Reproduce “Point-in-Time Financial RAG with Frozen LLMs and Market-Feedback Adaptive Retrieval” 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
Zijie Zhao, Roy E. Welsch
The authors adapt retrieval for financial event-impact prediction by keeping a language-model reader fixed and using delayed standardized residual-return feedback to adjust source preferences for similar event types and horizons. The system retrieves past-only news and filing evidence, then evaluates held-out 2021–2023 events. Frozen source memory raises macro-F1 from 0.438 to 0.471, but does not establish causation or deployment profitability.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.