Reproduce “QRAFTI: An Agentic Framework for Empirical Research in Quantitative Finance” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Terence Lim, Kumar Muthuraman, Michael Sury
Researchers from Graphen Inc. and The University of Texas at Austin developed QRAFTI, an agentic framework that utilizes a multi-agent LLM architecture and specialized tools to automate empirical equity factor research. It replicates established financial factors from natural language descriptions and generates new ones, producing comprehensive, standardized diagnostic reports that enhance auditability and reproducibility in quantitative finance.
Key insights
- A multi-agent architecture coupled with domain-specific tools can effectively automate and standardize complex, multi-stage empirical financial research workflows.
- Reflection-based planning significantly enhances the performance and consistency of agentic systems on complex tasks by enabling self-auditing and plan refinement.
- Integrating a panel-data abstraction with Model Context Protocol (MCP) servers and prioritizing tool calls over dynamic code generation improves robustness and auditability.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.