We consider the task of reinforcement learning in an environment in which rare significant events occur independently of the actions selected by the controlling agent. If these ev...
A new graph-based evolutionary algorithm named “Genetic Network Programming, GNP” has been proposed. GNP represents its solutions as directed graph structures, which can improv...
odel of Embodiment on Abstract Systems: from Hierarchy to Heterarchy Kohei Nakajima, Soya Shinkai, Takashi Ikegami A Behavior-Based Model of the Hydra, Phylum Cnidaria Malin Aktius...
A control of real processes requires different approach to neural network learning. The presented modification of backpropagation learning algorithm changes a meaning of learning...
Proactive assessment of computer-network vulnerability to unknown future attacks is an important but unsolved computer security problem where AI techniques have significant impact...