OpenAI Five
A Dota 2 game-playing AI developed by OpenAI, trained using reinforcement learning in a closed multi-agent environment. OpenAI Five is a representative achievement in multi-agent game AI, demonstrating the application of large-scale reinforcement learning to complex, real-time strategy games.
Core Facts
Timeline (last 90 days)
PPO is one of the underlying algorithmic choices for large-scale game AI projects such as OpenAI Five (Dota 2 AI) and AlphaStar.
OpenAI's Gym, DeepMind's Atari challenge, and Pokémon-series reinforcement learning experiments have long served as testing grounds in the field of reinforcement learning
DeepMind's AlphaStar, OpenAI Five, and other milestone projects are built on top of games.
OpenAI Five defeated Dota 2 world champions OG in 2019
Training the Dota 2 AI for OpenAI Five used 256 GPUs and 128,000 CPU cores, consuming hundreds of petaflop-days of compute in total.
Milestones in modern game AI research include DeepMind's DQN demonstrating deep reinforcement learning on Atari games in 2013, AlphaGo defeating the Go world champion in 2016, OpenAI Five defeating professional teams in Dota 2 in 2019, and AlphaStar reaching Grandmaster level in StarCraft II in 2019.
OpenAI Five在Dota 2中累计了超过45000年的游戏经验
OpenAI Five用强化学习训练出能击败职业电竞选手的Dota 2 AI
All Facts (10)
After the integration of deep reinforcement learning, AlphaGo defeating the human Go champion and OpenAI Five became its landmark applications.
75%UnverifiedPPO is one of the underlying algorithmic choices for large-scale game AI projects such as OpenAI Five (Dota 2 AI) and AlphaStar.
50%UnverifiedOpenAI's Gym, DeepMind's Atari challenge, and Pokémon-series reinforcement learning experiments have long served as testing grounds in the field of reinforcement learning
50%UnverifiedDeepMind's AlphaStar, OpenAI Five, and other milestone projects are built on top of games.
50%UnverifiedTraining the Dota 2 AI for OpenAI Five used 256 GPUs and 128,000 CPU cores, consuming hundreds of petaflop-days of compute in total.
50%UnverifiedMilestones in modern game AI research include DeepMind's DQN demonstrating deep reinforcement learning on Atari games in 2013, AlphaGo defeating the Go world champion in 2016, OpenAI Five defeating professional teams in Dota 2 in 2019, and AlphaStar reaching Grandmaster level in StarCraft II in 2019.
50%UnverifiedOpenAI Five在Dota 2中累计了超过45000年的游戏经验
50%UnverifiedOpenAI Five用强化学习训练出能击败职业电竞选手的Dota 2 AI
50%UnverifiedDeepMind的AlphaProof直接在Lean4环境中进行强化学习
50%UnverifiedOpenAI Five defeated Dota 2 world champions OG in 2019
50%