Table item information extraction with continuous machine learning through local and global models

A system, comprising:

• a processor;
• a non-transitory computer-readable medium;
• and
• stored instructions translatable by the processor for:
• displaying a user interface having a database table and an image, the database table having a plurality of columns, the image containing a table, the user interface including a user interface element for activating a table auto-completion algorithm;
• responsive to activation of the table auto-completion algorithm through the user interface element, performing:
• determining coordinates from a portion of the table highlighted on the image displayed through the user interface;
• determining data points using the coordinates;
• automatically extracting the data points from the portion of the table;
• storing information about the data points in a machine learning model as positive examples for on-the-fly training of the machine learning model to recognize a table layout;
• determining, based on the portion of the table, a next portion of the table;
• determining additional data points in the next portion of the table;
• automatically extracting the additional data points from the next portion of the table;
• receiving or obtaining, through the user interface, a correction to the additional data points automatically extracted from the next portion of the table;
• storing information about the correction in the machine learning model as a negative example for the on-the-fly training of the machine learning model;
• and
• automatically continuously extracting table information from the table utilizing the table layout thus recognized, the positive examples, and the negative example until extraction of the table information from the table is completed and the table information extracted from the table is entered into the database table.

A computer program product comprising a non-transitory computer-readable medium storing instructions translatable by a processor for:

• displaying a user interface having a database table and an image, the database table having a plurality of columns, the image containing a table, the user interface including a user interface element for activating a table auto-completion algorithm;
• responsive to activation of the table auto-completion algorithm through the user interface element, performing:
• determining coordinates from a portion of the table highlighted on the image displayed through the user interface;
• determining data points using the coordinates;
• automatically extracting the data points from the portion of the table;
• storing information about the data points in a machine learning model as positive examples for on-the-fly training of the machine learning model to recognize a table layout;
• determining, based on the portion of the table, a next portion of the table;
• determining additional data points in the next portion of the table;
• automatically extracting the additional data points from the next portion of the table;
• receiving or obtaining, through the user interface, a correction to the additional data points automatically extracted from the next portion of the table;
• storing information about the correction in the machine learning model as a negative example for the on-the-fly training of the machine learning model;
• and
• automatically continuously extracting table information from the table utilizing the table layout thus recognized, the positive examples, and the negative example until extraction of the table information from the table is completed and the table information extracted from the table is entered into the database table.