Riccardi, K., & Bandara, D. (2024, April), Masked Language Modeling for Predicting Missing Words in Damaged Ancient Greek Texts  Paper presented at 2024 ASEE North East Section, Fairfield, Connecticut. 10.18260/1-2--45774

\bibitem{asee_peer_45774}
  Riccardi, K., \& Bandara, D. (2024, April), \emph{Masked Language Modeling for Predicting Missing Words in Damaged Ancient Greek Texts}
  Paper presented at 2024 ASEE North East Section, Fairfield, Connecticut.
  10.18260/1-2--45774

Kyle Riccardi and Danushka Bandara.  "Masked Language Modeling for Predicting Missing Words in Damaged Ancient Greek Texts".  2024 ASEE North East Section, Fairfield, Connecticut, 2024, April.  ASEE Conferences, 2024.  https://peer.asee.org/45774 Internet.  15 Dec, 2024

\bibitem{asee_peer_45774}
  Kyle Riccardi and Danushka Bandara.
  "Masked Language Modeling for Predicting Missing Words in Damaged Ancient Greek Texts".
  \emph{2024 ASEE North East Section, Fairfield, Connecticut, 2024, April}.
  ASEE Conferences, 2024.
  https://peer.asee.org/45774 Internet.  15 Dec, 2024

@INPROCEEDINGS{asee_peer_45774
author = "Kyle Riccardi and Danushka Bandara"
title = "Masked Language Modeling for Predicting Missing Words in Damaged Ancient Greek Texts"
booktitle = "2024 ASEE North East Section"
year = "2024"
month = "April"
address = "Fairfield, Connecticut"
publisher = "ASEE Conferences"
note = {https://peer.asee.org/45774}
number = {10.18260/1-2--45774}
}

TY  - CPAPER
AB  - The ancient Greek texts are valuable for understanding and learning about the history, culture, and nuances of ancient Greek life. The texts come in many forms, including papyri, fragments of pottery, etc. Due to the nature of these materials and degradation over time, some of these texts are missing words, and even entire phrases. This makes it difficult for historians to interpret the texts. The data for this project was collected from the Perseus Collection and the 1KGreek collection, which contains  250,000 unique sentences of ancient Greek literature. The dataset was preprocessed using the import classical language toolkit(CLTK) and sentences were normalized for better encodings. After the encoding was done all of our data was split by sentences and then they were fed into a Disil Bert masked language model. The word piece tokenizer for this model was trained using a vocabulary list of 35,000 words. By using the DistilBert transformer model we were able to train a masked language model based on words to achieve a Hit@5 of 34 percent, Hit@10 of 35 percent, Hit@100 of 36 percent, and a perplexity of 1.04. This model can be a valuable aid for the historians' workflow in deciphering damaged ancient texts.
AU  - Kyle Riccardi
AU  - Danushka Bandara
CY  - Fairfield, Connecticut
DA  - 2024/04/19
PB  - ASEE Conferences
TI  - Masked Language Modeling for Predicting Missing Words in Damaged Ancient Greek Texts
UR  - https://peer.asee.org/45774
DO  - 10.18260/1-2--45774
ER  -