Employing deep reinforcement learning, we investigate how cells may have evolved to optimize their movement by selectively suppressing certain pseudopod growth directions, especially in shallow, ...