Artificial General Intelligence, Open-endedness, Self-Improvement, Reinforcement Learning
[Google Scholar] [dblp] [Semantic Scholar]