An example of Artificial Intelligence integration in Humanities research: Ithaca and the corpus of oracular lamellae of Dodona

Abstract

This paper presents a research project on the application of machine learning to the edition of ancient Greek inscriptions. More specifically, it implements Ithaca, a deep neural network architecture, for elaborating a new critical edition of the enquiries to the oracle of Zeus and Dione which are preserved on the lead tablets discovered in the sanctuary of Dodona, in northern Greece.

The goal of the project is twofold: first, it constitutes an attempt to incorporate a new technological tool into the epigraphist’s workflow, mainly state-of-the-art machine learning; second, it is conceived as a case study for evaluating the performance of deep learning for editing a corpus of Greek inscriptions which presents a high level of complexity.