Publications

GRAPHIA: Harnessing Social Graph Data to Enhance LLM-Based Social Simulation

Published in ACL, 2026

This work introduces Graphia, the first general LLM-based framework for simulating dynamic social graphs, enabling graph-supervised post-training that aligns agent behaviors with both local interactions and emergent network structure.

Leveraging LLM-based agents for social science research: insights from citation network simulations

Published in HSSCOMMS(SSCI Q1, AHCI Q1), 2025

This work introduces CiteAgent, an LLM-agent-based platform for simulating citation network dynamics, which enables controlled experimentation in academic environments.

Download Paper

LLM-Based Multi-Agent Systems are Scalable Graph Generative Models

Published in ACL-2025(findings), 2025

This work introduces GraphAgent-Generator (GAG), a scalable framework for generating social graphs using LLM-based multi-agent systems.

Download Paper | Download Bibtex

SRAP-Agent: Simulating and Optimizing Scarce Resource Allocation Policy with LLM-Based Agents

Published in EMNLP-2024(findings), 2024

This work introduces SRAP-Agent, a multi-agent frmaework for optimizing resource allocation policies.

Recommended citation: Ji, Jiarui, et al. "SRAP-Agent: Simulating and Optimizing Scarce Resource Allocation Policy with LLM-based Agent." Findings of the Association for Computational Linguistics: EMNLP 2024.
Download Paper | Download Bibtex