SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2109.04513 · EMNLP · 2021

Filling the Gaps in Ancient Akkadian Texts: A Masked Language Modelling Approach

Gabriel Stanovsky, Asaf Yehudai, Koren Lazar, Benny Saret, Wayne Horowitz, Nathan Wasserman

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 11 functions out of this paper's own repositories and ran 0 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
SLAB-NLP/Akk canonical 0 of 11
FunctionStatusWhere it lives
calc_wind_around_ind Not yet run SLAB-NLP/Akk/akkadian_bert/utils.py
code served (permissive licence) · get_code("dc0f42152bd71f09")
get_bert_path Not yet run SLAB-NLP/Akk/akkadian_bert/train_bert.py
code served (permissive licence) · get_code("85176e184025c2e0")
get_catalog Not yet run SLAB-NLP/Akk/RAW/get_langs.py
code served (permissive licence) · get_code("724cd694c6e42708")
get_concatenated_files Not yet run SLAB-NLP/Akk/akkadian_bert/utils.py
code served (permissive licence) · get_code("931f7b05fb434414")
get_enc_labels Not yet run SLAB-NLP/Akk/akkadian_bert/datasets_bert.py
code served (permissive licence) · get_code("26fa0b6bedc5dd1e")
get_files Not yet run SLAB-NLP/Akk/RAW/work_on_files.py
code served (permissive licence) · get_code("4f75983bc1a014c4")
get_language Not yet run SLAB-NLP/Akk/RAW/get_langs.py
code served (permissive licence) · get_code("983efcbf97af1179")
get_period Not yet run SLAB-NLP/Akk/RAW/get_langs.py
code served (permissive licence) · get_code("fb4bc8bf9816564c")
natural_number Not yet run SLAB-NLP/Akk/akkadian_bert/utils.py
code served (permissive licence) · get_code("5b9c23c30e55b67e")
read_file Not yet run SLAB-NLP/Akk/RAW/work_on_files.py
code served (permissive licence) · get_code("6b76badec7a176c5")
read_project Not yet run SLAB-NLP/Akk/RAW/work_on_files.py
code served (permissive licence) · get_code("621e7c8c7d006787")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We present models which complete missing text given transliterations of ancient Mesopotamian documents, originally written on cuneiform clay tablets (2500 BCE -100 CE). Due to the tablets' deterioration, scholars often rely on contextual cues to manually fill in missing parts in the text in a subjective and time-consuming process. We identify that this challenge can be formulated as a masked language modelling task, used mostly as a pretraining objective for contextualized language models. Following, we develop several architectures focusing on the Akkadian language, the lingua franca of the time. We find that despite data scarcity (1M tokens) we can achieve state of the art performance on missing tokens prediction (89% hit@5) using a greedy decoding scheme and pretraining on data from other languages and different time periods. Finally, we conduct human evaluations showing the applicability of our models in assisting experts to transcribe texts in extinct languages.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2109.04513")
get_code_for_paper("2109.04513")
have("2109.04513")

Connect an agent — have() is free.