A Fine-Tuned BERT Classifier for Personal-Letter Titles in Late-Ming and Early-Qing Collected Works
arXiv:2605.23103v1 Announce Type: cross Abstract: I present Lepton (Letter Prediction), a fine-tuned BERT classifier that predicts whether a title in a Classical Chinese wenji table of contents is a personal letter or a closely confusable preface (particularly the farewell-preface). Lepton fine-tunes bert-base-chinese on 5438 hand-labeled wenji titles from thirty-three late-Ming and early-Qing literati. I've deployed the model on Hugging Face and has been used at the China Biographical Database (CBDB) to identify approximately fifty-five thousand letters across mid-Ming through early-Qing wenj
This development leverages advancements in NLP and BERT models to address a specific historical and archival challenge.
It demonstrates how fine-tuned AI can be applied to niche academic fields, enhancing data discovery and research efficiency in humanities.
A new tool for historical researchers in Classical Chinese studies now significantly improves the identification of personal letters within large textual corpora.
- · Historians of Late-Ming and Early-Qing China
- · Digital humanities researchers
- · Institutions like China Biographical Database (CBDB)
- · Manual archival research methods
Historians can more rapidly identify and analyze personal correspondence from specific historical periods.
The increased availability of identified letters could lead to new insights into social networks, personal relationships, and intellectual trends of the Late-Ming and Early-Qing eras.
The methodology could be replicated for other historical languages or document types, accelerating digital cataloging and analysis across various cultural heritage domains.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI