Open-world continual learning: Unifying novelty detection and continual learning

Gyuhak Kim, Changnan Xiao, Tatsuya Konishi, Zixuan Ke, Bing Liu

Abstract: Highlights•This paper theoretically unifies novelty or out-of-distribution (OOD) detection, continual learning, and open world learning,•It proves that good within-task prediction and OOD detection are necessary and sufficient for successful open-world continual learning,•Several new algorithms have been designed for open-world continual learning based on the theory.