Entity Detection

Overview

This technical note describes Lexbe’s Entity Detection+ service which uses machine learning services to extract useful data and objects from unstructured text on an automated basis without the need for human coding.  This service is part of Lexbe’s ‘AI Insights’ - Artificial Intelligence tools for eDiscovery.

Background

eDiscovery in today's cases often includes a high volume of unstructured data. The traditional eDiscovery workflow of keyword searching can be time-consuming, challenging and 'hit or miss'.  Entity recognition can aid the eDiscovery process by making unstructured text structured, leading to easier evidence identification.

Entity recognition is a new technology that can help the legal review team to find important documents faster.  An ‘entity’ as used here is a textual reference to the unique name of real-world objects such as people, places, and commercial items, as well as measurement references such as dates and quantities.  Once entities are identified in unstructured text, they are extracted and placed into specified metadata fields. These metadata fields can then be searched and filtered to isolate particular documents of interest containing specified entities or objects.

Benefits

Lexbe’s Entity Detection service can assist in:

• • Discovering meaning and relationships within data

  • Creating organization of documents

  • Fast and efficient review

  • Privilege protection

Service Description

Lexbe’s Entity Detection service can be run on a Lexbe eDiscovery Platform (LEP) case of any size.  The following Entities will be extracted when found by the machine algorithm within each document in the case:

• • Person - Individuals, groups of people, nicknames, fictional characters

  • Organization - Large organizations, such as a government, company, religion, sports team, etc.

  • Dates -  A full date (for example, 11/25/2017), day (Tuesday), month (May), or time (8:30 a.m.)

  • Other - Entities that don't fit into any of the other entity categories

  • Location - A specific location, such as a country, city, lake, building, etc.

  • Quantity - A quantified amount, such as currency, percentages, numbers, bytes, etc.

  • Title - An official name given to any creation or creative work, such as movies, books, songs, etc.

For example, in the text "John moved to 1313 Mockingbird Lane in 2012," "John" might be recognized as a PERSON, "1313 Mockingbird Lane" might be recognized as a LOCATION, and "2012" might be recognized as a DATE.

Once the Service is run, then the extracted and fielded data can be seen from the Doc Tab in the LEP Document Viewer for each document, or in a table from the Browse View with the specified columns (fields) showing.

Entity Recognition+ objects in the LEP Doc Viewer:

Entity Recognition+ objects in the LEP Browse Tab:

Any of the data in the Entity Recognition+ fields can be filtered on to reduce the documents to only documents containing those entities or objects.  This is done from the filter tool accessible from the Browse and Search pages.

Entity Recognition+ objects in the Filter Tool:

How to Request

Lexbe’s Entity Detection service can be run on a case per request by Lexbe’s Professional Service team. There is no charge for one run Subscription Accounts and a small per document charge for Flex accounts. 

Additional Information

To request this service or additional information, please contact professionalservices@lexbe.com