How agents see things: On visual representations in an emergent language game

  title={How agents see things: On visual representations in an emergent language game},
  author={Diane Bouchacourt and M. Baroni},
  • Diane Bouchacourt, M. Baroni
  • Published in EMNLP 2018
  • Computer Science
  • There is growing interest in the language developed by agents interacting in emergent-communication settings. Earlier studies have focused on the agents' symbol usage, rather than on their representation of visual input. In this paper, we consider the referential games of Lazaridou et al. (2017) and investigate the representations the agents develop during their evolving interaction. We find that the agents establish successful communication by inducing visual representations that almost… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    Explore Further: Topics Discussed in This Paper

    Miss Tools and Mr Fruit: Emergent communication in agents learning about object affordances
    • 5
    • Open Access
    Anti-efficient encoding in emergent communication
    • 15
    • Open Access
    Word-order biases in deep-agent emergent communication
    • 7
    • Open Access
    Co-evolution of language and agents in referential games
    • 3
    • Open Access
    Avoiding hashing and encouraging visual semantics in referential emergent language games
    • 1
    • Open Access
    To Populate is To Regulate
    • 3
    • Open Access
    Information Minimization In Emergent Languages
    • 2


    Publications referenced by this paper.
    Multi-Agent Cooperation and the Emergence of (Natural) Language
    • 192
    • Open Access
    Emergence of Language with Multi-agent Games: Learning to Communicate with Sequences of Symbols
    • 108
    • Open Access
    Learning to Play Guess Who? and Inventing a Grounded Language as a Consequence
    • 39
    • Open Access
    Conceptual pacts and lexical choice in conversation.
    • 1,005
    • Open Access
    Emergent Translation in Multi-Agent Communication
    • 34
    • Open Access
    Visualizing and Understanding Convolutional Networks
    • 8,260
    • Open Access
    Categorical perception.
    • 237
    • Open Access
    Representational Similarity Analysis – Connecting the Branches of Systems Neuroscience
    • 1,653
    • Open Access