test-time augmentation

MIT researchers have developed a new AI technique that improves prediction reliability and reduces uncertainty, making machine learning models more trustworthy for high-stakes fields like healthcare.