Why do some people get cancer,
and not others?
Researchers are looking to artificial intelligence for the answer. Hospitals, research labs and public health agencies hold extensive collections of valuable health data, such as patient records and results from clinical trials.
AI models can process this data to identify patterns that can help us understand cancer’s development and spread, potentially leading to new treatments or even cures.
However, many of these records sit in digital fortresses to protect patient privacy—which is important, but limits the opportunities for analysis.
Enter federated learning. It’s a machine learning technique that bypasses this problem by allowing AI models to learn from patient data—without the need to directly access or move it.
Below, explore how federated learning is already making breakthrough discoveries—from how to predict the severity of a Covid-19 case to the gene that makes some astronauts more likely to develop cancer in space.