Collaborative AI: The path to fast, reliable concept maps

Creating taxonomies can be a tedious process, particularly for companies that already hoard masses of data. Artificial Intelligence could potentially lower cost and time, but can we really fully rely on auto-taxonomization? Collaborative AI promises the best of both worlds.

Text by Michael Wetzel


Image: © musicman/

Collaborative AI refers to processes in which humans and artificial systems work together. Our recent work has analyzed whether such a method can help terminologists turn existing “flat” terminology lists into a structured concept map, also known as a “knowledge graph”.

In this article, we outline our progress in using Collaborative AI to facilitate the speedy creation of concept maps. Which AI or Machine Learning (ML) algorithms help to draft such maps, and what are the strengths and weaknesses in terms of speed, approach, and quality?

In our research study, we compared a purely manual approach with a semi-automatic, collaborative AI-based knowledge graph creation implemented using the Coreon multilingual knowledge system. Our findings: The increase in efficiency makes the taxonomization of even large terminology databases feasible.


From an uncontrolled haystack of concepts to a ...