NLP Wiki

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== Learning to Communicate ======
There are some papers on artifical agents learning to communicate to play a game, etc.  They either learn a language from scratch (an emergent language) or use reinforcement learning to improve an existing model of language.

===== Overviews =====
  * Multi-agent communication
    * [[https://link.springer.com/article/10.1007/s10462-021-09996-w|2020 - Multi-agent deep reinforcement learning: a survey]]

===== Papers =====
  * **Multi-agent**
    * [[https://proceedings.neurips.cc/paper_files/paper/2017/file/70222949cc0db89ab32c9969754d4758-Paper.pdf|Havrylov & Titov 2017 - Emergence of Language with Multi-agent Games: Learning to Communicate with Sequences of Symbols]]
    * [[https://aclanthology.org/P19-1509.pdf|2019 - Word-order biases in deep-agent emergent communication]]
    * [[https://arxiv.org/pdf/2006.02419.pdf|2020 - Emergent Multi-Agent Communication in the Deep Learning Era]]
    * [[https://arxiv.org/pdf/2012.08630.pdf|2020 - Open Problems in Cooperative AI]]
  * **Emergence of Grammar**
    * [[https://arxiv.org/pdf/2503.01635?|Wechsler et al 2025 - The Emergence of Grammar through Reinforcement Learning]]

===== Related Pages =====
  * [[Dialog]]
  * [[Lifelong Learning]]
  * [[ml:Reinforcement Learning]]