Towards Open World Learning System

This thesis explores how to make AI systems learn and adapt on their own in a changing world. It develops new methods to help AI recognize unknown data, group it into meaningful categories, and add this new knowledge without forgetting what it already knows. By improving these three critical steps, the research aims to reduce the need for constant human input, making AI more autonomous and capable of handling real-world challenges. This work is important for creating AI systems that can independently learn and grow, improving their usefulness in everyday applications.