System to extract information from documents

A method of training a system to extract information from documents, the method comprising:

• feeding training documents in digital form to a layout preserving Optical Character Recognition (OCR) module;
• identifying, by the OCR module, within each of documents, multiple logical blocks in the document;
• identifying, by the OCR module, text present in each of the identified blocks;
• receiving, by a tagging module, one or more tags for the whole of the document, one or more tags for one or more of the logical blocks and one or more tags for one or more word tokens on the document;
• receiving, by a machine learning module, a text input comprising the text identified in the document and the tags for the whole of the document;
• and
• receiving, by the machine learning module, a first image of the document with layout of the one or more of the identified blocks superimposed and the tags for the logical blocks in the document, wherein the machine learning module receives the text input and the first image and the tags for the logical blocks corresponding to a plurality of the training documents.