Gender imbalance in medical imaging datasets produces biased classifiers for computer-aided diagnosis.
Larrazabal AJ, Nieto N, Peterson V, Milone DH, Ferrante E. · 2020
Read Gender imbalance in medical imaging datasets produces biased classifiers for computer-aided diagnosis. by Larrazabal AJ, Nieto N, Peterson V, Milone DH, Ferrante E. (2020) 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/europmc-pmc7293650.txt.
Description
Larrazabal and co-authors demonstrate that gender imbalance in medical image datasets produces biased diagnostic AI. Their study appeared in the Proceedings of the National Academy of Sciences. Artificial intelligence systems for reading scans are being adopted worldwide. The team tested what happens when training data contain far more images from one sex than the other.
Using three deep neural network designs and two public chest X-ray datasets for thoracic diseases, they varied the gender balance systematically. Performance consistently dropped for the underrepresented gender when a minimum balance was not met. The same algorithm could thus diagnose one sex worse than the other purely because of the data. The authors urge regulators to require gender balance and diversity checks before approving computer-aided diagnosis systems.
Catalog record
| CREATOR | Larrazabal AJ, Nieto N, Peterson V, Milone DH, Ferrante E. |
|---|---|
| YEAR | 2020 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |