Automated Social Science: Language Models as Scientist and Subjects

By Benjamin S. Manning, Kehang Zhu, John J. Horton

Rating

1545
Battle Count: 67

Relevance

6/10
While not directly applicable to quantitative trading, the methodology could be adapted for simulating market behaviors, testing trading strategies, or generating hypotheses about market dynamics. The auction scenario demonstrates potential relevance to market mechanisms.

Implementation Complexity

8/10
The system involves complex integration of LLMs, structural causal models, and automated experimental design. Implementing such a system would require significant expertise in AI, causal inference, and experimental design.

Reproducibility

4/5
The paper provides detailed information about the system implementation and experimental design, enhancing reproducibility. Code and data are stated to be available on the author's website.

About this paper

Methodology: Structural Causal Model-based Approach. Problem types: Causal Inference, Hypothesis Generation, Hypothesis Testing, Simulation.

The interactive Everscope explorer (charts, battles, favorites) loads below.