Emergence of Linguistic Communication from Referential Games with Symbolic and Pixel Input

Abstract

The ability of algorithms to evolve or learn (compositional) communication protocols has traditionally been studied in the language evolution literature through the use of emergent communication tasks. Here we scale up this research by using contemporary deep learning methods and by training reinforcement-learning neural network agents on referential communication games. We extend previous work, in which agents were trained in symbolic environments, by developing agents which are able to learn from raw pixel data, a more challenging and realistic input representation. We find that the degree of structure found in the input data affects the nature of the emerged protocols, and thereby corroborate the hypothesis that structured compositional language is most likely to emerge when agents perceive the world as being structured.

arXiv commentTo appear at ICLR 2018

References in corpus (5)

• Grounded Language Learning in a Simulated 3D World
• Emergent Complexity via Multi-Agent Competition
• Emergence of Language with Multi-agent Games: Learning to Communicate with Sequences of Symbols
• Emergent Communication in a Multi-Modal, Multi-Step Referential Game
• Analogs of Linguistic Structure in Deep Representations

Cited by in corpus (15)

• Dark, Beyond Deep: A Paradigm Shift to Cognitive AI with Humanlike Common Sense
• Emergent social conventions and collective bias in LLM populations
• Developmentally motivated emergence of compositional communication via template transfer
• Multi-agent Deep Reinforcement Learning with Extremely Noisy Observations
• Emergent Language: A Survey and Taxonomy
• Measuring non-trivial compositionality in emergent communication
• Towards Emergent Language Symbolic Semantic Segmentation and Model Interpretability
• A perspective on multi-agent communication for information fusion
• Symbolic Semantic Segmentation and Interpretation of COVID-19 Lung Infections in Chest CT volumes based on Emergent Languages
• From Language Games to Drawing Games
• Pommerman: A Multi-Agent Playground
• Expressivity of Emergent Language is a Trade-off between Contextual Complexity and Unpredictability
• Multi-Agent Reinforcement Learning as a Computational Tool for Language Evolution Research: Historical Context and Future Challenges
• Emergence of Writing Systems Through Multi-Agent Cooperation
• The emergence of visual semantics through communication games