DeepMind’s AlphaStar Benchmark Improves RL Offline Agent With 90% Win Rate Against SOTA AlphaStar Supervised Agent

In a new paper AlphaStar Unplugged: Large-Scale Offline Reinforcement Learning, a DeepMind research team presents AlphaStar Unplugged, an unprecedented challenging large-scale offline reinforcement learning benchmark that leverages a offline dataset from StarCraft II for RL agents training.

agents ai artificial intelligence benchmark dataset deepmind deepmind research deep-neural-networks machine learning machine learning & data science ml offline paper rate reinforcement reinforcement learning research research team scale sota team technology training