PAP_NER: A large-scale vietnamese administrative named entity recognition corpus and hybrid deep learning architecture
Dinh-Dien La, Tien-Bang Tran, Ngoc-Huy Du, Ngoc-Hung Dang, Trung-Nghia Phung, Van-Khanh Tran · 2026
Read PAP_NER: A large-scale vietnamese administrative named entity recognition corpus and hybrid deep learning architecture by Dinh-Dien La, Tien-Bang Tran, Ngoc-Huy Du, Ngoc-Hung Dang, Trung-Nghia Phung, Van-Khanh Tran (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0353166.txt.
Description
La and colleagues, writing in 2026, introduce a new language resource for Vietnamese government work. Their team built a large set of labeled administrative texts, meant to teach computers to find named items in official papers. Older Vietnamese data sets drew mostly on news or health texts, which left a gap for formal state documents.
The new corpus holds a very large number of sentences marked for five kinds of items, such as agencies, legal papers, dates, and amounts. Human checkers agreed closely on the labels, which supports trust in the data. The authors also test a model that pairs a Vietnamese language model with a step that keeps labels in a sensible order. That model beat older methods on their tests, and the data set is shared in public for other teams to use.
Catalog record
| CREATOR | Dinh-Dien La, Tien-Bang Tran, Ngoc-Huy Du, Ngoc-Hung Dang, Trung-Nghia Phung, Van-Khanh Tran |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 92397 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |