AlphaStar
E41648
AlphaStar is a DeepMind-created artificial intelligence system that achieved grandmaster-level performance in the real-time strategy game StarCraft II.
Observed surface forms (1)
| Surface form | Occurrences |
|---|---|
| AlphaStar (StarCraft II agent) | 1 |
Statements (48)
| Predicate | Object |
|---|---|
| instanceOf |
artificial intelligence system
ⓘ
game-playing AI ⓘ |
| achievedRank | Grandmaster ⓘ |
| achievedRankIn | StarCraft II European ladder ⓘ |
| achievedRankRegion | Europe ⓘ |
| announcedBy | DeepMind ⓘ |
| announcedIn | 2018 ⓘ |
| basedOnGame | StarCraft II ⓘ |
| collaborator | Blizzard Entertainment ⓘ |
| defeatedPlayer |
MaNa
ⓘ
TLO ⓘ |
| developer |
DeepMind
ⓘ
DeepMind ⓘ
surface form:
Google DeepMind
|
| developerHeadquarters |
London, England
ⓘ
surface form:
London
|
| evaluationSetting | Blizzard ladder environment ⓘ |
| field |
artificial intelligence
ⓘ
deep reinforcement learning ⓘ machine learning ⓘ |
| gameGenre | real-time strategy ⓘ |
| handles |
long-term planning
ⓘ
partial observability ⓘ real-time decision making ⓘ |
| notableFor |
defeating professional StarCraft II players
ⓘ
playing StarCraft II at grandmaster level ⓘ real-time strategy game performance ⓘ |
| organizationTypeOfDeveloper | AI research lab ⓘ |
| platform |
IBM PC
ⓘ
surface form:
PC
|
| playsRace |
Protoss
ⓘ
Terran ⓘ Zerg ⓘ |
| publiclyRevealedIn | 2019 ⓘ |
| relatedTo |
AlphaGo
ⓘ
AlphaZero ⓘ |
| researchArea |
imperfect information games
ⓘ
multi-agent systems ⓘ sequential decision making ⓘ |
| trainingMethod |
league-based training
ⓘ
reinforcement learning via self-play ⓘ supervised learning from human replays ⓘ |
| usesAlgorithm |
deep neural networks
ⓘ
imitation learning ⓘ multi-agent reinforcement learning ⓘ reinforcement learning ⓘ self-play ⓘ |
| usesComponent |
LSTM
ⓘ
attention mechanism ⓘ neural network policy ⓘ value network ⓘ |
Referenced by (2)
Full triples — surface form annotated when it differs from this entity's canonical label.
this entity surface form:
AlphaStar (StarCraft II agent)