Emergent Communication
A research branch of multi-agent reinforcement learning in which multiple agents, cooperating under limited communication bandwidth (for example, being able to send only discrete symbols), spontaneously evolve an internally consistent symbol system to transmit information efficiently.
Timeline (last 90 days)
Emergent communication is a research branch in multi-agent reinforcement learning, referring to agents spontaneously evolving an internally consistent symbol system when the communication channel is limited.
In classic experiments on emergent communication, the narrower the communication channel, the more agents tend to compress semantics and discard redundancy, forming highly contextualized 'argot.'
The phenomenon of participants in a variety show developing a minimalist collaborative language is similar to the emergent communication phenomenon in AI research and can be seen as a miniature simulation of the human version.
All Facts (3)
Emergent communication is a research branch in multi-agent reinforcement learning, referring to agents spontaneously evolving an internally consistent symbol system when the communication channel is limited.
50%UnverifiedIn classic experiments on emergent communication, the narrower the communication channel, the more agents tend to compress semantics and discard redundancy, forming highly contextualized 'argot.'
50%UnverifiedThe phenomenon of participants in a variety show developing a minimalist collaborative language is similar to the emergent communication phenomenon in AI research and can be seen as a miniature simulation of the human version.
50%