View in EDS HTML Full Text PDF Full Text

Human-level performance in 3D multiplayer games with population-based reinforcement learning.

Saved in:
Bibliographic Details
Title: Human-level performance in 3D multiplayer games with population-based reinforcement learning.
Authors: Jaderberg, Max (AUTHOR), Czarnecki, Wojciech M. (AUTHOR), Dunning, Iain (AUTHOR), Marris, Luke (AUTHOR), Lever, Guy (AUTHOR), Castañeda, Antonio Garcia (AUTHOR), Beattie, Charles (AUTHOR), Rabinowitz, Neil C. (AUTHOR), Morcos, Ari S. (AUTHOR), Ruderman, Avraham (AUTHOR), Sonnerat, Nicolas (AUTHOR), Green, Tim (AUTHOR), Deason, Louise (AUTHOR), Leibo, Joel Z. (AUTHOR), Silver, David (AUTHOR), Hassabis, Demis (AUTHOR), Kavukcuoglu, Koray (AUTHOR), Graepel, Thore (AUTHOR)
Source: Science (pre-March 2025). 5/31/2019, Vol. 364 Issue 6443, p859-865. 7p. 4 Diagrams.
Subjects: Reinforcement learning, Multiplayer games, Agent (Philosophy), Pixels, Multiagent systems, Artificial intelligence research, Video games
Abstract: Reinforcement learning (RL) has shown great success in increasingly complex single-agent environments and two-player turn-based games. However, the real world contains multiple agents, each learning and acting independently to cooperate and compete with other agents. We used a tournament-style evaluation to demonstrate that an agent can achieve human-level performance in a three-dimensional multiplayer first-person video game, Quake III Arena in Capture the Flag mode, using only pixels and game points scored as input. We used a two-tier optimization process in which a population of independent RL agents are trained concurrently from thousands of parallel matches on randomly generated environments. Each agent learns its own internal reward signal and rich representation of the world. These results indicate the great potential of multiagent reinforcement learning for artificial intelligence research. [ABSTRACT FROM AUTHOR]
Copyright of Science (pre-March 2025) is the property of American Association for the Advancement of Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Psychology and Behavioral Sciences Collection
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
Text:
  Availability: 1
Header DbId: pbh
DbLabel: Psychology and Behavioral Sciences Collection
An: 136815122
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Human-level performance in 3D multiplayer games with population-based reinforcement learning.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Jaderberg%2C+Max%22">Jaderberg, Max</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Czarnecki%2C+Wojciech+M%2E%22">Czarnecki, Wojciech M.</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Dunning%2C+Iain%22">Dunning, Iain</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Marris%2C+Luke%22">Marris, Luke</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Lever%2C+Guy%22">Lever, Guy</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Castañeda%2C+Antonio+Garcia%22">Castañeda, Antonio Garcia</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Beattie%2C+Charles%22">Beattie, Charles</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Rabinowitz%2C+Neil+C%2E%22">Rabinowitz, Neil C.</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Morcos%2C+Ari+S%2E%22">Morcos, Ari S.</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Ruderman%2C+Avraham%22">Ruderman, Avraham</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Sonnerat%2C+Nicolas%22">Sonnerat, Nicolas</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Green%2C+Tim%22">Green, Tim</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Deason%2C+Louise%22">Deason, Louise</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Leibo%2C+Joel+Z%2E%22">Leibo, Joel Z.</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Silver%2C+David%22">Silver, David</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Hassabis%2C+Demis%22">Hassabis, Demis</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Kavukcuoglu%2C+Koray%22">Kavukcuoglu, Koray</searchLink> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Graepel%2C+Thore%22">Graepel, Thore</searchLink> (AUTHOR)
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Science+%28pre-March+2025%29%22">Science (pre-March 2025)</searchLink>. 5/31/2019, Vol. 364 Issue 6443, p859-865. 7p. 4 Diagrams.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Reinforcement+learning%22">Reinforcement learning</searchLink><br /><searchLink fieldCode="DE" term="%22Multiplayer+games%22">Multiplayer games</searchLink><br /><searchLink fieldCode="DE" term="%22Agent+%28Philosophy%29%22">Agent (Philosophy)</searchLink><br /><searchLink fieldCode="DE" term="%22Pixels%22">Pixels</searchLink><br /><searchLink fieldCode="DE" term="%22Multiagent+systems%22">Multiagent systems</searchLink><br /><searchLink fieldCode="DE" term="%22Artificial+intelligence+research%22">Artificial intelligence research</searchLink><br /><searchLink fieldCode="DE" term="%22Video+games%22">Video games</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Reinforcement learning (RL) has shown great success in increasingly complex single-agent environments and two-player turn-based games. However, the real world contains multiple agents, each learning and acting independently to cooperate and compete with other agents. We used a tournament-style evaluation to demonstrate that an agent can achieve human-level performance in a three-dimensional multiplayer first-person video game, Quake III Arena in Capture the Flag mode, using only pixels and game points scored as input. We used a two-tier optimization process in which a population of independent RL agents are trained concurrently from thousands of parallel matches on randomly generated environments. Each agent learns its own internal reward signal and rich representation of the world. These results indicate the great potential of multiagent reinforcement learning for artificial intelligence research. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Science (pre-March 2025) is the property of American Association for the Advancement of Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=pbh&AN=136815122
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1126/science.aau6249
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 7
        StartPage: 859
    Subjects:
      – SubjectFull: Reinforcement learning
        Type: general
      – SubjectFull: Multiplayer games
        Type: general
      – SubjectFull: Agent (Philosophy)
        Type: general
      – SubjectFull: Pixels
        Type: general
      – SubjectFull: Multiagent systems
        Type: general
      – SubjectFull: Artificial intelligence research
        Type: general
      – SubjectFull: Video games
        Type: general
    Titles:
      – TitleFull: Human-level performance in 3D multiplayer games with population-based reinforcement learning.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Jaderberg, Max
      – PersonEntity:
          Name:
            NameFull: Czarnecki, Wojciech M.
      – PersonEntity:
          Name:
            NameFull: Dunning, Iain
      – PersonEntity:
          Name:
            NameFull: Marris, Luke
      – PersonEntity:
          Name:
            NameFull: Lever, Guy
      – PersonEntity:
          Name:
            NameFull: Castañeda, Antonio Garcia
      – PersonEntity:
          Name:
            NameFull: Beattie, Charles
      – PersonEntity:
          Name:
            NameFull: Rabinowitz, Neil C.
      – PersonEntity:
          Name:
            NameFull: Morcos, Ari S.
      – PersonEntity:
          Name:
            NameFull: Ruderman, Avraham
      – PersonEntity:
          Name:
            NameFull: Sonnerat, Nicolas
      – PersonEntity:
          Name:
            NameFull: Green, Tim
      – PersonEntity:
          Name:
            NameFull: Deason, Louise
      – PersonEntity:
          Name:
            NameFull: Leibo, Joel Z.
      – PersonEntity:
          Name:
            NameFull: Silver, David
      – PersonEntity:
          Name:
            NameFull: Hassabis, Demis
      – PersonEntity:
          Name:
            NameFull: Kavukcuoglu, Koray
      – PersonEntity:
          Name:
            NameFull: Graepel, Thore
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 31
              M: 05
              Text: 5/31/2019
              Type: published
              Y: 2019
          Identifiers:
            – Type: issn-print
              Value: 00368075
          Numbering:
            – Type: volume
              Value: 364
            – Type: issue
              Value: 6443
          Titles:
            – TitleFull: Science (pre-March 2025)
              Type: main
ResultId 1