Rethinking LLM Human Simulation: When a Graph is What You Need
This provides a lightweight, efficient alternative to LLMs for human simulation tasks, which is incremental but offers practical benefits in domains like survey prediction and decision-making.
The paper tackles the problem of simulating human choices among discrete options, showing that a graph neural network (GNN) can match or surpass large language models (LLMs) in accuracy while being three orders of magnitude smaller.
Large language models (LLMs) are increasingly used to simulate humans, with applications ranging from survey prediction to decision-making. However, are LLMs strictly necessary, or can smaller, domain-grounded models suffice? We identify a large class of simulation problems in which individuals make choices among discrete options, where a graph neural network (GNN) can match or surpass strong LLM baselines despite being three orders of magnitude smaller. We introduce Graph-basEd Models for human Simulation (GEMS), which casts discrete choice simulation tasks as a link prediction problem on graphs, leveraging relational knowledge while incorporating language representations only when needed. Evaluations across three key settings on three simulation datasets show that GEMS achieves comparable or better accuracy than LLMs, with far greater efficiency, interpretability, and transparency, highlighting the promise of graph-based modeling as a lightweight alternative to LLMs for human simulation. Our code is available at https://github.com/schang-lab/gems.