Self-learning AI

Self-learning AI refers to systems that iteratively improve using their own predictions or weak supervision signals. Approaches include self-training, pseudo-labeling, and continual learning with mechanisms to prevent drift and catastrophic forgetting.