When Optimal is the Enemy of Good: High-Budget Differential Privacy for Medical AI | Towards Data Science

Can we guarantee patient privacy without sacrificing model accuracy?