Less is more, they say. But when it comes to creating the world’s most advanced Semantic AI technology – more is more. In the ever expanding world of information, how to navigate within the vast universe of words and concepts?

Since 2000 we’ve been building an extremely powerful text analysis engine Leiki Focus to make a real difference in how people can discover information. After 18 years we’re quite proud of our semantic analysis technology that uses our very own Leiki Ontology – the world’s most comprehensive general-purpose ontology. It identifies over 200,000 language-independent concepts using more than 5 million rules, crafted by our dedicated team of professional  linguists using Leiki’s intelligent tools at our semantic AI center in Helsinki, Finland.

To help navigate in a structure built of millions of rules and concepts, the Leiki team uses a new visualization tool. It produces a graph of any chosen category, displaying the numerous connections to all the linked topics.

With the Leiki Ontology anyone can quickly and easily detect the meaning of all concepts in any piece of text. This goes much deeper than just picking up those exact concepts, or keywords, that are mentioned in the text. Once a concept is mentioned, also the related concepts are automatically activated.

As the image of the “dog training” category illustrates, text mentioning dog training automatically triggers several related categories, such as animal training, dog care, dogs, pets, etc.

Building audiences with Leiki Ontology

In addition to profiling text, the Leiki Ontology is a helpful tool in creating audience segments. First of all, we use the Ontology to build audience data by detecting users’ interests from their browsing. For example, reading a piece of text that has “dog training” adds the user’s browser to all relevant segments, such as pets, animals, dogs, and pet care.

In building segments, the Leiki Ontology helps the audience specialists to pick the right topics. The process is very simple: to create an interest-based audience such as “people interested in dogs”, you don’t need to include in the segment definition all imaginable dog-related keywords. Just by having the ontology category “dogs” there, the audience will already include all users that have interest in any dog related concepts.

Are you looking for a specific audience segment – we can help you!

Concept mapping with complex rules

The deeper you go in the ontology, the more connections you’ll be able to find between different topics. There are several types of connections that exist between concepts in the Leiki Ontology. The example graph of “Steven Mnuchin” shows both related concepts and those used in disambiguation. You’ll see the closest topic categories highlighted at the top of the image.

Semantic AI at your fingertips

Get the power of semantic AI with Leiki’s services, contact us! Our ontology covers 28 languages (and counting) with a range of category schemes from IAB Tier 1 and 2 (version 2) up to our own Hi-Def system of 200,000 topic categories as well as custom taxonomies.