Humanist Discussion Group

Humanist Archives: March 11, 2022, 6:19 a.m. Humanist 35.587 - pubs: paper on neural networks for ancient texts

              Humanist Discussion Group, Vol. 35, No. 587.
        Department of Digital Humanities, University of Cologne
                      Hosted by DH-Cologne
                Submit to:

        Date: 2022-03-10 10:31:05+00:00
        From: Thea Sommerschield <thea.sommerschield@CLASSICS.OX.AC.UK>
        Subject: New tool for restoring and attributing Greek inscriptions using Artificial Intelligence

Dear List Members,

Today Nature publishes our paper
<> “Restoring and
attributing ancient texts using deep neural networks”. The paper
introduces Ithaca: the first assistive tool using deep neural networks
to aid historians in not only restoring the missing text of ancient
Greek inscriptions, but also identifying their original location, and
establishing the date they were written.

This work is a collaboration between the Universities of Venice Ca’
Foscari <>, Oxford
<>and Athens AUEB
<>, and Google’s DeepMind <>.
To make this research widely available to researchers, educators, museum
staff and others, we partnered with Google Cloud
<>and Google Arts & Culture
<>to launch a interactive version
<>of Ithaca for historians to use for their
personal research.

Please find the abstract of the published paper below:

“Ancient History relies on disciplines such as Epigraphy, the study of
inscribed texts known as "inscriptions", for evidence of the thought,
language, society and history of past civilizations. However, over the
centuries many inscriptions have been damaged to the point of
illegibility, transported far from their original location, and their
date of writing is steeped in uncertainty.

We present Ithaca, the first Deep Neural Network for the textual
restoration, geographical and chronological attribution of ancient Greek
inscriptions. Ithaca is designed to assist and expand the historian’s
workflow: its architecture focuses on collaboration, decision support,
and interpretability. While Ithaca alone achieves 62% accuracy when
restoring damaged texts, as soon as evaluated historians use Ithaca
their accuracy leaps to 72%, confirming this synergistic research aid’s
impact. Ithaca can attribute inscriptions to their original location
with 71% accuracy and can date them with a distance of less than 30
years of their ground-truth ranges, redating key texts of Classical
Athens and contributing to topical debates in Ancient History. This work
shows how models like Ithaca can unlock the cooperative potential
between AI and historians, transformationally impacting the way we study
and write about one of the most significant periods in human history.”

   * Link to published paper (open access):
     <> and /Nature/'s
     cover issue


   * Link to Ithaca’s free online interface:

   * Link to the /Nature/ video:

   * Link to the open source code:

   * Ithaca's authors: Yannis Assael*, Thea Sommerschield*, Brendan
     Shillingford, Mahyar Bordbar, John Pavlopoulos, Marita
     Chatzipanagiotou, Ion Androutsopoulos, Jonathan Prag & Nando de Freitas

Thank you for circulating, and we look forward to your feedback on the tool!

Thea Sommerschield
Marie Skłodowska-Curie Fellow | Università Ca’ Foscari Venezia
Fellow in Hellenic Studies | CHS, Harvard University 

Unsubscribe at:
List posts to:
List info and archives at at:
Listmember interface at:
Subscribe at: