EvaCUN2023

• Evaluation of Natural Language Processing (NLP) tools for the ancient Cuneiform languages (3,400 BCE - 75 CE)
• Machine translation from ancient Akkadian to either modern English, Hebrew or Arabic
• Co-operated with ALT2023

Important Dates

• CFP in Dec. 2022 (depending on ALT's schedule)
• Registration for the shared task: January 1, 2023 ~ March 1, 2023
• Training data release: April 1, 2023
• Test data release: TBD
• Data submission : TBD
• Tech report submission: TBD

Task

Cuneiform is one of the earliest writing systems in recorded human history (ca. 3,400 BCE-75 CE). Hundreds of thousands of such texts were found over the last two centuries in the Middle East. Most of these texts are found on a clay or stone medium, and are written in Sumerian and Akkadian, beside relatively smaller corpora (still in the tens of thousands) in Elamite, Eblaite, Hittite, Hurrian, Urartian, Hattian, and Luwian, as well as languages which use alphabetic Cuneiform like Ugaritic and Old Persian.

Tracks

1. Translation from any ancient Cuneiform language into modern English
2. Translation from any ancient Cuneiform language into modern Hebrew
3. Translation from any ancient Cuneiform language into modern Arabic

Participants

Organizers

• Adam Anderson, Data Science Discovery Partner, UC Berkeley, California
• Shai Gordin, Digital Pasts Lab, Ariel University, Israel

Student Members

• Stav Klein, Digital Pasts Lab, Ariel University, Israel
• Morris Alper, School of Electrical Engineering, Tel-Aviv University, Israel
• Luis Sáenz, Seminar für Sprachen und Kulturen des Vorderen Orients, Heidelberg University
• Zaid Maayah, College of Letters & Science, Data Science Major, UC Berkeley, California
• Tina Chen, College of Letters & Science, Data Science Major, UC Berkeley, California

Data

Test Objective