FIELD OF THE INVENTION:

The present invention is related to a system and method of processing a data in real-time. Specifically, the present invention is related to a system and method of processing, in real-time, a data of a document to provide translation of one natural language to the other simplified natural language along with providing a clear and concise summary based on the translated information of the document.

BACKGROUND OF THE INVENTION:
Translation of language documents are generally provided from one language to the other language. The translation of the text in a document is done in multiple ways. However, accuracy of a translated document by changing the form of the text in the document without changing the intent is generally remains very low. For example, mostly the legal case judgements are published in English legal language and composed of several complex Legal-English keywords which are only meaningful for the legal professionals or the persons related from legal domain. However, a common person who is even associated with a legal case had no idea of complex legal keywords used in the legal document which finally creates misunderstanding and misinterpretation of the judgement. Further, the existing systems and methods fails to provide concise summary of the legal judgement documents in simple English language or the other desired language set by a user.

The existing systems and methods are lacking in integration of natural language processing with artificial intelligence wherein the existing methods are nowhere describes producing a meaningful document by determining the perfect synonyms along with sentiments of the complex legal or English words without deviating from the text of the original document. The existing systems and methods are only providing translation services wherein a cognitive intent is not being determined, thereby reduces the accuracy of the translated document. Further, the translation of the legal and English words simultaneously in order to provide more concise summary has not been disclosed by any existing system and/or method.

Therefore, there was a need for a method and system of processing a data wherein by determining the cognitive intent and entities, a translation as well as summary should be generated which may comprise of entirely different combination of text while retaining the initial cognitive intent of the text.

Further, there was a need to provide a summary of the text present in a document, which should focus on the most relevant facts first rather than emphasizing on each and every text. Additionally, there was a need for a system and method to provide a meaningful concise summary by focussing on the relevant and important text of the document.

In order to mitigate the drawbacks of the existing system and methods, as mentioned above, the present invention is aimed to provide a technical solution to the technical problem of “How to provide a meaningful translation and concise summary based on the cognitive intent and entities present in a document, without changing the initial intent of the document and by determining the relationship in between the text of a sentence of a paragraph of the document”.
The present invention not only provides a technical solution rather it provides an inventive technical solution wherein the system and method of the present invention creating a revised document, wherein the revised sentence is a combination of cognitive alternatives of extracted legal and grammar text as present in the document. By combining the cognitive alternatives of the extracted legal and grammar text, the claimed invention is able to perform translation in a way that the cognitive alternatives not only produces meaning sentence but a concise sentence also by which a person can save time and resources for processing the revised document.

SUMMARY OF THE INVENTION:

The following presents a simplified summary of the subject matter in order to provide a basic
understanding of some of the aspects of subject matter embodiments. This summary is not an
extensive overview of the subject matter. It is not intended to identify key/critical elements of the embodiments or to delineate the scope of the subject matter. Its sole purpose to present some concepts of the subject matter in a simplified form as a prelude to the more detailed description that is presented later.

In an embodiment, a method of processing plurality of data present in a document is described wherein the method identifying one or more paragraphs in the document; extracting each one of one or more sentences present in the plurality of paragraphs of the document; tokenizing each one of one or more words of each one of one or more sentences present in the plurality of paragraphs of the document, wherein the tokenizing is based on identification of natural language, grammar text and legal text; creating a Combination of plurality of tokenized words of each of one or more words present in the plurality of paragraphs of the document, wherein combining plurality of tokenized words are based on grammar knowledge of the plurality of tokenized words; tagging the one of one or more tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging; tagging the one of one or more combined tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging; extracting grammar text from the tagged tokenized words; extracting legal text from the tagged tokenized words; identifying a cognitive intent of the plurality of tokenized paragraphs, wherein the identifying the cognitive intent of the tokenized paragraph includes the determination of relationship in between tagged tokenized grammar text and tagged tokenized legal text; storing the extracted grammar text in a grammar dataset and extracted legal text in a legal dataset; mapping tagged tokenized grammar text with the text stored in the grammar dataset for creating a processed grammar text and mapping tagged tokenized legal text with the text stored in the legal dataset for creating a processed legal text; creating a revised sentence by combining the processed grammar text and processed legal text; extracting intents and entities from the revised sentence; applying weight vectors to the plurality of words of the revised sentence; determining the cognitive context and intent of the revised sentence; generating a summary by combining plurality of revised sentences.

In an embodiment, the method of the present invention determining the cognitive context and intent of the revised sentence by computing the relationship in between the plurality of sentences by a computational model wherein an entity, intent and context is being determined and classified as per the computed relationship between the plurality of revised sentences, wherein the part of speech tagging is applied to a part of speech tagged database to produce a part of speech tagged sentence.

In an embodiment, the method of the present invention extracting the grammar text from the tokenized paragraph by identifying nouns, pronouns, adjectives, adverbs, verbs, type of tenses, prepositions, conjunctions.

In an embodiment, the method of the present invention mapping the tagged tokenized grammar text with the text stored in the grammar dataset by evaluating location of grammatical text of the processed grammar text with the tagged tokenized legal text and by comparing the grammatical text of the processed grammar text with the grammar knowledge of the plurality of tokenized words. The grammar dataset is provided with cognitive alternatives of tagged tokenized grammar text and the legal dataset is provided with cognitive alternatives of the tagged tokenized legal text.

In an embodiment, the method of the present invention combining the processed grammar text and processed legal text by evaluating that the combination of the processed grammar text with the grammar knowledge of the plurality of tokenized words and the combination of processed grammar text with the processed legal text matches the cognitive intent of the plurality of tokenized paragraphs of the document.

In an embodiment, the method of the present invention tagging by preparing a corpus wherein a raw data is collected, prepared and labelled based on annotation guidelines, wherein the annotation guidelines comprise evaluation of cognitive intent and cognitive entity of the one or more combination of words of the plurality of sentences of the one or more paragraphs of the document.

In an embodiment, the method of the present invention performs tagging by updating and training the corpus with the evaluated cognitive intent and with the cognitive entity of the one or more combination of words of the plurality of sentences of the one or more paragraphs of the document.

In an embodiment, the method of the present invention performs the tagging by labelling the one or more words of the plurality of sentences of the one or more paragraphs of the document by combining a selected text of the corpus with the one or more words of the plurality of sentences of the one or more paragraphs wherein the selected text of the corpus is determined based on a query made by a machine learning technique processed on the combination of tokenized words.

In an embodiment, the machine learning technique of the present invention extracting one or more features of grammar text from the tokenized words and extracting one or more features of the legal text from the tokenized words.

In an embodiment, the machine learning technique of the present invention enables encoding and combining of the one or more extracted features of the grammar text and legal text.

In an embodiment, the machine learning technique evaluates a vector value of the extracted features of grammar text and of the extracted features of the legal text by combining the encoded features of the one or more extracted features of the grammar text and legal text.

In an embodiment, the machine learning technique enables decoding the combined encoded one or more extracted features of the grammar text and legal text.

In an embodiment, the method of the present invention performs the machine learning technique to produce an encoded map vector based on the evaluated vector values of the encoded features of the one or more extracted features of the grammar text and legal text, and wherein the encoded map vector represents the relationship in between the features of the grammar text and features of the legal text.

In an embodiment, the method of the present invention creating the revised sentence by determining complexity of the revised sentence with at least one sentence of the document, wherein upon determining the complexity of the revised sentence being high, the revised sentence is paraphrased with the grammar text and legal text of the corpus to achieve least complexity of the revised sentence by comparing with the sentence of the document.

In an embodiment, the method of the present invention creating a processed grammar text and creating a processed legal text comprises providing cognitive alternatives to the tagged tokenized grammar text and tagged tokenized legal text respectively, wherein the cognitive alternatives of the grammar text and legal text are synonyms of the tagged tokenized grammar text and tagged tokenized legal text respectively, wherein providing the cognitive alternatives to the tagged tokenized grammar text and tagged tokenized legal text are based on determination and selection of synonyms of the grammar text and legal text respectively, of the sentence.

In an embodiment, the method of the present invention creating the processed grammar text and processed legal text by identifying synonyms, abbreviation, active form and passive form of the one or more words present in the sentence; applying grammar rules to the evaluated location of the grammatical text as per the determined location of the grammatical text of the sentence; wherein the abbreviation is identified by performing clipping the tagged tokenized words wherein the cognitive intent of the tagged tokenized words is identical; contracting the tagged tokenized words; creating acronym of the tagged tokenized words, wherein the clipping is performed by mapping the one or more words of the sentence by comparing the one or more words with the words present in the grammar dataset and legal dataset and by evaluating the order of the grammatical and legal text of the sentence.

In an embodiment, the method of the present invention determining the cognitive context and intent of the revised sentence comprises calculating the similarity in between the revised sentence and sentence of the one or more paragraphs of the document, wherein the calculating the similarity in between the revised sentence and sentence of the one or more paragraphs comprising, matching of linguistic knowledge and encyclopaedia knowledge of natural language and with encyclopaedia knowledge of legal text, wherein by calculating similarity in between the revised sentence and sentence of the one or more paragraphs comprising assigning decimal vector values, wherein the least decimal value indicating similarity in between the revised sentence and sentence of the one or more paragraphs.

