Reproducibility and Transparency in AI
In a new perspective piece “Transparency and reproducibility in artificial intelligence” published in the journal Nature, an international group of scientists including AIM investigators discusses the need for reproducible and transparent AI research.
The authors use a recent publication in Nature as an example, which investigated the use of artificial intelligence to diagnose breast cancer from mammogram images more accurately than expert human radiologists. As there are restrictive data access procedures, lack of published computer code, and unreported model parameters, this makes it practically impossible for any other researchers to confirm or extend this work.
AIM Investigators
