second sequence of characters that are in a second arrangement that matches the first arrangement;

feeding the first and second sequences to the model as the reduced dataset; and

training the model with both the reduced dataset and the user specified first and second classifications.

2. The method of claim 1 , further comprising:

generating for display the first classification for the first symbol concurrently with the first subset to a second user to verify classification; and

receiving a further user input from the second user specifying a third classification for the first symbol, the third classification differing from the first classification.

3. The method of claim 1 , further comprising:

extracting first textual information and first visual clues from the first sequence of characters;

translating the first visual clues to a descriptive vector associated with the first sequence of characters;

processing the first textual information of the first sequence to associate the first subset with the first symbol and the second subset with the second symbol; and

extracting second textual information and second visual clues from the second sequence of characters,

wherein training the model further comprises training the model with both the reduced dataset and the descriptive vector.

4. The method of claim 3 , wherein the first and second visual clues comprise text color, bold font, italics, indentations, spacing, graphics, icons, and glyphs.

5. The method of claim 3 , wherein only one descriptive vector is translated from the first and second visual clues for each of the respective first and second sequences of characters.

6. The method of claim 3 , wherein the first and second types of characters are alphanumeric symbols, numbers, punctuations, scientific operators, non-Latin scripts, and Unicode symbols.

7. The method of claim 1 , wherein:

the first sequence of characters comprises alphanumeric characters;

the first subset comprises alphabetic characters only;

the first symbol comprises a symbol representing an alphabetic character string;

the first type of character comprises alphabetic characters;

the second subset comprises numeric characters only;

the second symbol comprises a symbol representing a numeric character string;

the second type of character comprises numeric characters; and

the first arrangement of the first and second symbols is a summarized representation of the first sequence of characters.

8. The method of claim 7 , wherein

the first classification is a media asset title;

the second classification is a media asset release year; and

the second arrangement is a summarized representation of the second sequence of characters, and is different from the first arrangement.

9. The method of claim 1 , further comprising translating the first sequence and generating a descriptive vector, wherein the automatically associating first and second classifications is based on the descriptive vector and the user input.

10. The method of claim 9 , further comprising displaying a user guide based on the descriptive vector and user classifications of symbols.

11. The method of claim 9 , wherein the descriptive vector is in a table form.

12. A system for training a model using a reduced dataset, comprising computer-readable memory, a user input device, and control circuity configured to:

receive a first sequence of characters;

associate a first subset of the first sequence of characters with a first symbol, wherein the first subset corresponds to a first type of character;

associate a second subset of the first sequence of characters with a second symbol, wherein the second subset corresponds to a second type of character;

determine a first arrangement of the first and second symbols based on the first and second subsets within the first sequence of characters;

receive user input specifying a first classification for the first symbol and a second classification for the second symbol;

automatically associate the first and second classifications with a second sequence of characters that are in a second arrangement that matches the first arrangement;

feed the first and second sequences to the model as the reduced dataset; and

train the model with both the reduced dataset and the user specified first and second classifications.

13. The system of claim 12 , wherein the control circuitry is further configured to:

extract first textual information and first visual clues from the first sequence of characters;

translate the first visual clues to a descriptive vector associated with the first sequence of characters;

process the first textual information of the first sequence to associate the first subset with the first symbol and the second subset with the second symbol; and

extract second textual information and second visual clues from the second sequence of characters,

wherein training the model further comprises training the model with both the reduced dataset and the descriptive vector.

14. The system of claim 13 , wherein the first and second visual clues comprise text color, bold font, italics, indentations, spacing, graphics, icons, and glyphs.

15. The system of claim 13 , wherein the control circuitry translates only one descriptive vector from the first and second visual clues for each of the respective first and second sequences of characters.

16. The system of claim 12 , wherein the first and second types of characters are alphanumeric symbols, numbers, punctuations, scientific operators, non-Latin scripts, and Unicode symbols.

17. The system of claim 12 , wherein:

the first sequence of characters comprises alphanumeric characters;

the first subset comprises alphabetic characters only;

the first symbol comprises a symbol representing an alphabetic character string;

the first type of character comprises alphabetic characters;

the second subset comprises numeric characters only;

the second symbol comprises a symbol representing a numeric character string;

the second type of character comprises numeric characters; and

the first arrangement of the first and second symbols is a summarized representation of the first sequence of characters.

18. The system of claim 17 , wherein

the first classification is a media asset title;

the second classification is a media asset release year; and

the second arrangement is a summarized representation of the second sequence of characters, and is different from the first arrangement.

19. The system of claim 12 , further comprising translating the first sequence and generating a descriptive vector, wherein the automatically associating the first and second classifications is based on the descriptive vector and the user input.

20. The system of claim 19 , further comprising displaying a user guide based on the descriptive vector and user classifications of symbols.