Neural Networks and Graphological Research in the Literature

Penskaya, Elena Naumovna

D.Sc. in Philology, Professor, School of Philology, Academic Supervisor, Faculty of Humanities, HSE University, Head of Research Group, Interdisciplinary Research Center, MIPT University, Moscow, Russia

The talk will discuss ways to build neural networks for the literary sources’ graphological study; the author focuses on artificial intelligence’ possibilities in the handwriting expertise. The problem of authenticity, evidence in determining the original or forgery in art, the process of attribution of literary sources, the compilation of “portraits” and “libraries” of handwriting is discussed.
Different methods of using large artificial neural networks for analyzing the handwritten text’s scanned fragments are discussed.
The general objectives: 1) to provide an expert opinion, 2) predict the errors.
The assessment of the decision’s reliability is a new and extremely important aspect that provides the artificial intelligence’ trustworthy.

Keywords: manuscripts, neural networks, artificial intelligence, literary sources, expertise