Clause discovery for validation of documents

1. A method for managing documents over a network using one or more processors that execute instructions to perform actions, comprising:

instantiating a document engine to identify one or more clauses in a document, wherein the document is associated with one or more document types; and

instantiating a validation engine to perform actions, including:

determining one or more evaluations for the one or more clauses, wherein the one or more evaluations employ one or more of machine learning (ML) models based on one or more categories associated with each clause, semantic evaluations based on the one or more document types associated with the document, and text evaluations based on content of the clause;

assigning a category and a confidence score to the one or more clauses based on the one or more evaluations performed by one or more of textual evaluators or the semantic evaluators;

determining the one or more ML models based on the category assigned to the one or more clauses;

classifying the one or more clauses based on the one or more ML models;

employing a result of the classification to modify a confidence score associated with the one or more clauses; and

identifying one or more inconsistent evaluations, wherein each inconsistent evaluation is associated with a clause that is both associated with a confidence score that exceeds a high threshold value and a quantity of updates to content of the clause that exceeds another high threshold value; and

instantiating a machine learning (ML) engine to perform actions including:

retraining a portion of the one or more ML models based on updated content of the one or more clauses associated with the one or more inconsistent evaluations, wherein the retrained portion includes one or more increased confidence scores with each of the one or more clauses that already include content that is equivalent to the content updates of the one or more clauses associated with the inconsistent evaluations, and wherein the retrained portion of the one or more ML models associate one or more decreased confidence scores with the one or more clauses associated with the one or more inconsistent evaluations.

2. The method of claim 1 , wherein the validation engine performs further actions including: employing a result of the classification to modify the confidence score associated with the one or more clauses.

3. The method of claim 1 , wherein the one or more clauses are included in a data object as one or more self-referential records to improve the performance of computing resources employed to execute the instructions by providing for one or more higher-order relationships to be represented in a same data object rather than requiring separate data objects to maintain the one or more relationships.

4. The method of claim 1 , wherein determining the one or more evaluations, further comprises:

employing one or more text evaluators to evaluate one or more of grammar, parts-of-speech, word-counts, or character counts of the one or more clauses;

employing one or more semantic evaluators to evaluate the one or more document types associated with the one or more clauses; and

assigning the category and the confidence score to the one or more clauses based on the one or more evaluations performed by one or more of the text evaluators or the semantic evaluators.

5. The method of claim 1 , further comprising, instantiating a feedback engine that performs actions, including:

displaying a report that includes the document and the one or more clauses, wherein the category and the confidence score associated with each clause is included in the report; and

capturing the one or more actions that are stored in an activity journal.

6. The method of claim 1 , wherein the validation engine performs further actions, comprising:

providing document training data;

employing the document training data to train the one or more ML models; and

modify the one or more ML models based on evaluation of the one or more trained ML models.

7. A processor readable non-transitory storage media that includes instructions for managing documents over a network, wherein execution of the instructions by one or more processors on one or more network computers performs actions, comprising:

instantiating a document engine to identify one or more clauses in a document, wherein the document is associated with one or more document types; and

instantiating a validation engine to perform actions, including:

determining one or more evaluations for the one or more clauses, wherein the one or more evaluations employ one or more of machine learning (ML) models based on one or more categories associated with each clause, semantic evaluations based on the one or more document types associated with the document, and text evaluations based on content of the clause;

assigning a category and a confidence score to the one or more clauses based on the one or more evaluations performed by one or more of textual evaluators or the semantic evaluators;

determining the one or more ML models based on the category assigned to the one or more clauses;

classifying the one or more clauses based on the one or more ML models;

employing a result of the classification to modify a confidence score associated with the one or more clauses; and

identifying one or more inconsistent evaluations, wherein each inconsistent evaluation is associated with a clause that is both associated with a confidence score that exceeds a high threshold value and a quantity of updates to content of the clause that exceeds another high threshold value; and

instantiating a machine learning (ML) engine to perform actions including:

retraining a portion of the one or more ML models based on updated content of the one or more clauses associated with the one or more inconsistent evaluations, wherein the retrained portion includes one or more increased confidence scores with each of the one or more clauses that already include content that is equivalent to the content updates of the one or more clauses associated with the inconsistent evaluations, and wherein the retrained portion of the one or more ML models associate one or more decreased confidence scores with the one or more clauses associated with the one or more inconsistent evaluations.

