Evolving neural networks

摘要: Neuroevolution, i.e. evolution of artificial neural networks, has recently emerged as a powerful technique for solving challenging reinforcement learning problems. Compared to traditional (e.g. value-function based) methods, neuroevolution is especially strong in domains where the state world not fully known: The can be disambiguated through recurrency, and novel situations handled pattern matching. In this tutorial, I will review (1) methods that evolve fixed-topology network topologies, construction processes, (2) ways combining algorithms with evolutionary (3) applications control, robotics, life, games.