System and method for machine-readable electronic document

A method for converting a first document including formatting cues designed for a human reader to a machine-readable electronic document, the method comprising:

• receiving by one or more processors, via a user interface (UI), user input to the first document towards creating the machine-readable electronic document;
• executing a neural network to analyze the user input to generate the machine-readable electronic document, wherein the neural network is trained using training dataset by iteratively adjusting at least one parameter of the neural network during training, wherein the neural network comprises an input layer, at least one mid layer, and an output layer, and wherein the analyzing includes the one or more processors:
• determining, from the user input, first data representing a structure of the machine-readable electronic document;
• determining, from the user input, second data representing semantics of the machine-readable electronic document;
• and
• determining, from the user input, third data representing a context of the machine-readable electronic document;
• labeling the first data, the second data, and the third data using metadata and storing the metadata in a metadata file associated with machine-readable electronic document, wherein the machine-readable electronic document is in a data-interchange format;
• responsive to determining the second data representing semantics, calculating, using a semantic relation map that includes a network of expressions being connected based on their meanings, semantic similarity scores measuring similarity between the second data and each of a plurality of linguistic expressions stored in a semantic database and comparing the semantic similarity scores with a threshold to determine a set of linguistic expressions that are similar to the second data;
• determining a class for the machine-readable electronic document based on a feature associated with the machine-readable electronic document;
• calculating a corresponding class participation rate for each one in the set of linguistic expressions in the class;
• determining an annotation for the machine-readable electronic document based on the class participation rates, the annotation including suggested replacement expressions as an addition to the machine-readable electronic document to a user, wherein the suggested replacement expressions are selected from the set of linguistic expressions and are associated with corresponding class participation rates within a range of the class participation rate associated with the second data;
• responsive to identifying the annotation, presenting the annotation on the UI, wherein the annotation includes at least one of text data, audio data, and image data to be stored in a machine-readable format;
• and
• responsive to identifying a user selection of a first replacement expression on the UI, adding the first replacement expression to the machine-readable electronic document;
• generating the machine-readable electronic document, by the one or more processors, based on the first data, the second data, and the third data and providing the machine-readable electronic document and the associated metadata file in the data-interchange format to a downstream software application for filtering, classification and ranking;
• receiving a request from the downstream software application for the machine-readable electronic document, wherein the request comprises a profile including a privilege level associated with a user;
• generating, based on the privilege level associated with the user, a version of the machine-readable electronic document, wherein the version of the machine-readable electronic document is one of a complete version of the machine-readable electronic document or a customized version of the machine-readable electronic document;
• and
• providing the determined version of the machine-readable document to the downstream software application.

A machine-readable non-transitory storage media encoded with instructions that, when executed by one or more processing devices, cause the one or more processing devices to perform operations for converting a first document including formatting cues designed for a human reader to a machine-readable electronic document, the one or more processing devices to:

• receive, via a user interface (UI), user input to the first document towards creating the machine-readable electronic document;
• execute a neural network to analyze the user input to generate the machine-readable electronic document, wherein the neural network is trained using training dataset by iteratively adjusting at least one parameter of the neural network during training, wherein the neural network comprises an input layer, at least one mid layer, and an output layer, and wherein to analyze the user input, the one or more processors are further to:
• determine, from the user input, first data representing a structure of the machine-readable electronic document;
• determine, from the user input, second data representing semantics of the machine-readable electronic document;
• and
• determine, from the user input, third data representing a context of the machine-readable electronic document;
• label the first data, the second data, and the third data using metadata and store the metadata in a metadata file associated with machine-readable electronic document, wherein the machine-readable electronic document is in a data-interchange format;
• responsive to determining the second data representing semantics, calculate, using a semantic relation map that includes a network of expressions being connected based on their meanings, semantic similarity scores measuring similarity between the second data and each of a plurality of linguistic expressions stored in a semantic database and compare the semantic similarity scores with a threshold to determine a set of linguistic expressions that are similar to the second data;
• determine a class for the machine-readable electronic document based on a feature associated with the machine-readable electronic document;
• calculate a corresponding class participation rate for each one in the set of linguistic expressions in the class;
• determine an annotation for the machine-readable electronic document based on the class participation rates, the annotation including suggested replacement expressions as an addition to the machine-readable electronic document to a user, wherein the suggested replacement expressions are selected from the set of linguistic expressions and are associated with corresponding class participation rates within a range of the class participation rate associated with the second data;
• responsive to identifying the annotation, present the annotation on the UI, wherein the annotation includes at least one of text data, audio data, and image data to be stored in a machine-readable format;
• and
• responsive to identifying a user selection of a first replacement expression on the UI, add the first replacement expression to the machine-readable electronic document;
• and
• generate the machine-readable electronic document based on the first data, the second data, and the third data and provide the machine-readable electronic document and the associated metadata file in the data-interchange format to a downstream software application for filtering, classification and ranking;
• receiving a request from the downstream software application for the machine-readable electronic document, wherein the request comprises a profile including a privilege level associated with a user;
• generate, based on the privilege level associated with the user, a version of the machine-readable electronic document, wherein the version of the machine-readable electronic document is one of a complete version of the machine-readable electronic document or a customized version of the machine-readable electronic document;
• and
• provide the determined version of the machine-readable document to the downstream software application.