8. The media of claim 7 , wherein the validation engine performs further actions including: employing a result of the classification to modify the confidence score associated with the one or more clauses.

9. The media of claim 7 , wherein the one or more clauses are included in a data object as one or more self-referential records to improve the performance of computing resources employed to execute the instructions by providing for one or more higher-order relationships to be represented in a same data object rather than requiring separate data objects to maintain the one or more relationships.

10. The media of claim 7 , wherein determining the one or more evaluations, further comprises:

employing one or more text evaluators to evaluate one or more of grammar, parts-of-speech, word-counts, or character counts of the one or more clauses;

employing one or more semantic evaluators to evaluate the one or more document types associated with the one or more clauses; and

assigning the category and the confidence score to the one or more clauses based on the one or more evaluations performed by one or more of the text evaluators or the semantic evaluators.

11. The media of claim 7 , further comprising, instantiating a feedback engine that performs actions, including:

displaying a report that includes the document and the one or more clauses, wherein the category and the confidence score associated with each clause is included in the report; and

capturing the one or more actions that are stored in an activity journal.

12. The media of claim 7 , wherein the validation engine performs further actions, comprising:

providing document training data;

employing the document training data to train the one or more ML models; and

modify the one or more ML models based on evaluation of the one or more trained ML models.

13. A system for managing documents comprising:

a network computer, comprising:

a transceiver that communicates over the network;

a memory that stores at least instructions; and

one or more processors that execute instructions that perform actions, including:

instantiating a document engine to identify one or more clauses in a document, wherein the document is associated with one or more document types; and

instantiating a validation engine to perform actions, including:

determining one or more evaluations for the one or more clauses, wherein the one or more evaluations employ one or more of machine learning (ML) models based on one or more categories associated with each clause, semantic evaluations based on the one or more document types associated with the document, and text evaluations based on content of the clause;

assigning a category and a confidence score to the one or more clauses based on the one or more evaluations performed by one or more of textual evaluators or the semantic evaluators;

determining the one or more ML models based on the category assigned to the one or more clauses;

classifying the one or more clauses based on the one or more ML models;

employing a result of the classification to modify a confidence score associated with the one or more clauses; and

identifying one or more inconsistent evaluations, wherein each inconsistent evaluation is associated with a clause that is both associated with a confidence score that exceeds a high threshold value and a quantity of updates to content of the clause that exceeds another high threshold value; and

instantiating a machine learning (ML) engine to perform actions including:

retraining a portion of the one or more ML models based on updated content of the one or more clauses associated with the one or more inconsistent evaluations, wherein the retrained portion includes one or more increased confidence scores with each of the one or more clauses that already include content that is equivalent to the content updates of the one or more clauses associated with the inconsistent evaluations, and wherein the retrained portion of the one or more ML models associate one or more decreased confidence scores with the one or more clauses associated with the one or more inconsistent evaluations.

14. The system of claim 13 , wherein the one or more clauses are included in a data object as one or more self-referential records to improve the performance of computing resources employed to execute the instructions by providing for one or more higher-order relationships to be represented in a same data object rather than requiring separate data objects to maintain the one or more relationships.

15. The system of claim 13 , wherein determining the one or more evaluations, further comprises:

employing one or more text evaluators to evaluate one or more of grammar, parts-of-speech, word-counts, or character counts of the one or more clauses;

employing one or more semantic evaluators to evaluate the one or more document types associated with the one or more clauses; and

assigning the category and the confidence score to the one or more clauses based on the one or more evaluations performed by one or more of the text evaluators or the semantic evaluators.

16. The system of claim 13 , further comprising, instantiating a feedback engine that performs actions, including:

displaying a report that includes the document and the one or more clauses, wherein the category and the confidence score associated with each clause is included in the report; and

capturing the one or more actions that are stored in an activity journal.

17. The system of claim 13 , wherein the validation engine performs further actions, comprising:

providing document training data;

employing the document training data to train the one or more ML models; and

modify the one or more ML models based on evaluation of the one or more trained ML models.