SwarmWorld: Stigmergic technological evolution in societies of language-model agents

ID: b98e8ee7
原文链接: https://arxiv.org/abs/2608.26081
PDF: https://arxiv.org/pdf/2608.26081v1
作者: Subhadeep Pal, Fiona Y. Wang, Markus J. Buehler
日期: 2026-08-26
更新: 2026-08-26
分类: agents
来源类型: paper
标签: multi-agent、emergence、stigmergy、collective-intelligence、arxiv
质量评分: 4/5
抓取时间: 2026-08-28T05:15:03Z

中文导读

初始同质的 LLM agent 在 SwarmWorld 中不分配角色不给配方，自组织成演化的技术社会：探索空间环境加工资源测试材料建造持久工件并编写可执行控制器，控制器在 agent 全部撤场后由确定性模拟器在未见扰动下评审共享社会发展出比强 best-of-N 独立搜索更宽更抗扰的技术组合，但最强单件工件上独立搜索仍有竞争力分工自发涌现为探索建造维护协调四种行为；技术通过协作建造可执行继承与持久的 agent-工件网络积累，且复用多从物理观察而非通信开始；显式文化机制放大协作，但收益取决于结果与时间尺度

为什么值得关注

无角色分配的 LLM 群体自组织出演化技术社会，是 emergence/stigmergy 方向少见的可复现实验设置；"共享搜索的组合韧性超过 best-of-N，但最强单件工件上独立搜索仍有竞争力"这条边界，对多 agent 系统何时值得上群体给出了可操作的判据。

关键信息

论文标题：SwarmWorld: Stigmergic technological evolution in societies of language-model agents
作者：Subhadeep Pal, Fiona Y. Wang, Markus J. Buehler
arXiv：https://arxiv.org/abs/2608.26081
发布时间：2026-08-26
arXiv 分类：cs.AI, cond-mat.mtrl-sci, cs.CL
关联标签：multi-agent、emergence、stigmergy、collective-intelligence、arxiv

English Abstract

Collective intelligence can emerge when individuals coordinate through a shared environment, allowing local actions to accumulate into durable social organization. Language-model agents offer a new substrate for this process, yet most multi-agent systems rely on direct conversation, predefined roles, or centralized workflows. It remains unclear whether decentralized agents can build functional technologies and outperform independent search. Here, initially homogeneous LLM agents in SwarmWorld self-organize without assigned roles or recipes into evolving technological societies. Agents explore a spatial environment, process resources, test materials, construct persistent artifacts, and write executable controllers evaluated by a deterministic simulator under unseen disturbances after the agents are removed. SwarmWorld splits cognition from consequence: agents propose architectures and controllers within fixed action and material schemas, while the simulated world determines function. Shared societies develop broader, more resilient technological portfolios than a strong best-of-N isolated-search baseline, although isolated search remains competitive for the strongest artifact. Agents differentiate into exploration, construction, maintenance, and coordination behaviors, transitioning as the world matures. Technologies accumulate through collaborative construction, executable inheritance, and persistent agent-artifact networks, with most reuse beginning through physical observation rather than communication. Explicit cultural mechanisms amplify collaboration and organization, but functional benefits depend on outcome and timescale. Physical stigmergy alone supports capable societies, while interaction drives persistent technological ecologies rather than universally superior individual inventions.

English Summary

Initially homogeneous LLM agents self-organize, without assigned roles or recipes, into evolving technological societies in SwarmWorld: they explore a spatial environment, process resources, test materials, build persistent artifacts, and write executable controllers judged by a deterministic simulator under unseen disturbances after the agents are removed. Shared societies develop broader, more resilient technological portfolios than a strong best-of-N isolated-search baseline, while isolated search stays competitive for the strongest single artifact. Agents differentiate into exploration, construction, maintenance, and coordination behaviors; technologies accumulate via collaborative construction, executable inheritance, and persistent agent-artifact networks, with most reuse starting through physical observation rather than communication.

Obsidian Notes

内容由 opencli arxiv paper 拉取 arXiv 元数据与摘要生成。
中文导读与价值判断均锚定在条目已有摘要、论文摘要、作者、日期与分类信息上；未补充论文摘要之外的实验细节。