In an embodiment, the method of the present invention generating the summary comprises by identifying parameters and sequence of the one or more words of the revised sentence of the one or more paragraphs of the document; creating a pre-trained model with the identified parameters and sequence of the one or more words of the revised sentence of the one or more paragraphs of the document wherein the pre-trained model process the cognitive intent and context of the revised sentence by identifying the existed identified parameters and sequence in the model.

In an embodiment, the method of the present invention generating the summary by identifying legal text parameters and sequence in the one or more revised sentences of the one or more paragraphs of the document; creating a pre-trained model with the identified legal text parameters and sequence of the revised sentence of the one or more paragraphs of the document wherein the pre-trained model process the cognitive intent and context of the revised sentence by identifying the existed identified legal text parameters and sequence in the model.

In an embodiment, the method of the present invention generating the summary by identifying plurality of legal text present in the one or more revised sentences of the one or more paragraphs of the document, wherein the plurality of legal text is a selection of: prime facts, prime evidences, prime arguments; creating a pre-trained model with the identified plurality of legal text.

In an embodiment, the method of the present invention applying weight vectors to the plurality of words of the revised sentence comprises ranking the revised sentences based on highest weight vector value by combining the one or more words of the revised sentence; creating a parse tree of the revised sentence based on the context and grammar knowledge of the revised sentence; identifying an important context of the revised sentence based on the parse tree value;
Storing the parse tree value along with the important context of the revised sentence; ranking the important context of the revised sentence based on the parse tree value; presenting the highest ranking important context plurality of words of the revised sentence.

In an embodiment, the method of the present invention applying the cognitive alternative to the tagged tokenized grammar and legal words of the sentence of the paragraph comprises applying cognitive alternatives to the immediate one of one or more grammar and legal text of the sentence wherein applying the cognitive alternative to the tagged tokenized grammar and legal words tokenized words of the sentence of the paragraph comprises applying cognitive alternatives to the later one of one or more text of the sentence, wherein the cognitive alternatives of the grammar text and legal text are ranked based on the determination of the alternative cognitive intent of the tokenized words of the sentence of the paragraph and wherein the cognitive alternatives of the grammar text and legal text with higher ranking are formatted and stored in a cognitive database, wherein the formatted cognitive alternatives of the grammar text and legal text with higher ranking are applied more weightage.

In an embodiment, the method of the present invention creating the revised sentence by mapping formatted cognitive alternatives of the grammar text with the formatted cognitive alternatives of the legal text with more weightage.

In an embodiment, the method of the present invention creating the revised sentence by mapping formatted cognitive alternatives of the grammar text of more weightage with the tokenized legal text and simultaneously mapping formatted cognitive alternatives of the legal text of more weightage with the tokenized grammar text.

In an another embodiment of the present invention there is provided a system for processing plurality of data present in a document. The system identifying, by an identifier, one or more paragraphs in the document; extracting, by an extractor, each one of one or more sentences present in the paragraph of the document; tokenizing, by a tokenizer, each one of one or more words of each one of one or more sentences present in the plurality of paragraphs of the document, wherein the tokenizing is based on identification of grammar text and legal text; creating, by a pre-processor, a Combination of plurality of tokenized words of each of one or more words present in the plurality of paragraphs of the document, wherein the combining plurality of tokenized words are based on grammar knowledge of the plurality of tokenized words; tagging, by an AI model, the one of one or more tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging; tagging, by the AI model, the one of one or more combined tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging; extracting, by a processor, grammar text from the tagged tokenized words; extracting, by the processor, legal text from the tagged tokenized words; identifying, by a computing module, a cognitive intent and cognitive context of the plurality of tokenized paragraphs, wherein the identifying the cognitive intent of the tokenized paragraph includes the determination of relationship in between tagged tokenized grammar text and tagged tokenized legal text; storing, in a database, the extracted grammar text in a grammar dataset and extracted legal text in a legal dataset; mapping, by a mapping module, the tagged tokenized grammar text with the text stored in the grammar dataset for creating a processed grammar text and tagged tokenized legal text with the text stored in the legal dataset for creating a processed legal text; creating, by a generating tool, a revised sentence by combining the processed grammar text and processed legal text; extracting, by the processor, intents and entities from the revised sentence; applying, by the processor, weight vectors to the plurality of words of the revised sentence; determining, by the processor, the cognitive context and intent of the revised sentence; generating, by the generating tool, a summary by combining plurality of revised sentences.

In an embodiment of the present invention, the processor determining the cognitive context and intent of the revised sentence by computing the relationship in between the plurality of sentences by a computational model wherein an entity, intent and context is being determined and classified as per the computed relationship between the plurality of revised sentences.

In an embodiment of the present invention, the mapping module mapping the tagged tokenized grammar text with the text stored in the grammar dataset by evaluating location of grammatical text of the processed grammar text with the tagged tokenized legal text and by comparing the grammatical text of the processed grammar text with the grammar knowledge of the plurality of tokenized words wherein the grammar dataset is provided with cognitive alternatives of tagged tokenized grammar text and the legal dataset is provided with cognitive alternatives of the tagged tokenized legal text.

In an embodiment of the present invention, the generating tool combining the processed grammar text and processed legal text by evaluating that the combination of the processed grammar text with the grammar knowledge of the plurality of tokenized words and the combination of processed grammar text with the processed legal text matches the cognitive intent of the plurality of tokenized paragraphs of the document.

In an embodiment of the present invention, the AI model performs tagging by preparing a corpus wherein a raw data is collected, prepared and labelled based on annotation guidelines wherein the annotation guidelines comprise evaluation of cognitive intent and cognitive entity of the one or more combination of words of the plurality of sentences of the one or more paragraphs of the document.

In an embodiment of the present invention, the AI model performs tagging by updating and training the corpus with the evaluated cognitive intent and with the cognitive entity of the one or more combination of words of the plurality of sentences of the one or more paragraphs of the document.

In an embodiment of the present invention, the AI model performs tagging by labelling the one or more words of the plurality of sentences of the one or more paragraphs of the document by combining a selected text of the corpus with the one or more words of the plurality of sentences of the one or more paragraphs wherein the selected text of the corpus is determined based on a query made by a machine learning technique processed on the combination of tokenized words.

In an embodiment of the present invention, the AI model performs the machine learning technique by extracting one or more features of grammar text from the tokenized words and extracting one or more features of the legal text from the tokenized words.

In an embodiment of the present invention, the AI model performs the machine learning technique by encoding and combining the one or more extracted features of the grammar text and legal text.

In an embodiment of the present invention, the AI model performs the machine learning technique by evaluating a vector value of the extracted features of grammar text and of the extracted features of the legal text by combining the encoded features of the one or more extracted features of the grammar text and legal text.

In an embodiment of the present invention, the AI model performs the machine learning technique by decoding the combined encoded one or more extracted features of the grammar text and legal text.

In an embodiment of the present invention, the AI model performs the machine learning technique to produce an encoded map vector based on the evaluated vector values of the encoded features of the one or more extracted features of the grammar text and legal text wherein the encoded map vector represents the relationship in between the features of the grammar text and features of the legal text.

In an embodiment of the present invention, the system is creating the revised sentence by determining complexity of the revised sentence with at least one sentence of the document, wherein upon determining the complexity of the revised sentence being high, the revised sentence is paraphrased with the grammar text and legal text of the corpus to achieve least complexity of the revised sentence by comparing with the sentence of the document.

In an embodiment of the present invention, the system of the present invention creating the processed grammar text and creating a processed legal text by providing cognitive alternatives to the grammar text and legal text respectively.

In an embodiment of the present invention, the system of the present invention providing the cognitive alternatives to the grammar text and legal text are based on determination and selection of synonyms of the grammar text and legal text of the sentence, wherein the cognitive alternatives of the grammar text and legal text are synonyms of the grammar text and legal text respectively.

In an embodiment of the present invention, the system of the present invention creating a processed grammar text and processed legal text by identifying synonyms, abbreviation, active form and passive form of the one or more words present in the sentence; applying grammar rules to the evaluated location of the grammatical text as per the determined location of the grammatical text of the sentence, wherein the abbreviation is identified by performing clipping the tagged tokenized words wherein the cognitive intent of the tagged tokenized words is identical; contracting the tagged tokenized words; and by creating acronym of the tagged tokenized words.

In an embodiment of the present invention, the system of the present invention performing clipping by mapping the one or more words of the sentence by comparing the one or more words with the words present in the grammar dataset and legal dataset and by evaluating the order of the grammatical and legal text of the sentence.

In an embodiment of the present invention, the system of the present invention determining the cognitive context and intent of the revised sentence by calculating the similarity in between the revised sentence and sentence of the one or more paragraphs of the document, wherein the calculating the similarity in between the revised sentence and sentence of the one or more paragraphs comprising, matching of linguistic knowledge and encyclopaedia knowledge of natural language and with encyclopaedia knowledge of legal text and wherein by calculating similarity in between the revised sentence and sentence of the one or more paragraphs comprising assigning decimal vector values, wherein the least decimal value indicating similarity in between the revised sentence and sentence of the one or more paragraphs.

In an embodiment of the present invention, the generating tool generating the summary by identifying parameters and sequence of the one or more words of the revised sentence of the one or more paragraphs of the document; creating a pre-trained model with the identified parameters and sequence of the one or more words of the revised sentence of the one or more paragraphs of the document wherein the pre-trained model process the cognitive intent and context of the revised sentence by identifying the existed identified parameters and sequence in the model.

In an embodiment of the present invention, the generating tool generating the summary by identifying legal text parameters and sequence in the one or more revised sentences of the one or more paragraphs of the document; creating a pre-trained model with the identified legal text parameters and sequence of the revised sentence of the one or more paragraphs of the document wherein the pre-trained model process the cognitive intent and context of the revised sentence by identifying the existed identified legal text parameters and sequence in the model.

In an embodiment of the present invention, the generating tool generating the summary by identifying plurality of legal text present in the one or more revised sentences of the one or more paragraphs of the document, wherein the plurality of legal text is a selection of: prime facts, prime evidences, prime arguments; creating a pre-trained model with the identified plurality of legal text.

In an embodiment of the present invention, the processor applying weight vectors to the plurality of words of the revised sentence by ranking the revised sentences based on highest weight vector value by combining the one or more words of the revised sentence; creating a parse tree of the revised sentence based on the context and grammar knowledge of the revised sentence; identifying an important context of the revised sentence based on the parse tree value; storing the parse tree value along with the important context of the revised sentence; ranking the important context of the revised sentence based on the parse tree value; presenting the highest ranking important context plurality of words of the revised sentence.

BRIEF DESCRIPTION OF DRAWINGS:
The figures as flowcharts are the block diagrams of the preferred embodiment of the current submitted invention. The following drawings are illustrative of particular examples for enabling systems and methods of the present disclosure, are descriptive of some of the methods and mechanism, and are not intended to limit the scope of the invention. The drawings are not to scale (unless so stated) and are intended for use in conjunction with the explanations in the following detailed description.
Figure 1 is a block diagram of the proposed system in which automated language translation system and finally the summary of the source input text will be generated at last.
Figure 2 is the internal system in which key sections of the proposed system is displayed.
Figure 3 is the flowchart of the internal system in which general idea of translation is displayed.
Figure 4 is the block diagram displaying the summarization of the key steps of the proposed system
Figure 5 is the block diagram of part of speech tagging showing various tags of the grammar
Figure 6 is the block diagram of the system with respect to the end user
Figure 7 is the block diagram of labelling the sentence
Figure 8 is the block diagram of the sentence creation with the help of new labelled tags and the grammar rules of natural language
Figure 9 is the flowchart of creation of new sentence from various previous information of original sentence and tagged information of legal and language words.
Figure 10 is the generation of annotated corpus process for labelling of input data
Figure 11 is the flow diagram showing machine learning model processing on various inputs
Figure 12 is the internal process of the transformer model
Figure 13 is the flow diagram of the paraphrase generation of complex legal sentence in normalized form
Figure 14 is the flow diagram of swapping word with its synonym
Figure 15 is the flow diagram of how to perform swapping of word with its synonym
Figure 16 is the process of paraphrasing of a sentence in normalized form
Figure 17 is the flow diagram of swapping various words with its abbreviations
Figure 18 is the flow diagram of clipping various words with its shorter form
Figure 19 is the flow diagram of performing active and passive word translation
Figure 20 is the flow diagram of checking similarity value of original and paraphrased sentences
Figure 21 is the block diagram showing the internal process of the assigning weights to the legal sentences inside a transformer
Figure 22 is the block diagram showing attributes needed to check similarity based on the information combined
Figure 23 is the process of segmentation of original sentence and its paraphrase sentence
Figure 24 is the process of summary generation of a legal document in various ways
Figure 25 is a system diagram of the invention
DETAILED DESCRIPTION:
Exemplary embodiments now will be described with reference to the accompanying figures. The exposure may, however, be embodied in numerous different forms and shouldn't be illustrated as limited to the embodiments set forth herein; rather, these figures are provided so that this description will be thorough and complete, and will thoroughly convey its scope to those proficient in the art. The terminology used in the detailed description of the particular exemplary figures illustrated in the accompanying delineations isn't intended to be limiting.

It's to be noted, still, that the reference numerals in claims illustrate only typical embodiments of the present subject matter, and are thus, not to be considered for limiting of its scope, for the subject matter may admit to other correspondingly effective embodiments.

The specification may relate to “an”, “one “or “some” embodiment(s) in several places. This doesn't inevitably indicate that each cognate reference is to the same embodiment(s), or that the feature only applies to a single embodiment.

As used herein, the singular forms “a”, “an” and “the” are intended to include the plural forms as well, unless expressly stated else. It'll be further understood that the terms “includes”, “comprises”, “including” and/ or “comprising” when used in this specification, specify the presence of stated features, integers, way, operations, elements, and/ or components, but don't obviate the presence or addition of one or additional other features, integers, steps, operations, elements, constituents, and/ or groups thereof.

It'll be understood that when an element is appertained to as being “connected” or “coupled” to another element, it can be directly connected or coupled to the other element or intermediating elements may be present. likewise, “connected” or “coupled” as used herein may include operatively connected or coupled. As used herein, the term “and/ or” includes any and all combinations and arrangements of one or further of the associated listed particulars. Unless else defined, all terms (including specialized and scientific terms) used herein have the same meaning as generally understood by one of ordinary skill in the art to which this exposure pertains. It'll be further understood that terms, similar as those defined in generally used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains.

The functions described herein may be enforced in hardware, software executed by a processor, firmware, or any combination thereof. However, the functions may be stored on or transmitted over as one or additional instructions or code on a computer- readable medium, if implemented in software executed by a processor. Other exemplifications and executions are within the scope of the description and adjoined claims. For illustration, due to the nature of software, functions described herein may be implemented using software executed by a processor, hardware, firmware, hardwiring, or combinations of any of these. Features implementing functions may also be physically located at multi positions, including being distributed such that portions of functions are implemented at different physical locations.
It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein. The figures depict a simplified structure only showing some elements and functional entities, all being logical units whose implementation may differ from what is shown. The connections shown are logical connections; the actual physical connections may be different. It is apparent to a person skilled in the art that the structure may also comprise other functions and structures.

Computer-readable media includes both non-transitory computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A non-transitory storage medium may be any available medium that may be accessed by a general-purpose or special-purpose computer.

As used herein, including in the claims, "or" as used in a list of items (e.g., a list of items prefaced by a phrase such as "at least one of or "one or more of) indicates an inclusive list such that, for example, a list of at least one of X, Y, or Z means X or Y or Z or XY or XZ or YZ or XYZ (i.e., X and Y and Z). Also, as used herein, the phrase "based on" shall not be construed as a reference to a closed set of conditions. For example, an example step that is described as "based on condition X" may be based on both a condition X and a condition Y without departing from the scope of the present disclosure. In other words, as used herein, the phrase "based on" shall be construed in the same manner as the phrase "based at least in part on."
In the appended figures, similar components or features may have the same reference label. Further, various components of the same type may be distinguished by following the reference label by a dash and a second label that distinguishes among the similar components. If just the first reference label is used in the specification, the description is applicable to any one of the similar components having the same first reference label irrespective of the second reference label, or other subsequent reference label.
The description set forth herein, in connection with the appended figures, describes exemplar configurations and doesn't represent all the illustrations that may be executed or that are within the scope of the claims. The term" example" used herein means" serving as an example, case, or illustration," and not" preferred" or" beneficial over other examples." The detailed description includes specific details for the purpose of providing an understanding of the described techniques. These techniques, still, may be exercised without these specific details. In some cases, known structures and devices are shown in block figure form in order to avoid obscuring the conceptions of the described examples.
The description herein is provided to enable a person having ordinary skill in the art to make or use the description. Various variations to the disclosure will be apparent to a person having ordinary skill in the art, and the general principles defined herein may be applied to other variations without departing from the scope of the description. accordingly, the description isn't limited to the illustrations and designs described herein, but is to be accorded the broadest scope consistent with the principles and new features disclosed herein.

In an embodiment of the present invention as shown in Figure. 1, a method of processing a data, in real-time, is described.

As shown, at step 101, the method as implemented by a system 2500, identifies one or more paragraphs in a document. The proposed system will take input text from the legal judgment in particular natural language. At step 101 only the each one or more sentences present in the plurality of paragraphs of the document is being extracted. At step 102, the method of the present invention performs tokenization process wherein tokenization is performed at sentence wise of the document. It should be noted that during the tokenization process the each one or more words of the each one or more sentences of the document are tokenized on the basis of the identification of natural language, grammar text and legal text. At step 103, the method of the present invention creating a combination of plurality of tokenized words wherein the combination of plurality of tokenized words are based on grammar knowledge of the plurality of tokenized words. At step 104, the process of tagging is performed. The tagging is performed by way of part of speech tagging. Further, tagging is also performed on the one or more combined tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging. In the tagging process, the grammar text is extracted from the tagged tokenized words and simultaneously legal text is also being extracted from the tagged tokenized words. A person skilled in the art should appreciate that in the tagging process a cognitive intent of the plurality of tokenized paragraphs is also being identified, wherein the method of identification of the cognitive intent of the tokenized paragraph may include the determination of relationship in between tagged tokenized grammar text and tagged tokenized legal text. Further, upon extraction process the grammar text is stored in a grammar dataset and extracted legal text is being stored in a legal dataset. At step 105, mapping is performed wherein the tagged tokenized grammar text is mapped with the text stored in the grammar dataset for creating a processed grammar text and mapping tagged tokenized legal text with the text stored in the legal dataset for creating a processed legal text. Upon completion of the mapping process, at step 106, a revised sentence is created wherein the revised sentence is a combination of the processed grammar text and processed legal text. At step 107, the intents and entities from the revised sentence is again extracted wherein at step 108, weight vectors are applied to the plurality of words of the revised sentence. It should be noted that the purpose of identifying the intents entities and applying weight vectors is to determine the revised sentence meaning in terms of why, where, what, which, when and how, which can be helpful in understanding the main idea behind that sentence or a paragraph. Further, the weighted graph made from the weigh vectors are able to fetch the main ideas and main points from the paragraphs of the document. Further, the present invention also able to find the internal patters, in which how the main ideas are related to each other in a complete document. At step 109, after determination of the cognitive context and intent of the revised sentence, a final tree is also created in which how the information is conveyed from one node to another node is indicated on the basis of which at step 109, a summary based on the information is generated which may be dependent on one or more condition.
In an embodiment of the present invention, the figure 2 explaining the process of tokenization wherein it is shown that how entity and intent classification can be performed in the text sentence and in paragraphs of the document respectively.

At step 201, the sentence is presented and subsequently at step 202, the tokenization is performed on all words of the sentence wherein the words are tokenized in a sentence based on the natural language. In the next phase at step 203, with the use of annotated NER corpus part of speech tagging of incoming sentence is done. Entity in a sentence is detected, at step 204, with the help of internal computation model. Various relationships between various sentences which have entities dependencies are computed in the step 205. Finally, intent classification in a given sentence is computed, at step 206, based on the internal intent classification module.

In an embodiment of the present invention, the figure 3 explaining the general idea of processing the data of a document. Specifically, the figure 3 is explaining a translation process of the document which is being processed. display the process of converting input sentence into target sentence. The aim of the proposed method and system is to translate a document passage in the refined and normal form. At step 301, the input source text is first fed to the above system of the present invention. It should be noted that the translation is performed on paragraph basis. Therefore, the first paragraph is always taken into consideration first. Then subsequently the other paragraphs are being taken into consideration. Additionally, in the paragraph, each of the sentence of the corresponding paragraph is processed and may be translated into the target language. A person skilled in the art would appreciate that the target language refers to the revised sentences and revised paragraphs as produced by the present invention, wherein the revised sentences and revised paragraphs could be in same language or in other language. For the sake of clarity, the target language may also refer to paraphrased language wherein the simplest form of words is used in order to explain the document which is to be processed. Therefore, the target language may be combination of revised grammar text and revised legal text as generated by the present invention.
Once the translation of all the sentences of all the paragraphs are completed, the proposed invention checks, at step 305, for the next passage in the document. If any sentence is left out or pending for conversion, then the entire process from step 302 is repeated. After completing all the text of the document, the present invention, at step 306, outputs the new translated language to the system.
In an embodiment of the present invention, the figure 4 explaining that by which elements of the present invention the translation is being performed. The Fig. 4 is explaining the process of translation of complex legal natural language into more general natural language for the common person. A person skilled in the art would appreciate that the document which need to be processed could be scanned by a scanner, wherein the scanner could be a dedicated scanner or the scanning process could be performed by a mobile device by taking snapshot or by taking picture from a camera of a mobile device. Once the document is fetched from the scanner or from any input source, the document is presented for processing. After the input processing, the document is passed through the hardware circuitry. The document may have complex to normal legal language translation mechanism which may be applied on the text content of the document. After translating the input text, a summarizer may use its internal computation to perform the various computational techniques to get main idea of the document and combining the idea with various main points and generate distinct types of summaries for the end user. Finally, the translated document could be stored wherein the paraphrasing and as well as the summaries are documented for the user of the invention.

In an embodiment of the present invention, the figure 5 explaining that how the part of speech tagging is performed wherein different tags of grammar language is performed. As shown, a person skilled in the art would appreciate in order to perform the part of speech tagging, the input sentence is fed to the system of the present invention, for checking the tags of each word with the help of annotated sentence. As the annotated sentence has various tags information. Sentence is evaluated for the presence of various tag like present of noun, presence of pronoun, presence of adverb, presence of adjective, presence of verb, presence of preposition, presence of conjunctions and interjections in a sentence. Here, it should be noted that the grammar knowledge like the identification of the above elements are not limited and it could be performed on the other forms which may be combination of any elements in any form. Also, the present invention may check the tense of the verb in which sentence is written. In the end, finally tagged sentence can be fed to the input of next system of the present invention for further evaluation.

In an embodiment of the present invention, Fig 6 is an exemplary embodiment showing block diagram of the system with respect to an end user. As shown, the block diagram is used to display the process of converting input sentence into target sentence with respect to a perspective of the end user. End user may interact with the user interface where a computational process may be running on the system of the present invention. As shown, the internal working of the language translation is being explained and described, wherein the system of the present invention translates the legal passage in the refined and normal form. As shown and could be understood by a person skilled in the art that an input source text is first fed to the elements of the present invention, wherein first of all, first paragraph is taken into consideration. Each sentence of this paragraph will be translated into the target language. After completing translation of all the sentences of paragraph, the present invention checks for the next passage in the document. Finally, the proposed system, outputs the new translated language to the next system.
In an embodiment of the present invention figure 7 is the block diagram of explaining the labelling process for a sentence of the document. It should be noted that the labelling of the input text is based on the earlier annotated Named Entity Recognition dataset which has all the annotated data of full encyclopaedia of specific natural language. A complete paragraph from a single page is inputted to the above system. From a paragraph a single sentence is taken as the single unit for labelling process. Single sentence is then fed to the labelled system and with the help of annotated NER corpus all the words present in the input sentence are labelled with the seven types of part of speech tagging which are Noun, Pronoun, Adverb, Adjective, Preposition, Conjunction, Interjection. The labelled data may be further processed to create and to aid the processing of data of the document.
In an embodiment of the present invention figure 8 is showing the process of sentence creation with the help of new labelled tags and the grammar rules of natural language. As shown, the figure 8 is explaining the process of sentence creation for a particular natural language which may have knowledge of grammar of that particular language. To perform the sentence creation process, first of all words having tagged information are processed by the system of the present invention. In the next step, Natural language grammar system which have full understanding of the grammar of English language may be used to modify the sentence as per the syntax of the language. Finally, the sentence with same tag of words may be created based on the identification of the correct form syntax and semantics. A person skilled in the art should appreciate that the identification of the correct syntax and semantics is based on the natural language grammar database which is fed and updated with the knowledge of the updated grammar knowledge in real-time and such knowledge could be processed and used for future process of translation of the document. Further, the order of tagged words could be determined by the present invention by considering the cognitive context of the order of words.

In an embodiment of the present invention figure 9 is showing the diagram of creation of new sentence from various previous information of original sentence and tagged information of legal and language words. As indicated the process of creating a combined sentence from the respective two different parts of the sentence in natural language and the legal sentence in normalized form is explained herein. As shown the two parts of a sentence are joined together based on two conditions. First condition is joining general English natural language part of speech words with the general legal part of speech words based on the grammar knowledge of that specific natural language. Second condition, which is the most important, as it checks the joining of two part of speech tagged words of general natural language and legal do not break the structure of the original sentence. This step will generate the output sentence in synchronization with the input sentence. A person skilled in the art should appreciate that in case if a sentence is not having combination of legal and general natural language, then again the process as explained above, could be performed on based on first condition or could be moved to the second condition if required to be.

In an embodiment of the present invention, figure 10 is showing the process of generation of annotated corpus process for labelling of input data. The process of annotating words of a sentence which is very helpful in creating corpus of the legal as well as other words of the natural language. Main aim of the process is to set the system of the present invention for the labelling of the input sentence which can be annotated with the annotated corpus. As shown that the input to the system of the present invention, first of all raw data is collected. A person skilled in the art should appreciate the raw data is of two forms. Here the importance could be given to the legal sentences which have various words uttered in different ways. Second one may be the English natural language sentences which have various words of different complexities. It is important to note that the data is prepared in the structured manner having different sentences and the data could be stored in a text or a csv file. In the next step, the process of annotating the data with the help of annotation guidelines is shown, wherein it specifies how parts of speech are tagged in different entities. There could be seven parts of speech in which various words are tagged. In the next step, the labelling of the sentences are done step by step with words by words. If another new sentence appears in the labelling process, then it is updated and finally an updated corpus is ready to fed to the proposed system of the present invention wherein part of speech tagging of the input sentences of a textual paragraph is also performed.

