Distillation of a Locally Deployable 8B Language Model for Lung Histology Abstraction: A 33-Report Proof-of-Concept Study
Dian Tri Wiyanti, Min-Huei Hsu, Daniel C A Nugroho · 2026
Read Distillation of a Locally Deployable 8B Language Model for Lung Histology Abstraction: A 33-Report Proof-of-Concept Study by Dian Tri Wiyanti, Min-Huei Hsu, Daniel C A Nugroho (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/pmc-13588119.txt.
Description
Dian Tri Wiyanti, Min-Huei Hsu, and Daniel Nugroho shrink a big AI reader into a small one a hospital can run on site, in a 2026 study. Cancer registries need facts pulled from lung reports, a slow hand task that lags care. Cloud models read well but send private records to outside servers. The team taught a compact model using a cloud model's worked answers as lessons.
A six-tier rule frame, drawn from world cancer standards, steadied the small model's logic. Training ran in one pass on a home-grade graphics card, using thousands of public records. The setup then ran on real hospital records in Taipei, and experts blind-checked 33 reports closely. The authors offer the pilot as proof that private, local reading of reports can work without sending data away.
Catalog record
| CREATOR | Dian Tri Wiyanti, Min-Huei Hsu, Daniel C A Nugroho |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 57 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |