chatbots, open domain dialog agents, reinforcement learning, neural memory, continual learning