In an embodiment of the present invention, figure 11 is showing the flow diagram wherein machine learning model processing on various inputs is shown. As per the flow diagram, the system of the present invention is used to locate the label as entity as well as the text classification based on the intent present in the text sentence. A person skilled in the art would appreciate that the word embeddings are the most authentic way to get the closeness of the words in a most closed space. The Word embeddings may be generated from large text corpus where vectors are present in the low dimensional space. Here the result of training, as shown, is that relation in semantic similarity or sentences which are nearby in context are very close together. Further, to find the intent of the word in a sentence, word embedding of the training data with known label are combined together to the machine learning techniques. Then the text may be classified with word embedding and may be inputted to the machine learning model and by which with the help of earlier knowledge, the system of the present invention is able to produce the intent specific to that label.

In an embodiment of the present invention, figure 12 is the internal process of mapping, wherein the system of the present invention is used to map the one type of information into different type of information. A given sentence may be composed of various words and these words can contain various types of entities and having some intent associated with each sentence. The function of single unit of encoder is to accept a single unit of the source sentence and gain insights from it and move it to next section. Number of hidden layers depends on the dimensions or the features of the data. In between, there is a combiner which acts as a mapping vector which contains mapped values between the input and outputs features. Function of next unit which is decoder is used to output feature based on the encoding vector values. So, the above system is used to produce an encoding map vector which is derived from the relation of input features of a sentence and the output features of a sentence.

In an embodiment of the present invention, figure 13 is the flow diagram of the paraphrase generation of complex legal sentence in normalized form. The paraphrasing of the input source NLP text and the output of the final text is performed after performing the similarity evaluation tests on the original sentence and the paraphrased sentence. Purpose of the paraphrase generator of the present invention is to generate standard natural language so that common people can get the ideas behind the text in normal form as compared to text in the original form. As shown, the method may check the complexity level of the original sentence and then paraphrase the sentence based on the complexity scores of the sentence. Complexity scores can vary from 0 to 1 based on the low and high complexity level of words in an original sentence. Number of words based on the complexity chart present in the dataset will be used for calculating the value of each word complexity and will be summed together. After checking the complexity of the system, if the complexity is high then the paraphrased generator may perform the generation of new sentence with the help of its internal computational techniques. Otherwise, system may not perform the paraphrasing of the sentence. The benefit of this complexity calculation system is that it may not perform the computation without any specific reason. After, the generation of the paraphrased sentence the similarity evaluation may be performed. Similarity evaluation can be performed based on the various semantic level checking. Euclidean distance principle as well as the correlation may be used for evaluating the distance between two sentences and the extent to which two variables are linearly related respectively. If similarity evaluation result is very low, then the paraphrasing output system may finally output the paraphrased sentence. Otherwise, next iteration will be performed with another set of words which will be in the dataset present in both the legal and as well as English natural language datasets. In the second iteration it may fetch the words which are high in similarity as compared to previous words which were used in the first iteration of the paraphrasing system. After performing this second iteration the output may be again evaluated for the similarity comparison. The system of the present invention may be paraphrased twice only. The output generated as paraphrased sentence can be further used as input for next level system.

In an embodiment of the present invention, figure 14 is showing the flow diagram of swapping word with its synonym. The purpose of the system of the present invention is to create an output word based on the synonym rule of the natural language. The logic behind using the synonym rule is that in certain scenarios the use of synonym helps to decrease the complexity of sentence and finally perform the normalized alteration of the word. Input word with tag and position of the word in the previous sentence may be taken into consideration. Word with tag is inputted into the synonym dataset system. This system in a specialized way, finds the synonym with the specific tag information corresponding to the original word with tag information. After finding the synonym information system may send the new word with tag information to the next output system. If the system is not able to find the synonym of the word, then the system may send the original word as it is to the output system for next level.

In an embodiment of the present invention, figure 15 is showing that how to perform swapping of word with its synonym. Here, as an input words with part of speech tag may be used to convert into the corresponding synonym with the original tag information. Internal system may have a dataset which may be composed of mapping of words with their specific synonyms with increasing order of complexities. Input POS tagged word is first submitted to the synonym generator system with word and tags as different inputs. Word with effective tag like adjective or adverb or other will be searched in the dataset and finds the corresponding synonym with specific tag information. After finding the synonym, the system will send the synonym word with tag information to the output module and finally new word with the same part of speech tag information will be sent as the output to the next module of the system.

In an embodiment of the present invention, figure 16 is the process of paraphrasing of a sentence in normalized form. The purpose of performing of the process of paraphrasing is to output a sentence based on natural language rules with grammar rules which solves the purpose of paraphrasing of complex and long sentence into simple and concise sentence in normal form. Therefore, in order to achieve the desired level of ease and least complexity of a sentence, an input source text sentence may be fed to the paraphrasing system. This paraphrasing system is comprising of three units. First unit is used for generating the synonym of the first word with specific tag order. Second unit is used to replace words with their specific abbreviations. This step makes the system more flexible. Last unit is used to find the active and passive form of the sentence in specific order. Overall system may produce a single sentence. After this step the grammar rules may be applied on the words and finally the words may be joined as per grammar rule of the natural language. Further, after completing the joining of words as sentence, the system may output the final paraphrased sentence to the next input system for further processing.

In an embodiment of the present invention, figure 17 is showing the flow diagram of swapping various words with its abbreviations. In order to perform the process as shown in fig 17, i.e. to find and replace words which can be clipped or contracted or the place where acronym can be fit into the words of original sentence. Benefit of this system is that in paraphrasing or in summarization words can be normalized or can be get shortened or precise which will also state the context without altering the meaning of a sentence. Sentence for the paraphrasing may be sourced into the abbreviation system of the subpart of the system of the present invention. A person skilled in the art should appreciate that first the words which can be easily clipped without changing the meaning of the words in a sentence are morphed. In the next stage contraction of the words in a sentence is performed so that meaning of the words in a sentence will remain same. In the last stage word, which can be swapped with acronym are swapped with acronym. For an example, like Government can be replaced with Govt or Supreme Court of India can be replaced with SCI. After performing all the tasks of internal systems, the paraphrased sentence will be checked again with the syntax of the grammar of the corresponding natural language. After joining all the words, a final sentence will be outputted from the present system.

In an embodiment of the present invention, figure 18 is the flow diagram of clipping various words with its shorter form. A person skilled in the art should be appreciate that it is the part of paraphrasing unit, which may act as clipper so that new words can be used in place of long words so that this process may not alter the meaning of the words in a sentence. Here the sentence for paraphrasing may be inputted to the system so that clipping operation can be performed on the system. Internal unit may have a dataset which has a map of words with their clipped versions. Words with order are mapped with this dataset and may check for the specific clipped version of the word. The output of this internal unit may output as the new clipped words. There can be scenarios, where there may be no words available for clipping so in that case same original word will be the output of the system.
In an embodiment of the present invention, figure 19 is showing the flow diagram of performing active and passive word translation. A person skilled in the art may appreciate that there can be scenarios where passive voice is more suitable in place of active voice and also there may be scenarios where active voice is more suitable in place of passive voice. The main aim of this process is to bring paraphrased sentences in a manner that the output of the system may produce normalized sentence which have same context as original sentence. As an input for the system of the present invention, the words of a sentence in same manner are fed to the system. System of the present invention may have an active to passive transformer, which is capable of converting active to passive or passive to active speech. After converting the sentences, the system checks whether the narrated and the original sentence have the same context or not. If it has same context, then the system sends the narrated text to the output part of the system. Otherwise, system will send the same original source sentence text to the output part.
In an embodiment of the present invention, figure 20 is the flow diagram of checking similarity value of original and paraphrased sentences wherein the process is used to evaluate the similarity amid two text sentences. Prime aim of the process is to accept the sentences which have meaning or context closed to the reality in real world. There may be two inputs to the present system. First input will be the original sentence and the second input may be the paraphrased sentence. Both the text sentences may be checked by the similarity checking computation system. The process may be capable of generating similarity score value in the range of 0 to 1 which can be decimal values also. If the output of decision unit is low means close to 0 then it means paraphrased sentence is closer in meaning to the original sentence and can be accepted in place or original sentence. Otherwise, if the value is high then it means paraphrased sentence is not closed together in meaning of the original sentence.

In an embodiment of the present invention, figure 21 is showing the block diagram of the internal process of the assigning weights to the legal sentences inside a transformer. The block diagram is showing that to find out the weights of various entities in a complete document. The system may take the input text from a complete document. A document can be composed of many pages. In each page there can be more than one paragraph. Each paragraph can compose of more than many sentences. So, the main aim of the process is to create weights of different entities. At the lowest part, each word is replaced by its specific weight in a sentence. In the next step, weight of each sentence in a paragraph is calculated. Finally, the weight of each page in a document is calculated. This process will provide the most important page in addition to most important paragraphs in the corresponding page and as well as the most important sentences in those paragraphs. And finally, the most weighted words in those specific sentences. So, by this mechanism it may be easy to check the main idea of the paragraph and as well as main idea inside the single page and finally the main idea inside the complete document and with the additional information related to the main idea.

In an embodiment of the present invention, figure 22 is the block diagram showing attributes needed to check similarity based on the information combined. As the present invention is for the common man who has little or no knowledge of the legal judgments, it is desired that the language used in the legal judgements should be made simpler for a common man to understand. Therefore, if one person wants to read the judgments then that person will come across various complex legal words as well as the way in which document is written is also not easy to understand with respect to perspective of that man. So, the present invention tries to solve above problem by using three blocks of knowledge. First one is linguistic knowledge of the natural language in which judgment is written. In the second block there is full encyclopaedia knowledge of that natural language and final block is the encyclopaedia knowledge of the entire legal language. These three knowledge blocks together will serve the purpose of solving the above stated problem. The paraphrased and the summary system may use these blocks to find the sentence similarity so that new sentences can be created and similarity vector values can be further used for processing other units.

In the other embodiment of the present invention figure 23 is showing the process of segmentation of original sentence and its paraphrase sentence. The figure has shown the basic steps needed to solve the complexity problems in the legal sentences. The main aim of these steps is to create a sentence with very low complex words with concise and information in normal or in standard form. An original sentence is composed of some English and as well as legal words which can be complex or hard to interpret in terms of the meaning. In the process, the words may be tagged in the seven parts of the grammar of the natural language. After this step, natural language words with parts of speech tagging are segmented from the legal language words. After this phase, translation process is used to check the complexity of the parts of speech words. Upon determination of presence of complexity, then words are translated with more similar words in the meaning and making sure that the part of speech tagging information will remain same. After performing the translation of words of natural language and as well as the legal language, words are combined together with help of attached parts of speech tags and with the help of English grammar. This output is the paraphrased sentence of the original text which is refined and normal form of the original source text.
In the other embodiment of the present invention, the Figure 24 showing the process of creating a summary of a legal document. The summary could be generated in various ways.
For creation of summary, the method need to create a pre-trained model which may be efficient in accepting text as sequence and can understand the parts of sentence with their POS tagged named entities, intents like issues, arguments, reasoning, facts and other information with the flow of sentence. Initially the pre-trained model can take entire sequence once at a time and understand the context of every part of the sentence. This model could understand the context in terms of what, why, which, where, when and how terms are intended and utilized inside the sentence. The main functions of the model are to check the input sequence of each paragraph. With the aid of pre-trained model, it may be checked that what is the main idea or intent in that sequence. If it gets the main idea, then it moves forwards to another sequence. Otherwise, it will take another sequence in addition to the previous sequence and try to fetch the main idea from it using model. After getting the main idea of the paragraph, model may check for the main points related with that idea in a paragraph. It means how the entities in the main idea tags are related to the other sentences in a paragraph. Similarly, each paragraph will be checked with the help of already trained model for the main points and corresponding main idea. In each page, all main ideas with their corresponding main points are ranked based on the model ranking process. Similarly, for each page above process is followed.
Now, for creating a first kind of summary, main points which are ranked higher may be considered for summary which will reveal most significant facts.
For second type of summary, key points are selected as in the summary which reveals the evidences and reasons along with main points context.
For third type of summary which is outline summary, arguments in addition to main points are stated. This type of summary may be the shortest in all the above types of summaries.
Finally, a detailed summary may be generated in bulleted points with the help of techniques to generate all the facts, evidences, arguments and other important information in the judgment.
By determining the intention of terms and utilization thereof, the present invention can create translation in natural language with simplified versions and can create summary in cognitive way.
A person skilled in the art would appreciate that the ranking mechanism may comprises the following steps, wherein first of all, the main ideas in the judgment are listed as same level of tree. Further, the model will check the connectivity between these main ideas. Additionally, the connected graph is created that may show the movement of information from one main idea to another main idea with the help of trained model. Here it should be noted that the some of the nodes may possible may not be connected to any other node in that case the main points related to that idea is searched in the entire text of the judgment. The nodes also tries to generate another, ideas relative to the original ideas and related main points to it.
The ranking of the nodes may be given by model based on the most centric information node which have maximum paths and if there is only one path from a node then it can be either parent node or a leaf node. The model may check whether the given node can be parent or leaf node depending on the weight vectors given by model.

By the above process the text as a sequence can be ranked and therefore more accurate results could be provided in order to translate the document in desired target language.

In an embodiment of the present invention, figure 25 is a system diagram of the invention. The system 2500 consists of various components/modules in order to perform the invention. As shown, the system 2500 is processing plurality of data present in a document. The system comprises of an identifier 2501. The identifier 2501 is identifying the text which may be present in many sentences and many paragraphs of the document. A person skilled in the art would be able to understand that the identifier could be input device used to get the document which is required to be processed. It could be completely hardware element or a combination of software and hardware modules. The system also comprises of an extractor 2502. The main job of the extractor is to extract each one of one or more sentences present in the paragraph of the document. It should be noted that the process of extraction could be the same process as explained in above paragraphs or it could be the combination of other processes by which extraction of the text could be performed within the scope of the invention. The system us also having a tokenizer 2503. The tokenizer 2503 tokenizes each one of one or more words of each one of one or more sentences present in the plurality of paragraphs of the document, wherein the tokenizing is based on identification of grammar text and legal text. A person skilled in the art should understand that the process of tokenization is already explained in above paragraphs and for the sake of clarity the same is not being repeated herein. However, the process of tokenization may be performed by tokenizer 2503 in conjunction with the other processes being performed by the other units of the system 2500. The pre-processor 2504 is creating a Combination of plurality of tokenized words of each of one or more words present in the plurality of paragraphs of the document, wherein the combining plurality of tokenized words are based on grammar knowledge of the plurality of tokenized words. Here it should be noted that the grammar knowledge of the plurality of tokenized words are being updated and stored in a database in real-time, which may be used for further processing of the same document or the other document. The system is further having an AI Model 2505, which performs the functions of tagging. Since, the tagging of words in translation process is required to be processed accurately, the AI model 2505 tagging the one of one or more tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging. The AI model 2505 is also performing the tagging wherein the one of one or more combined tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging. It is important for the AI model 2505 to perform the tagging to the individual words or to the combination of words for proper translation as it is described above. The system further comprises of a processor 2506, which performs the important steps, wherein the processor extracts the grammar text from the tagged tokenized words and also the legal text from the tagged tokenized words. The processor is in communication with a computing module which identifies a cognitive intent and cognitive context of the plurality of tokenized paragraphs, wherein the identifying the cognitive intent of the tokenized paragraph includes the determination of relationship in between tagged tokenized grammar text and tagged tokenized legal text. Here, it should be noted that the extracted grammar text is stored in a grammar dataset and the extracted legal text is stored in a legal dataset. For the sake of clarity, it is explained herein that both the grammar dataset and legal dataset are the part of a database 2509. The database 2509, is partitioned and is programmed in a way that the legal text may not be stored in grammar dataset and similarly the grammar dataset could not be stored in the legal dataset. An important aspect of the invention is performed by a mapping module 2508, wherein the tagged tokenized grammar text with the text stored in the grammar dataset is mapped for creating a processed grammar text and tagged tokenized legal text is mapped with the text stored in the legal dataset for creating a processed legal text. Once the mapping process is completed that text is then fed to a generating tool which creates a revised sentence by combining the processed grammar text and processed legal text. Once the revised sentence is created, the revised sentence is again provided to the processor. At this time the processor extracts the intents and entities from the revised sentence and depending upon the extracted data, the processor is then applied weight vectors to the plurality of words of the revised sentence. The idea behind applying the weight vectors is to identify which word will be optimum to use and therefore ranked accordingly. Further, the processor 2506 is also determines the cognitive context and cognitive intent of the revised sentence. Once, the determination of the cognitive intent and cognitive context then the generating tool generates a summary by combining plurality of revised sentences. Here it should be noted that to create the summary from the revised sentences, it is important to evaluate the cognitive context and cognitive intent as the text with highest ranking in terms of evaluation of the required intents and entities along with cognitive context and cognitive intent are presented in a way to provide more clear and concise summary of the document.

It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. Instructions may also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function/act specified in the flowchart and/or block diagram block or blocks. Instructions may also be loaded onto a computer or other programmable data processing apparatus like a scanner/check scanner to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

It should also be noted that in other implementations, the function(s) noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. In the drawings and specification, there have been disclosed exemplary embodiments of the invention. Although specific terms are employed, they are used in a generic and descriptive sense only and not for purposes of limitation.

We Claim:

1. A method (100) of processing plurality of data present in a document, comprising the steps of:
Identifying (101) one or more paragraphs in the document;
Extracting (101) each one of one or more sentences present in the plurality of paragraphs of the document;
Tokenizing (102) each one of one or more words of each one of one or more sentences present in the plurality of paragraphs of the document, wherein the tokenizing is based on identification of natural language, grammar text and legal text;
Creating (103) a Combination of plurality of tokenized words of each of one or more words present in the plurality of paragraphs of the document, wherein combining plurality of tokenized words are based on grammar knowledge of the plurality of tokenized words;
Tagging (104) the one of one or more tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging;
Tagging (104) the one of one or more combined tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging;
Extracting grammar text from the tagged tokenized words;
Extracting legal text from the tagged tokenized words;
Identifying a cognitive intent of the plurality of tokenized paragraphs, wherein the identifying the cognitive intent of the tokenized paragraph includes the determination of relationship in between tagged tokenized grammar text and tagged tokenized legal text;
Storing the extracted grammar text in a grammar dataset and extracted legal text in a legal dataset;
Mapping (105) tagged tokenized grammar text with the text stored in the grammar dataset for creating a processed grammar text and mapping tagged tokenized legal text with the text stored in the legal dataset for creating a processed legal text;
Creating (106) a revised sentence by combining the processed grammar text and processed legal text;
Extracting (107) intents and entities from the revised sentence;
Applying (108) weight vectors to the plurality of words of the revised sentence;
Determining (109) the cognitive context and intent of the revised sentence;
Generating (110) a summary by combining plurality of revised sentences.

2. The method as claimed in claim 1, wherein determining the cognitive context and intent of the revised sentence comprises computing the relationship in between the plurality of sentences by a computational model wherein an entity, intent and context is being determined and classified as per the computed relationship between the plurality of revised sentences.

3. The method as claimed in claim 1, wherein the part of speech tagging is applied to a part of speech tagged database to produce a part of speech tagged sentence.

4. The method as claimed in claim 1, wherein the extracting the grammar text from the tokenized paragraph comprises identifying nouns, pronouns, adjectives, adverbs, verbs, type of tenses, prepositions, conjunctions.

5. The method as claimed in claim 1, wherein mapping tagged tokenized grammar text with the text stored in the grammar dataset comprises evaluating location of grammatical text of the processed grammar text with the tagged tokenized legal text and by comparing the grammatical text of the processed grammar text with the grammar knowledge of the plurality of tokenized words.

6. The method as claimed in claim 1, wherein the grammar dataset is provided with cognitive alternatives of tagged tokenized grammar text and the legal dataset is provided with cognitive alternatives of the tagged tokenized legal text.

7. The method as claimed in claim 1, wherein combining the processed grammar text and processed legal text comprises evaluating that the combination of the processed grammar text with the grammar knowledge of the plurality of tokenized words and the combination of processed grammar text with the processed legal text matches the cognitive intent of the plurality of tokenized paragraphs of the document.

8. The method as claimed in claim 1, wherein the tagging comprises preparing a corpus wherein a raw data is collected, prepared and labelled based on annotation guidelines.

9. The method as claimed in claim 8, wherein the annotation guidelines comprises evaluation of cognitive intent and cognitive entity of the one or more combination of words of the plurality of sentences of the one or more paragraphs of the document.

10. The method as claimed in claim 8, wherein the tagging comprises updating and training the corpus with the evaluated cognitive intent and with the cognitive entity of the one or more combination of words of the plurality of sentences of the one or more paragraphs of the document.

11. The method as claimed in claim 10, wherein the tagging comprises labelling the one or more words of the plurality of sentences of the one or more paragraphs of the document by combining a selected text of the corpus with the one or more words of the plurality of sentences of the one or more paragraphs wherein the selected text of the corpus is determined based on a query made by a machine learning technique processed on the combination of tokenized words.

12. The method as claimed in claim 11, wherein the machine learning technique comprises extracting one or more features of grammar text from the tokenized words and extracting one or more features of the legal text from the tokenized words.

13. The method as claimed in claim 11, wherein the machine learning technique comprises encoding and combining the one or more extracted features of the grammar text and legal text.

14. The method as claimed in claim 11, wherein the machine learning technique evaluates a vector value of the extracted features of grammar text and of the extracted features of the legal text by combining the encoded features of the one or more extracted features of the grammar text and legal text.

15. The method as claimed in claim 11, wherein the machine learning technique comprises decoding the combined encoded one or more extracted features of the grammar text and legal text.

16. The method as claimed in claim 15, wherein the machine learning technique produces an encoded map vector based on the evaluated vector values of the encoded features of the one or more extracted features of the grammar text and legal text, and wherein the encoded map vector represents the relationship in between the features of the grammar text and features of the legal text.

17. The method as claimed in claim 1, wherein the creating the revised sentence comprises determining complexity of the revised sentence with at least one sentence of the document, wherein upon determining the complexity of the revised sentence being high, the revised sentence is paraphrased with the grammar text and legal text of the corpus to achieve least complexity of the revised sentence by comparing with the sentence of the document.

18. The method as claimed in claim 1, wherein the creating a processed grammar text and creating a processed legal text comprises providing cognitive alternatives to the tagged tokenized grammar text and tagged tokenized legal text respectively, wherein the cognitive alternatives of the grammar text and legal text are synonyms of the tagged tokenized grammar text and tagged tokenized legal text respectively, wherein providing the cognitive alternatives to the tagged tokenized grammar text and tagged tokenized legal text are based on determination and selection of synonyms of the grammar text and legal text respectively, of the sentence.

19. The method as claimed in claim 18, wherein creating the processed grammar text and processed legal text comprises:
identifying synonyms, abbreviation, active form and passive form of the one or more words present in the sentence;
applying grammar rules to the evaluated location of the grammatical text as per the determined location of the grammatical text of the sentence.

20. The method as claimed in claim 19, wherein the abbreviation is identified by performing:
Clipping the tagged tokenized words wherein the cognitive intent of the tagged tokenized words is identical;
Contracting the tagged tokenized words;
Creating acronym of the tagged tokenized words, wherein the clipping is performed by mapping the one or more words of the sentence by comparing the one or more words with the words present in the grammar dataset and legal dataset and by evaluating the order of the grammatical and legal text of the sentence.
21. The method as claimed in claim 1, wherein determining the cognitive context and intent of the revised sentence comprises calculating the similarity in between the revised sentence and sentence of the one or more paragraphs of the document, wherein the calculating the similarity in between the revised sentence and sentence of the one or more paragraphs comprising, matching of linguistic knowledge and encyclopaedia knowledge of natural language and with encyclopaedia knowledge of legal text.

22. The method as claimed in claim 21, wherein by calculating similarity in between the revised sentence and sentence of the one or more paragraphs comprising assigning decimal vector values, wherein the least decimal value indicating similarity in between the revised sentence and sentence of the one or more paragraphs.

23. The method as claimed in claim 1, generating the summary comprises:
Identifying parameters and sequence of the one or more words of the revised sentence of the one or more paragraphs of the document;
Creating a pre-trained model with the identified parameters and sequence of the one or more words of the revised sentence of the one or more paragraphs of the document wherein the pre-trained model process the cognitive intent and context of the revised sentence by identifying the existed identified parameters and sequence in the model.

24. The method as claimed in claim 1, generating the summary comprises:
Identifying legal text parameters and sequence in the one or more revised sentences of the one or more paragraphs of the document;
Creating a pre-trained model with the identified legal text parameters and sequence of the revised sentence of the one or more paragraphs of the document wherein the pre-trained model process the cognitive intent and context of the revised sentence by identifying the existed identified legal text parameters and sequence in the model.

25. The method as claimed in claim 1, generating the summary comprises:
Identifying plurality of legal text present in the one or more revised sentences of the one or more paragraphs of the document, wherein the plurality of legal text is a selection of: prime facts, prime evidences, prime arguments;
Creating a pre-trained model with the identified plurality of legal text.

26. The method as claimed in claim 1, wherein applying weight vectors to the plurality of words of the revised sentence comprises ranking the revised sentences based on highest weight vector value by combining the one or more words of the revised sentence;
Creating a parse tree of the revised sentence based on the context and grammar knowledge of the revised sentence;
Identifying an important context of the revised sentence based on the parse tree value;
Storing the parse tree value along with the important context of the revised sentence;
Ranking the important context of the revised sentence based on the parse tree value;
Presenting the highest ranking important context plurality of words of the revised sentence.

27. The method of processing the data in the document as claimed in claim 6, wherein applying the cognitive alternative to the tagged tokenized grammar and legal words of the sentence of the paragraph comprises applying cognitive alternatives to the immediate one of one or more grammar and legal text of the sentence.

28. The method of processing the data in the document as claimed in claim 6, wherein applying the cognitive alternative to the tagged tokenized grammar and legal words tokenized words of the sentence of the paragraph comprises applying cognitive alternatives to the later one of one or more text of the sentence, wherein the cognitive alternatives of the grammar text and legal text are ranked based on the determination of the alternative cognitive intent of the tokenized words of the sentence of the paragraph.

29. The method of processing the data in the document as claimed in claim 28, wherein the cognitive alternatives of the grammar text and legal text with higher ranking are formatted and stored in a cognitive database, wherein the formatted cognitive alternatives of the grammar text and legal text with higher ranking are applied more weightage.

