Back to results
US 8,788,523 B2Granted
Systems, methods and software for processing phrases and clauses in legal documents
AssigneeThomson Reuters Global Resources
InventorsKingsley Martin, Trace Liggett, Gary Berosik, Thomas Zielund, Dietmar Dorr, Jochen L. Leidner
Life of this patent
Estimating the term…
Abstract
Exemplary systems for processing a corpus of documents, such as legal contracts or agreements, are disclosed. The systems include a phrase discovery engine which derives statistics and phrase equivalence classes, groups of phrase equivalence classes, and uberphrases (clauses) bounded by phrases. These can be used to determine origins of phrases or clauses within given legal contract or to suggest alternative phrases and clauses.
Claims25
A computer-implemented method of classifying text in documents comprising:
- electronically receiving for storing in a memory a set of electronic documents each comprising electronic text;
- parsing by a processor in electrical communication with the memory the electronic text from each of the set of electronic documents in a plurality of phrases or clauses;
- associating each of the set of electronic documents with a transactional document type;
- associating at least a set of the plurality of phrases or clauses with a clause type classification, the clause type classification being associated with one or more transactional document type;
- and
- generating by the processor a reference set of clauses for each clause type classification and organizing the reference set of clauses into a clause index structure, whereby the reference set of clauses is subsequently used in processing user inputs, the processing user inputs comprising:
- receiving at a server a set of user input electronic text input by a user via a remote user device over a network;
- parsing by the server the set of user input electronic text to yield a set of user input phrases or clauses;
- comparing by the server the set of user input phrases or clauses to the reference set of clauses, the comparing the set of user input phrases or clauses including computing a set of similarity metrics;
- identifying a set of one or more match clauses from the reference set of clauses, the set of match clauses determined to be similar to one or more of the set of user input phrases or clauses based at least in part on the computed set of similarity metrics;
- and
- generating for output to the remote user device a signal representing the set of match clauses.
9 more elements
A system comprising a processor adapted to access and process machine readable instructions stored in a memory and access data stored in a database, the system in electronic communication with a network and further comprising:
- an input adapted to receive a set of electronic documents each comprising electronic text;
- a parser executed by the processor to parse the electronic text from each of the set of electronic documents and yield a plurality of phrases or clauses;
- a classifier executed by the processor to associate each of the set of electronic documents with a transactional document type
- and associate at least a set of the plurality of phrases or clauses with a clause type classification, the clause type classification being associated with one or more transactional document type;
- a clause index structure stored in a database accessible by the processor and comprising a reference set of clauses for each clause type classification
- the input further adapted to receive a set of user input electronic text from a remote user device via a network in electrical communication with the system;
- the parser further executed by the processor to parse the set of user input electronic text to yield a set of user input phrases or clauses;
- the processor configured to access from the database the clause index structure and to access from memory and execute a set of code adapted to:
- compare the set of user input phrases or clauses to the reference set of clauses:
- compute a set of similarity metrics:
- and identify a set of one or more match clauses from the reference set of clauses based at least in part on the computed set of similarity metrics;
- and
- a signal generated by the processor and representing the set of match clauses;
- and
- an output in electrical communication with the network and adapted to output the generated signal via the network for intended receipt at the remote user device.
12 more elements
A transactional document drafting system implemented by a computer comprising a processor adapted to process a set of instructions stored in a memory, the system comprising:
- an input adapted to receive from a user an input representing an identified transactional document type of interest to the user;
- a first set of instructions executed by the processor to present via a user display a clause index structure representing a library of clauses, the clause index structure and library of clauses being associated with the identified transactional document type, the library of clauses comprised of clauses previously identified from a reference set of documents, and wherein the library of clauses is organized based on a clause type classification and the clause index structure presents the library of clauses based at least in part on the clause type classification;
- upon receiving a signal representing a user selection of a clause from the presented library of clauses, a second set of instructions executed by the processor to present for display the selected clause from the library of clauses;
- a user interface adapted to insert the selected clause into an electronic document active in a word processing application operating on a user computing device;
- the input further adapted to receive the reference set of documents;
- a parser executed by the processor to parse electronic text from each of the reference set of documents to identify a set of phrases or clauses;
- the processor adapted to process a set of instructions stored in the memory to compare identified phrases or clauses from a first of the reference set of documents with phrases or clauses identified from one or more of the other reference set of documents and to determine a set of default clauses corresponding, respectively, to a set of clause classification types based on the comparison.
4 more elements
A computer-implemented method for drafting transactional documents, the method comprising:
- receiving from a user an input representing an identified transactional document type of interest to the user;
- executing by a processor a first set of instructions stored in a memory accessible by the processor to present to a user a clause index structure representing a library of clauses, the clause index structure representing a library of clauses being associated with the identified transactional document type, the library of clauses comprised of clauses previously identified from a reference set of documents, and wherein the library of clauses is organized based on a clause type classification and the clause index structure presents the library of clauses based at least in part on the clause type classification;
- receiving a signal representing a user selection of a clause from the presented library of clauses;
- executing a second set of instructions to present the selected clause from the library of clauses;
- incorporating the selected clause into an electronic document active in a word processing application;
- and
- receiving the reference set of documents;
- parsing electronic text from each of the reference set of documents to identify a set of phrases or clauses;
- comparing identified phrases or clauses from a first of the reference set of documents with phrases or clauses identified from one or more of the other reference set of documents;
- and
- determining a set of default clauses corresponding, respectively, to a set of clause classification types based on the comparison.
8 more elements
Dependencies are inferred from the claim text.