With the ever growing bodies of data in intranets and on the Internet, a fundamental problem of traditional search engines becomes all the more apparent. The traditional search engine is not able to understand the search query or the searched documents on a semantic level. Only phrases that are identical to the query or very similar on the surface level can be found. This is why documents that would be a perfect match on the semantic level cannot be found if divergent wording and phrasing is used. Millions of examples exist for that particular problem. The following is one of them.
The answer to this problem is the cognitive search engine SEMPRIA-Search, a multi-year development effort departing from semantic search. By using modern language technology and AI, the search query as well as the underlying documents can be understood on a semantic level. Wording and phrasing becomes almost irrelevant, as divergent wording can be countered with knowledge of synonyms, different phrasing with paraphrased data. Varied terminology can be met with knowledge of hierarchies and jargon. SEMPRIA-Search can explain every found match and each inference step. Hence, SEMPRIA-Search finds the relevant documents more precisely and more completely and presents the matches in a very comprehensible way.