The first agent is primed with a start-up language and produces utterances., The next generation hears only a sample of utterances., A new agent forms a hypothesis about the language underlying the utterances it hears., The new agent produces its own utterances for the following generation., The process continues for as many generations as the simulation is designed to explore., A learning bottleneck may force the language to become more generalisable., The simulation measures how compositional the language has become.

Iterated learning model

由
列印
嵌入

排行榜

視覺風格

選項

切換範本

恢復自動保存: ?