Treparel KMX

Evalueserve Insightloupe is powered by Treparel KMX text mining and visualization technology

Smart algorithms to simplify tasks in IP and R&D Intelligence

By acquiring Treparel in 2015 - and it's KMX Patent Mining application - Evalueserve gained access to AI technology that effectively underpins human expertise with best-in-class technology using smart algorithms to simplify key tasks in the field of IP and R&D Intelligence.

When Analysts work on a project, KMX can be used for some specific tasks
Explorer Maps

This unique type of visualization, also known as Topographical Maps, provides a data-driven overview of the most prominent topics covered by the documents in a project. They provide another perspective on the project document set in addition to the curation and enrichment (technology tags or taxonomy labels) of the documents done by the experts.

Using the explorer maps, a client can quickly identify documents with a similar terminology. Explorer Maps are generated without any human intervention or effort and can provide a non-human supervised perspective on a project.

Machine-assisted quality review of documents

Using Machine Learning, powered by KMX, Evalueserve Project Managers are assisted in ensuring quality of the work output. The KMX algorithm helps the Project Managers to identify documents that might be wrongly tagged or labelled for relevancy. The benefit is improved retrieval accuracy and consistent quality in the Analysis & Tagging phase of the project on Insightloupe. We call this Quality-check-as-a-Service.

KMX Classic in IPR&D operational toolbox

Evalueserve analysts use the original KMX application for bespoke projects for more advanced studies like IP licensing.

Machine assisted identification of relevant documents (prototype)

Evalueserve analysts also use KMX on Insightloupe to enhance the comprehensiveness (recall) of the study. Once an analyst has reviewed a first set of documents for relevancy on Insightloupe, he can use the KMX algorithm (as a part of secondary searching feature) to identify if there are relevant documents within a second set of documents. This enhances the retrieval of relevant documents and hence the recall of the study.



IPR&D Blog, news & samples
Blog post
What makes a good patent search? Learning from zero-fault-tolerance processes in aviation and medicine
Blog post
Patent licensing in the United States and beyond - Part 2:
Defining a new path for future technology patent licensing
Happy 5th Birthday Insightloupe
Evalueserve attending IPBC Global in Boston: June 16-18, 2019
Insightloupe delivers IP and R&D clarity
Get the highest-quality view of your IP and R&D data when and how you need it
Contact Us