[1Win4] Poster session 1

[1Win4-25] Evolutionary ecology of words

〇Reiji Suzuki¹, Takaya Arita¹ (1.Nagoya University)

Keywords:large language model, agent-based evolutionary model, artificial life

We propose an evolutionary ecology model of words that realizes virtual evolutionary and ecological processes by utilizing large language models (LLMs) for their creativity and ability to infer complex relationships between words. Each agent possesses a short word (or phrase) generated by an LLM and moves within a 2D environment. When agents become adjacent, the outcome of their interaction is determined by the LLM based on the relationship between their words, with the loser's word being replaced by the winner's. Word mutations, also based on LLM outputs, may occur. We conducted pilot experiments assuming that "strong animal species" would survive. From an initial population consisting of well-known species, many species emerged both gradually and in a punctuated equilibrium manner. Each trial demonstrated the evolution of diverse populations, with one type of large species becoming dominant, such as terrestrial animals, marine life, or extinct species, which were ecologically specialized and adapted ones across diverse extreme habitats. We also conducted a long-term experiment with a large population, demonstrating the emergence and coexistence of diverse species.