Nils Alberti


pv-avatar ©EUV

I studied computer science and sociology at the TU Dortmund University and the Humboldt University of Berlin. I completed my studies with a diploma thesis entitled „Latent Semantic Indexing zum Aufschlüsseln philosophischer Werke“. My central idea is how knowledge in the humanities, which is gathering dust in libraries in the form of uncounted dissertations, conference proceedings and monographs, can not only be digitised but also made usable. The long-term goal is a search engine that finds all passages of any work in which this point is also discussed: A search for § 261 from Hegel's Grundlinien der Philosophie des Rechts would then not only bring to light Marx's thesis that Hegel poses an unjust antinomy there, but also mentions in the secondary literature would be found, as would a short note on this in a habilitation thesis. A simple word comparison, as used by Google Books, reaches its limits with the word power of Schelling, Adorno or Sartre. Semantic algorithms, which capture the meaning of individual words or entire texts, offer progress here.
While my thesis was about the extent to which semantic methods are suitable for philosophical texts, I am now interested in how these methods can be adapted for very large corpora and how distortions resulting from the linguistic idiosyncrasies of the authors can be avoided.