Machine learningML & AI Methods

Adversarial Learning for Feature Shift Detection and Correction

The study investigates the use of adversarial learning to identify and correct feature shifts in datasets, demonstrating that it can outperform current statistical and neural network-based methods when combined with mainstream supervised classifiers.