Google DeepMind Researchers Apply Semantic Evolution to Create Non Intuitive VAD-CFR and SHOR-PSRO Variants for Superior Algorithmic Convergence

Google DeepMind researchers use semantic evolution to develop non-intuitive variants of CFR and PSRO algorithms, achieving superior convergence in multi-agent reinforcement learning. Rather than manually refining algorithms through trial-and-error, the team automated algorithm discovery, navigating vast spaces of update rules. This breakthrough shifts MARL research beyond human intuition limitations.