Label-Efficient Chest X-ray Diagnosis via Partial CLIP Adaptation
Heet Nitinkumar Dalsania · 2025
Read Label-Efficient Chest X-ray Diagnosis via Partial CLIP Adaptation by Heet Nitinkumar Dalsania (2025) 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/arxiv-2507-07254.txt.
Description
Chest X-rays are cheap and common, but the expert labels that teach a model to read them are dear and scarce. Heet Dalsania tackles that gap in 2025 with a model that already links images and words. Only part of its image side is tuned; the rest stays as trained.
The test bed is a known set of chest films with fourteen disease labels. The tuned model is tried with no labels at all, then with one to sixteen labelled films per class. Mean scores rose by more than twenty percent over the no-label base. The setup mimics a hospital's real shelf: stacks of films, few labels. A method that learns from a handful of marks could serve rare diseases most.
Catalog record
| CREATOR | Heet Nitinkumar Dalsania |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 17 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-07254