The vast quantity of texts online and in large company databases often means that polysemous or varying names of people, places, companies or events cannot be clearly categorized. When reading, people generally determine who or what is meant from the context. Reading the sentence “Müller was the top goal scorer in South Africa,” a well-informed reader would recognize from the context whether the text was referring to Gerd Müller, the one-time “nation’s bomber”, Thomas Müller, his worthy successor and top goal scorer in 2010, or the namesake of the former footballer Gerd Müller, the current German Minister for Development.
Machines cannot automatically understand this but such knowledge is nevertheless essential for a more in-depth comprehension of digital content. Ambiverse, a spin-off of the Max Planck Institute for Informatics in Saarbrücken, has developed an intelligent software solution for the automatic and accurate recognition and interpretation of words and phrases with several meanings in texts. The software relates words in the text to specific entities, such as people, companies, films or books, in reference repositories, free knowledge databases like Wikipedia and WikiData, or company-specific databases. This innovative technology enables the creation of more high-performance search and analysis tools for news archives, company documents, product reviews and much more. Ambiverse’s goal is now to make the software market-ready using the funding received from EXIST and to produce a viable business model.
Dr. Florian Kirschenhofer
Tel.: +49 (0)89 / 29 09 19-31