Computational Approaches to Ancient Writing Systems

In this talk, David Osgarby, Wout Sinnaeve and Michelle Waldispühl, introduce DESCRYPT and present computational case studies on deciphering Pictish symbols and detecting, transcribing, and restoring runic inscriptions. 

Image may contain: Font, Logo, Symbol, Graphics.

Abstract

We first present shortly the interdisciplinary research project Echoes of History: Analysis and Decipherment of Historical Writings (DESCRYPT) followed by studies using compuational approaches for the decipherment of Pictish symbols (David Osgarby) and the analysis and restoration of runic inscriptions (Wout Sinnaeve).Funded by the Swedish Riksbankens Jubileumsfond, DESCRYPT is an international and interdisciplinary project that combines philological approaches with AI and computational methods to study historical texts written in rare, unusual, or undeciphered scripts. The project will create digital collections and tools to identify, transcribe, and interpret these sources. Osgarby’s and Sinnaeve’s projects are case studies within the project.Osgarby discusses an implementation of a computational decipherment tool to his Pictish symbol data and shows the challenges and insights from a philological perspective. Sinnaeve presents his work of the application of computer vision to the detection and transcription of runic inscription using object detection and graph-based models.

Bio

David Osgarby is a PhD-student in Celtic at the University of Glasgow, affiliated with DESCRYPT as guest researcher, Wout Sinnaeve is a PhD-student in linguistics at ILOS/DESCRYPT and Michelle Waldispühl is a professor of German linguistics and co-PI of DESCRYPT. 

Published Oct. 16, 2026 10:48 AM - Last modified Oct. 16, 2026 10:50 AM