Corpus ID: 3457087

Compositional Obverter Communication Learning From Raw Visual Input

@article{Choi2018CompositionalOC,
  title={Compositional Obverter Communication Learning From Raw Visual Input},
  author={E. Choi and A. Lazaridou and N. D. Freitas},
  journal={ArXiv},
  year={2018},
  volume={abs/1804.02341}
}
  • E. Choi, A. Lazaridou, N. D. Freitas
  • Published 2018
  • Computer Science
  • ArXiv
  • One of the distinguishing aspects of human language is its compositionality, which allows us to describe complex environments with limited vocabulary. Previously, it has been shown that neural network agents can learn to communicate in a highly structured, possibly compositional language based on disentangled input (e.g. hand- engineered features). Humans, however, do not learn to communicate based on well-summarized features. In this work, we train neural agents to simultaneously develop… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    Paper Mentions

    Learning by Abstraction: The Neural State Machine
    • 44
    • Open Access
    Measuring Compositionality in Representation Learning
    • 33
    • Open Access
    Compositional Languages Emerge in a Neural Iterated Learning Model
    • 3
    • Highly Influenced
    • Open Access
    Emergence of Compositional Language with Deep Generational Transmission
    • 14
    • Open Access
    Word-order biases in deep-agent emergent communication
    • 7
    • Open Access
    Developmentally motivated emergence of compositional communication via template transfer
    • 1
    • Open Access
    Emergent Language Generalization and Acquisition Speed are not tied to Compositionality
    • 2
    • Open Access
    Miss Tools and Mr Fruit: Emergent communication in agents learning about object affordances
    • 5
    • Open Access
    Emergent Multi-Agent Communication in the Deep Learning Era
    • 1
    • Open Access
    Emergence of Communication in an Interactive World with Consistent Speakers
    • 15
    • Highly Influenced
    • Open Access

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 28 REFERENCES
    Learning to Play Guess Who? and Inventing a Grounded Language as a Consequence
    • 39
    • Open Access
    Show and tell: A neural image caption generator
    • 3,394
    • Open Access
    Natural Language Does Not Emerge 'Naturally' in Multi-Agent Dialog
    • 91
    • Highly Influential
    • Open Access
    Multi-Agent Cooperation and the Emergence of (Natural) Language
    • 192
    • Open Access
    Deep visual-semantic alignments for generating image descriptions
    • 1,832
    • Open Access
    Emergence of Grounded Compositional Language in Multi-Agent Populations
    • 263
    • Highly Influential
    • Open Access
    Toward Controlled Generation of Text
    • 445
    • Open Access
    VQA: Visual Question Answering
    • 1,833
    • Open Access
    Learning to Communicate with Deep Multi-Agent Reinforcement Learning
    • 521
    • Open Access
    The emergence of linguistic structure: an overview of the iterated learning model
    • 334