[R4830] "Can You Read This One?" New Advances and Issues in Digital Recognition of Printed/Handwritten Arabic Characters

Created by M. Safa Saracoglu
Tuesday, 11/21/17 8:00am


This round-table will generate a discussion among scholars who work on digital recognition of printed/hand-written Ottoman and Arabic texts through machine learning. The purpose is to highlight particular technical challenges the researchers are facing to generate a platform on the close connection of technical issues with historical projects. Limited attention to commonality of certain issues that the computer-scientist collaborators of digital humanities projects face is a pressing problem in our field. Historians tend to dominate topic-focused debates as "technical issues" take less emphasis. By modeling a methodology-focused debate around concrete problems we encounter in text-recognition, this round-table will bring technicians and historians involved in advanced projects and let the former group play the lead role in identifying specific topics to discuss.

With this particular design, this round-table will achieve two things:

1. The "humanities" half of the projects can pay more attention to their "digital" partners as they discuss specific machine-learning issues and identify common barriers, bottlenecks and best practices those different projects share. The round-table will focus on already established projects because they present a unique opportunity: people involved in them have already spent extensive amounts of time on "problem solving" operations that provide solutions to the particular demands of each project. This is usually because the technical side of these projects respond to the demands of the historians. Quite often, the historians are not focusing on the technical questions as primary issues although they have some understanding of them. Shifting the focus on concrete machine-learning aspects of these problems can help us connect different projects around technical axes rather than textual ones and see if such common mathematical problems are suitable to seek support from grant-giving organizations that focus on "natural sciences"--NSF etc.

2. Participants can discuss methods to handle "the disconnect" between the two sides of these projects: Historians and computer scientists may find it quite agonizing to communicate with each other at the beginning of such projects. After repeated misunderstandings and prolonged silences spent on understanding the meaning of the other sides' disciplinary speak, we, the participants of such projects come to an understanding about each other's concerns. Such communication takes time and a round-table discussion can help identify best practices to help eliminate misunderstandings and frustrations in communication.

Problem-oriented conversation among established teams can help us identify guidelines that can benefit the larger scholarly community and explore new problems and funding sources.





M. Safa Saracoglu

(Bloomsburg University)
Panel Participating Role(s): Chair; Organizer; Presenter;

Maxim Romanov

(Leipzig University)
Maxim Romanov is a Universitätsassistent für Digital Humanities at the Institute for History, University of Vienna. His dissertation (Near Eastern Studies, U of Michigan, 2013) explored how modern computational techniques of text analysis can be applied...
Panel Participating Role(s): Presenter;

M. Erdem Kabadayi

(Koç University, Istanbul)
Panel Participating Role(s): Presenter;

Mehmet Can Yavuz

(Koç University)
Panel Participating Role(s): Presenter;

Kursat Aker

(Miletos R&D - METU)
Panel Participating Role(s): Presenter;