IFC-HFlowVAE: A self-enhancing generative framework with structural anchoring for imbalanced clinical data augmentation
Lu Yuwen, Shuyu Chen · 2026
Read IFC-HFlowVAE: A self-enhancing generative framework with structural anchoring for imbalanced clinical data augmentation by Lu Yuwen, Shuyu Chen (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-0357260.txt.
Description
Lu Yuwen and Shuyu Chen tackle a common headache in medical machine learning: rare disease groups have too few patient records to train a good model. Their 2026 framework invents extra records for the small group using a generative model, then improves those records in rounds. Borrowed real-style samples act as anchors so the invented data do not drift away from true patient patterns during the rounds.
Across seven clinical datasets, models trained with the boosted data sorted patients better than models using rival boosting methods. Checks also showed the invented records kept tricky traits of the rare group, such as skewed and multi-peaked value patterns. Yuwen and Chen offer the method for settings where collecting more real rare-case records is slow or impossible.
Catalog record
| CREATOR | Lu Yuwen, Shuyu Chen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47853 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |