##article.return## Symmetry Heuristics for Stable Reinforcement Learning Design Agents Download Download PDF