Learning to cooperate in a changing world: How caring about the future promotes cooperation across scales
Abstract
In social dilemmas, individuals need to forgo short-term temptations to achieve synergistic collective outcomes through cooperation. Previous work has examined mechanisms through which cooperation can evolve, including direct reciprocity, indirect reciprocity, environmental stochasticity, network reciprocity, and demographic stochasticity. These mechanisms have largely been studied under natural selection or social learning, where strategies with higher payoffs are more likely to spread. It is equally important to study cooperation under self-learning, where individuals adapt through their own experiences. Here, we focus on multi-agent reinforcement learning and derive analytical conditions under which these mechanisms stabilize cooperation under learning dynamics across scales. We find that reinforcement learning can steer self-interested individuals toward cooperation when they value the future over short-term temptation. Our work provides a unified approach to identifying the determinants of learning to cooperate in a changing world, thereby paving the way for the advancement of cooperative artificial intelligence.
Related
- Explaining human cooperation through a dual mechanism of individual and social learning
- Social comparison shapes the evolution of cooperation in structured populations
- The Butterfly Defects: Tracing Cooperation in LLM Agent Networks
- Nonlinear Opinion Dynamics of Stubborn Individuals in Time-Varying Cooperation–Competition Networks
- Non-Hermitian Cooperation Induced Chern Insulator
BibTeX
@misc{geng2026learning,
title = {{Learning to cooperate in a changing world: How caring about the future promotes cooperation across scales}},
author = {Yuxin Geng and Xingru Chen and Xin Wang and Hongwei Zheng and Longzhao Liu and Shaoting Tang and Feng Fu},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2609.34005},
eprint = {2609.34005},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.34005},
}
RIS
TY - UNPB
TI - Learning to cooperate in a changing world: How caring about the future promotes cooperation across scales
AU - Yuxin Geng
AU - Xingru Chen
AU - Xin Wang
AU - Hongwei Zheng
AU - Longzhao Liu
AU - Shaoting Tang
AU - Feng Fu
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2609.34005
UR - https://arxiv.org/abs/2609.34005
AB - In social dilemmas, individuals need to forgo short-term temptations to achieve synergistic collective outcomes through cooperation. Previous work has examined mechanisms through which cooperation can evolve, including direct reciprocity, indirect reciprocity, environmental stochasticity, network reciprocity, and demographic stochasticity. These mechanisms have largely been studied under natural selection or social learning, where strategies with higher payoffs are more likely to spread. It is equally important to study cooperation under self-learning, where individuals adapt through their own experiences. Here, we focus on multi-agent reinforcement learning and derive analytical conditions under which these mechanisms stabilize cooperation under learning dynamics across scales. We find that reinforcement learning can steer self-interested individuals toward cooperation when they value the future over short-term temptation. Our work provides a unified approach to identifying the determinants of learning to cooperate in a changing world, thereby paving the way for the advancement of cooperative artificial intelligence.
ER -