Free Information Center
Residual reinforcement learning is an approach that combines model-based control methods with reinforcement learning by learning a residual policy to improve performance. It integrates prior knowledge and data-driven adaptation to enhance learning efficiency and robustness in complex environments.
MAPPO is an extension of Proximal Policy Optimization (PPO) designed for multi-agent environments, enhancing coordination among agents.
Batch-constrained Q-learning (BCQ) is an advanced reinforcement learning algorithm designed to optimize decision-making using limited data samples.
DeepNash is an artificial intelligence system designed to play Stratego, a complex strategy board game. It utilizes advanced game-theoretic and reinforcement learning techniques to achieve high-level play against human and AI opponents.
Behavior-regularized actor-critic (BRAC) is a reinforcement learning algorithm designed to enhance the stability and performance of policy learning through behavior regularization.
DeepMind is a British artificial intelligence company known for its advancements in machine learning and neural networks. Founded in 2010 and acquired by Google in 2015, it focuses on developing AI technologies with applications in various scientific and practical domains.
Information-theoretic exploration is a strategy in machine learning and artificial intelligence that guides agents to explore their environment by maximizing information gain. It leverages concepts from information theory to improve learning efficiency and decision-making in uncertain or unknown environments.
The StarCraft AI competition is an ongoing challenge in artificial intelligence research where AI agents compete in the real-time strategy game StarCraft. It serves as a benchmark for testing AI capabilities in complex decision-making and strategy.
Reinforcement learning from AI feedback (RLAIF) is a machine learning approach where reinforcement learning algorithms are guided by feedback generated from AI systems. This method enables models to improve based on evaluations provided by other AI agents rather than relying solely on human feedback.
Prioritized experience replay is a technique in reinforcement learning that improves learning efficiency by prioritizing important experiences.