30. The method of processing the data in the document as claimed in claim 29, wherein creating the revised sentence comprises mapping formatted cognitive alternatives of the grammar text with the formatted cognitive alternatives of the legal text with more weightage.

31. The method of processing the data in the document as claimed in claim 30, wherein creating the revised sentence comprises mapping formatted cognitive alternatives of the grammar text of more weightage with the tokenized legal text and simultaneously mapping formatted cognitive alternatives of the legal text of more weightage with the tokenized grammar text.

32. A system for processing plurality of data present in a document, comprising:
Identifying, by an identifier, one or more paragraphs in the document;
Extracting, by an extractor, each one of one or more sentences present in the paragraph of the document;
Tokenizing, by a tokenizer, each one of one or more words of each one of one or more sentences present in the plurality of paragraphs of the document, wherein the tokenizing is based on identification of grammar text and legal text;
Creating, by a pre-processor, a Combination of plurality of tokenized words of each of one or more words present in the plurality of paragraphs of the document, wherein the combining plurality of tokenized words are based on grammar knowledge of the plurality of tokenized words;
Tagging, by an AI model, the one of one or more tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging;
Tagging, by the AI model, the one of one or more combined tokenized words of one or more sentences present in the plurality of paragraphs, wherein the tagging comprises part of speech tagging;
Extracting, by a processor, grammar text from the tagged tokenized words;
Extracting, by the processor, legal text from the tagged tokenized words;
Identifying, by a computing module, a cognitive intent and cognitive context of the plurality of tokenized paragraphs, wherein the identifying the cognitive intent of the tokenized paragraph includes the determination of relationship in between tagged tokenized grammar text and tagged tokenized legal text;
Storing, in a database, the extracted grammar text in a grammar dataset and extracted legal text in a legal dataset;
Mapping, by a mapping module, the tagged tokenized grammar text with the text stored in the grammar dataset for creating a processed grammar text and tagged tokenized legal text with the text stored in the legal dataset for creating a processed legal text;
Creating, by a generating tool, a revised sentence by combining the processed grammar text and processed legal text;
Extracting, by the processor, intents and entities from the revised sentence;
Applying, by the processor, weight vectors to the plurality of words of the revised sentence;
Determining, by the processor, the cognitive context and intent of the revised sentence;
Generating, by the generating tool, a summary by combining plurality of revised sentences.

33. The system as claimed in claim 32, wherein the processor determining the cognitive context and intent of the revised sentence by computing the relationship in between the plurality of sentences by a computational model wherein an entity, intent and context is being determined and classified as per the computed relationship between the plurality of revised sentences.

34. The system as claimed in claim 32, wherein the mapping module mapping the tagged tokenized grammar text with the text stored in the grammar dataset by evaluating location of grammatical text of the processed grammar text with the tagged tokenized legal text and by comparing the grammatical text of the processed grammar text with the grammar knowledge of the plurality of tokenized words.

35. The system as claimed in claim 34, wherein the grammar dataset is provided with cognitive alternatives of tagged tokenized grammar text and the legal dataset is provided with cognitive alternatives of the tagged tokenized legal text.

36. The system as claimed in claim 32, wherein the generating tool combining the processed grammar text and processed legal text by evaluating that the combination of the processed grammar text with the grammar knowledge of the plurality of tokenized words and the combination of processed grammar text with the processed legal text matches the cognitive intent of the plurality of tokenized paragraphs of the document.

37. The system as claimed in claim 32, wherein the AI model performs tagging by preparing a corpus wherein a raw data is collected, prepared and labelled based on annotation guidelines.

38. The system as claimed in claim 37, wherein the annotation guidelines comprises evaluation of cognitive intent and cognitive entity of the one or more combination of words of the plurality of sentences of the one or more paragraphs of the document.

39. The system as claimed in claim 32, wherein the AI model performs tagging by updating and training the corpus with the evaluated cognitive intent and with the cognitive entity of the one or more combination of words of the plurality of sentences of the one or more paragraphs of the document.

40. The system as claimed in claim 32, wherein the AI model performs tagging by labelling the one or more words of the plurality of sentences of the one or more paragraphs of the document by combining a selected text of the corpus with the one or more words of the plurality of sentences of the one or more paragraphs wherein the selected text of the corpus is determined based on a query made by a machine learning technique processed on the combination of tokenized words.

41. The system as claimed in claim 40, wherein the AI model performs the machine learning technique by extracting one or more features of grammar text from the tokenized words and extracting one or more features of the legal text from the tokenized words.

42. The system as claimed in claim 40, wherein AI model performs the machine learning technique by encoding and combining the one or more extracted features of the grammar text and legal text.

43. The system as claimed in claim 40, wherein AI model performs the machine learning technique by evaluating a vector value of the extracted features of grammar text and of the extracted features of the legal text by combining the encoded features of the one or more extracted features of the grammar text and legal text.

44. The system as claimed in claim 40, wherein the AI model performs the machine learning technique by decoding the combined encoded one or more extracted features of the grammar text and legal text.

45. The system as claimed in claim 44, wherein the AI model performs the machine learning technique to produce an encoded map vector based on the evaluated vector values of the encoded features of the one or more extracted features of the grammar text and legal text.

46. The system as claimed in claim 45, wherein the encoded map vector represents the relationship in between the features of the grammar text and features of the legal text.

47. The system as claimed in claim 1, wherein the creating the revised sentence comprises determining complexity of the revised sentence with at least one sentence of the document.

48. The system as claimed in claim 47, wherein upon determining the complexity of the revised sentence being high, the revised sentence is paraphrased with the grammar text and legal text of the corpus to achieve least complexity of the revised sentence by comparing with the sentence of the document.

49. The system as claimed in claim 1, wherein the creating the processed grammar text and creating a processed legal text comprises providing cognitive alternatives to the grammar text and legal text respectively.

50. The system as claimed in claim 49, wherein the cognitive alternatives of the grammar text and legal text are synonyms of the grammar text and legal text respectively.

51. The system as claimed in claim 50, wherein providing the cognitive alternatives to the grammar text and legal text are based on determination and selection of synonyms of the grammar text and legal text of the sentence.

52. The system as claimed in claim 50, wherein creating a processed grammar text and processed legal text comprises:
identifying synonyms, abbreviation, active form and passive form of the one or more words present in the sentence;
applying grammar rules to the evaluated location of the grammatical text as per the determined location of the grammatical text of the sentence.

53. The system as claimed in claim 52, wherein the abbreviation is identified by performing:
Clipping the tagged tokenized words wherein the cognitive intent of the tagged tokenized words is identical;
Contracting the tagged tokenized words;
Creating acronym of the tagged tokenized words.

54. The system as claimed in claim 54, wherein the clipping is performed by mapping the one or more words of the sentence by comparing the one or more words with the words present in the grammar dataset and legal dataset and by evaluating the order of the grammatical and legal text of the sentence.

55. The system as claimed in claim 1, wherein determining the cognitive context and intent of the revised sentence comprises calculating the similarity in between the revised sentence and sentence of the one or more paragraphs of the document.

56. The system as claimed in claim 55, wherein the calculating the similarity in between the revised sentence and sentence of the one or more paragraphs comprising, matching of linguistic knowledge and encyclopaedia knowledge of natural language and with encyclopaedia knowledge of legal text.

57. The system as claimed in claim 56, wherein by calculating similarity in between the revised sentence and sentence of the one or more paragraphs comprising assigning decimal vector values, wherein the least decimal value indicating similarity in between the revised sentence and sentence of the one or more paragraphs.

58. The system as claimed in claim 1, wherein the generating tool generating the summary comprises:
Identifying parameters and sequence of the one or more words of the revised sentence of the one or more paragraphs of the document;
Creating a pre-trained model with the identified parameters and sequence of the one or more words of the revised sentence of the one or more paragraphs of the document wherein the pre-trained model process the cognitive intent and context of the revised sentence by identifying the existed identified parameters and sequence in the model.

59. The system as claimed in claim 1, wherein the generating tool generating the summary comprises:
Identifying legal text parameters and sequence in the one or more revised sentences of the one or more paragraphs of the document;
Creating a pre-trained model with the identified legal text parameters and sequence of the revised sentence of the one or more paragraphs of the document wherein the pre-trained model process the cognitive intent and context of the revised sentence by identifying the existed identified legal text parameters and sequence in the model.

60. The system as claimed in claim 1, wherein the generating tool generating the summary comprises:
Identifying plurality of legal text present in the one or more revised sentences of the one or more paragraphs of the document, wherein the plurality of legal text is a selection of: prime facts, prime evidences, prime arguments;
Creating a pre-trained model with the identified plurality of legal text.

61. The system as claimed in claim 1, wherein the processor applying weight vectors to the plurality of words of the revised sentence by:
ranking the revised sentences based on highest weight vector value by combining the one or more words of the revised sentence;
Creating a parse tree of the revised sentence based on the context and grammar knowledge of the revised sentence;
Identifying an important context of the revised sentence based on the parse tree value;
Storing the parse tree value along with the important context of the revised sentence;
Ranking the important context of the revised sentence based on the parse tree value;
Presenting the highest ranking important context plurality of words of the revised sentence.