Augmented Close Reading for Classical Latin
Intertextuality is a critical literary concept in the analysis of Latin texts from classical antiquity. This project contributes a reading interface that combines Latin BERT (Bamman and Burns 2020), a BERT model trained on classical Latin, with design choices from popular online Latin reading platforms, so readers can select a word in a passage and surface semantically similar uses across other texts.
The accompanying paper was accepted to NLP4DH 2025, co-located with NAACL.
Forsan et haec olim meminisse iuvabit.
Vergil, Aeneid 1.203 (AI has progressed so far as to recommend epigraphs for me now :0)
Status
The interactive tool is offline while the project is paused. The code, model, and documentation are preserved, and it may return in an updated form.
Read more
- Source code Snapshot of the thesis code, shared for research
- Paper NLP4DH 2025, co-located with NAACL (ACL Anthology)
- Latin BERT Bamman and Burns (2020), the model behind the search