Learning Whom to Trust: Market-Feedback Adaptive Retrieval for Frozen LLMs in Event-Driven Financial RAG
arXiv:2605.31201v1 Announce Type: new Abstract: Financial retrieval-augmented generation (RAG) systems typically rank evidence by textual relevance, but in financial markets the useful evidence source depends on event type, forecast horizon, and market context. We study news-triggered event-impact prediction as a point-in-time financial RAG problem. For each company-news anchor, the system retrieves related financial news and SEC filing passages, appends a pre-decision market-context card, and predicts multi-horizon residual-return signals. Our method keeps the large language model (LLM) reade
The paper provides a practical application for advanced RAG systems in event-driven financial markets, leveraging frozen LLMs to focus on efficient, context-aware information retrieval.
This work demonstrates a significant step towards more accurate and dynamic financial analysis by refining how LLMs access and interpret real-time market data.
Traditional financial RAG systems that rely solely on textual relevance are being superseded by methods incorporating market feedback and event-specific context for improved predictive power.
- · Financial data analytics firms
- · Quantitative hedge funds
- · Event-driven traders
- · AI-powered investment platforms
- · Legacy financial analysis software
- · Purely text-based RAG approaches
- · Human financial researchers relying on static data without market feedback
Improved accuracy and speed of financial event-impact prediction using AI.
Increased capital allocation to AI-driven trading strategies due to enhanced predictive capabilities.
Potential for new financial instruments and market structures that leverage minute-by-minute AI-generated insights.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL