Statements (22)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:AI_research_project
|
| gptkbp:announced |
2018
|
| gptkbp:architecture |
LSTM neural networks
|
| gptkbp:citation |
https://arxiv.org/abs/1912.06680
|
| gptkbp:crew |
5 AI agents
|
| gptkbp:demonstrates |
gptkb:The_International_2018
|
| gptkbp:developedBy |
gptkb:OpenAI
|
| gptkbp:focusesOn |
gptkb:Dota_2
|
| gptkbp:goal |
to defeat human professional Dota 2 teams
|
| gptkbp:language |
gptkb:Python
|
| gptkbp:notableAchievement |
defeated OG (Dota 2 world champions) in April 2019
|
| gptkbp:notableEvent |
OpenAI Five Arena (public matches against humans)
|
| gptkbp:notableFor |
first AI to defeat world champions in a complex esport
|
| gptkbp:openSource |
partial code and research
|
| gptkbp:relatedTo |
gptkb:OpenAI_Gym
gptkb:OpenAI_Codex |
| gptkbp:trainer |
self-play
|
| gptkbp:uses |
deep reinforcement learning
|
| gptkbp:website |
https://openai.com/research/openai-five
|
| gptkbp:bfsParent |
gptkb:OpenAI,_Inc.
|
| gptkbp:bfsLayer |
7
|
| https://www.w3.org/2000/01/rdf-schema#label |
OpenAI Five Research
|