Back to results
US 2021/0065320 A1Published application
Collaborative matter management and analysis
OwnerOfer Bleiweiss, Jeremiah Kelmaninventors
Life of this patent
Estimating the term…
Abstract
A collaborative matter management and analysis tool is presented. Each matter has its own workspace that is built around a collaborative backbone. Users can provide access to the matter workspace to anyone who they want to collaborate with, including others at their organization and persons outside the organization, such as clients, lawyers, experts, vendors and other third parties. Processing of the document is automatic and based on the content and document type. The document, when uploaded, is OCR'd and based on the content, citations are converted to hyperlinks and the cited documents are linked. Documents and events are associated to enable enhanced productivity tools such as searching and e-brief.
Claims20
A method for collaborative matter management comprising:
- continuously gathering intelligence relevant to a matter from all collaborative elements of the matter, wherein said gathering intelligence comprises listening and collecting user activities and data in the collaborative elements, wherein the collaborative elements comprise an evidence database, feeds, witness lists, Master Files, Players, and notes, wherein the evidence database comprises one or more key documents and wherein the notes are associated with the one or more key documents;
- actively training a machine learning algorithm using the intelligence as training data;
- continuously applying said trained machine learning algorithm to a plurality of documents in a review tool to identify and elevate one or more potentially relevant documents in said plurality of documents to a queue;
- the machine learning algorithm assigning a confidence level and providing a justification for each of the one or more documents elevated to the queue, wherein the justification is an indication of a basis for elevation to the queue such as similarities in language with highest rated documents that have already been promoted to the evidence database and relationship with players already associated with documents in the evidence database;
- the machine learning algorithm continuously reviewing and adjusting the confidence level of each of the one or more documents in the queue and ordering the documents in the queue based on the confidence level;
- obtaining inputs from a user's evaluation of documents in the queue to identify any document in the queue for promotion to the evidence database or for return to the review tool;
- updating said intelligence for said active training of said machine learning algorithm if the user identifies one or more of the documents in the queue for promotion to the evidence database or for return to the review tool;
- returning to the review tool the one or more documents in the queue identified for return by the user;
- and
- promoting to the evidence database said one or more documents in the queue identified for promotion by the user.
7 more elements
A method for collaborative matter management comprising:
- continuously gathering intelligence comprising user activities and data on all collaborative elements of a matter, wherein the collaborative elements comprise an evidence database, feeds, witness lists, Master Files, Players and notes, wherein the evidence database comprises one or more key documents and wherein the notes are associated with the one or more key documents;
- actively using the intelligence as training data for a machine learning algorithm;
- continuously applying said trained machine learning algorithm to a plurality of documents in a review tool to identify and elevate one or more potentially relevant documents in said plurality of documents to a queue;
- the machine learning algorithm assigning a confidence level and providing a justification for each of the one or more documents elevated to the queue, wherein the justification is an indication of a basis for elevation to the queue such as similarities in language with highest rated documents that have already been promoted to the evidence database and relationship with players already associated with documents in the evidence database;
- and
- the machine learning algorithm continuously reviewing and adjusting the confidence level of each document in the queue and ordering the documents in the queue based on the confidence level.
3 more elements
A non-transitory computer-readable medium comprising computer-readable instructions for collaborative matter management, wherein execution of said computer-readable instructions by one or more processors causes said one or more processors to:
- continuously gather intelligence relevant to a matter from all collaborative elements of the matter, wherein said gathering intelligence comprises listening and collecting user activities and data in the collaborative elements, wherein the collaborative elements comprise an evidence database, feeds, witness lists, Master Files, Players, and notes, wherein the evidence database comprises one or more key documents and wherein the notes are associated with the one or more key documents;
- actively train a machine learning algorithm using the intelligence as training data;
- continuously apply said trained machine learning algorithm to a plurality of documents in a review tool to identify and elevate one or more potentially relevant documents in said plurality of documents to a queue;
- the machine learning algorithm assigns a confidence level and provides a justification for each of the one or more documents elevated to the queue, wherein the justification is an indication of a basis for elevation to the queue such as similarities in language with highest rated documents that have already been promoted to the evidence database and relationship with players already associated with documents in the evidence database;
- and
- the machine learning algorithm continuously reviews and adjusts the confidence level of each document in the queue and order the documents in the queue based on the confidence level.
3 more elements
Dependencies are inferred from the claim text.