AI Data Discovery Interface Using Term Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data discovery platforms are inefficient in organizing and managing large volumes of documents during legal discovery processes, leading to overwhelming tasks for users who need to identify relevant information from thousands of documents.
Innovation Solution
A computer-implemented method that uses artificial intelligence to generate a map of terms and words from a training set of documents, allowing users to identify key concepts and keywords, and provides a user interface to guide the search process, including suggestions for document review and tagging by custodians with limited privileges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users manually search and review thousands of documents to identify relevant information, then comprehensive document review can be achieved, but the time and effort required becomes overwhelming
Solution Approach 1:
The system performs preliminary actions by automatically generating a map of terms and words from documents before the user conducts their search. This pre-processing organizes document content into structured term maps that enable faster subsequent searches, allowing users to quickly locate relevant information without manually reviewing thousands of documents.
Solution Approach 2:
The patent introduces an intermediary mechanism - a map of terms and words - that mediates between the large volume of documents and the user's search needs. This intermediate structure acts as a bridge, transforming unstructured document content into an organized format that facilitates efficient retrieval while maintaining comprehensive review capabilities.
2Loss of information
If a comprehensive user interface provides all available document information and search options, then users can access complete data, but the interface complexity increases making it harder to use
Solution Approach 1:
The user interface is segmented into functional sections based on the map of terms and words. Instead of presenting all document information simultaneously, the interface divides content into organized segments that users can access through term-based navigation. This segmentation maintains information completeness while improving usability through structured presentation.
Solution Approach 2:
Different sections of the user interface provide different levels and types of information quality based on user needs. The interface adapts its presentation - providing detailed term maps in some areas, summary views in others, and contextual information where relevant. This local quality approach ensures users receive appropriate information without being overwhelmed by unnecessary details.
3Productivity
If custodians with limited privileges can freely tag and organize documents, then document organization improves, but system security and control may be compromised
Solution Approach 1:
The system performs preliminary organization through the automatically generated map of terms and words before custodians perform tagging. This pre-organized structure provides a framework that guides custodian actions, ensuring they work within established organizational parameters. This maintains system control while enabling efficient document tagging and organization by custodians with limited privileges.
Data Source
AI summary
A method includes receiving a first set of documents that correspond to data discovery documents. The method further includes generating a map of terms and words based on the first set of documents, the map of terms and words corresponding to concepts. The method further includes receiving from a first user an initial document relating to a data discovery issue. The method further includes requesting the user to provide an identification of at least one of an initial name of a first person, an initial date, and an initial keyword related to the legal issue. The method further includes generating at least one of a similar name of a second person, a similar date, and a similar keyword based on the map of terms and words. The method further includes identifying a review document based on the review document including at least one of the similar name of the second person, the similar date, and the similar keyword.