The Ecosystem Path to General AI

We start by discussing the link between ecosystem simulators and general AI. Then we present the open-source ecosystem simulator Ecotwin, which is based on the game engine Unity and operates on ecosystems containing inanimate objects like mountains and lakes,\ua0as well as organisms such as animals and plants. Animal cognition is\ua0modeled by integrating three separate networks: (i) a reflex network for\ua0hard-wired reflexes; (ii) a happiness network that maps sensory data such\ua0as oxygen, water, energy, and smells, to a scalar happiness value; and (iii)\ua0a policy network for selecting actions. The policy network is trained with\ua0reinforcement learning (RL), where the reward signal is defined as the\ua0happiness difference from one time step to the next. All organisms are\ua0capable of either sexual or asexual reproduction, and they die if they run out of critical resources. We report results from three studies with\ua0Ecotwin, in which natural phenomena emerge in the models without being\ua0hardwired. First, we study a terrestrial ecosystem with wolves, deer,\ua0and grass, in which a Lotka-Volterra style population dynamics emerges.Second, we study a marine ecosystem with phytoplankton, copepods, and krill, in which a diel vertical migration behavior emerges. Third, we\ua0study an ecosystem involving lethal dangers, in which certain agents that combine RL with reflexes outperform pure RL agents