Skip to content

Source paper

OASIS: Open Agent Social Interaction Simulations with One Million Agents

Ziyi Yang, Zaibin Zhang, Zirui Zheng and 20 more

Ziyi Yang, Zaibin Zhang, Zirui Zheng, Yuxian Jiang, Ziyue Gan, Zhiyu Wang, Zijian Ling, Jinsong Chen, Martz Ma, Bowen Dong, Prateek Gupta, Shuyue Hu, Zhenfei Yin, Guohao Li, Xu Jia, Lijun Wang, Bernard Ghanem, Huchuan Lu, Chaochao Lu, Wanli Ouyang, Yu Qiao, Philip Torr, Jing Shao

arXiv 2411.11581 · 2024

Research ideas from this paper (4)

Generated by AI. The paper’s authors did not write these.

Implementing Collaborative Filtering and Advanced Graph-Based Recommendation Systems in Multi-Agent Social Simulations

Integrate advanced collaborative filtering and graph-based recommendation systems into multi-agent social simulations to capture multi-hop network interactions.

OASIS: Open Agent Social Interaction Simulations with One Million Agents
🟠Medium💡Moderate InnovationLarge clusterRelated work found#Agents#Retrieval & memory

Multimodal Agent Perception: Integrating Vision and Audio into Social Media Simulations

Upgrade multi-agent social simulation frameworks to process multimodal post attachments by integrating vision-language models into agent perception modules.

OASIS: Open Agent Social Interaction Simulations with One Million Agents
🔴Hard🔬Significant ResearchMultiple GPUsRelated work found#Agents#Multimodal

Optimizing High-Concurrency Database and Inference Pipelines for Million-Agent Simulations

Optimize million-agent simulation concurrency by replacing relational databases and standard execution with in-memory graph stores and continuous batching inference engines.

OASIS: Open Agent Social Interaction Simulations with One Million Agents
🟠Medium💡Moderate InnovationLarge clusterNo close match found#Agents#Efficiency

Simulating Digital Economies: Expanding Action Spaces to Include Monetization and E-Commerce

Expand agent action spaces and simulation environments to include monetization and e-commerce behaviors driven by virtual currency dynamics.

OASIS: Open Agent Social Interaction Simulations with One Million Agents
🟡Fairly easy🔧Minor TweakMultiple GPUsNo close match found#Agents

AI summary

This paper introduces OASIS, a generalizable and scalable agent-based model (ABM) for simulating social media platforms like X and Reddit with up to one million large language model (LLM) agents. By accurately replicating real-world social phenomena such as information spreading, group polarization, and herd effects, OASIS provides a powerful, risk-free environment for studying complex societal systems and collective behaviors.

How it works (AI summary)

OASIS is built upon five modular components: an Environment Server tracking relational database states, a Recommendation System (RecSys) filtering information flow based on platform-specific algorithms, an Agent Module using LLMs with memory and Chain-of-Thought reasoning, a Time Engine for probabilistic temporal activation, and an asynchronous Scalable Inferencer to manage millions of concurrent agent requests efficiently.