FinCARDS: Card-Based Analyst Reranking for Financial Document Question Answering
For financial analysts and QA systems, FinCards provides a more stable and auditable reranking method for long corporate filings, addressing the limitations of semantic-only rerankers.
FinCards improves early-rank retrieval and reduces ranking variance in financial document QA by reframing evidence selection as constraint satisfaction with a finance-aware schema, outperforming lexical and LLM-based rerankers without fine-tuning.
Financial question answering (QA) over long corporate filings requires evidence to satisfy strict constraints on entities, financial metrics, fiscal periods, and numeric values. However, existing LLM-based rerankers primarily optimize semantic relevance, leading to unstable rankings and opaque decisions on long documents. We propose FinCards, a structured reranking framework that reframes financial evidence selection as constraint satisfaction under a finance-aware schema. FinCards represents filing chunks and questions using aligned schema fields (entities, metrics, periods, and numeric spans), enabling deterministic field-level matching. Evidence is selected via a multi-stage tournament reranking with stability-aware aggregation, producing auditable decision traces. Across two corporate filing QA benchmarks, FinCards substantially improves early-rank retrieval over both lexical and LLM-based reranking baselines, while reducing ranking variance, without requiring model fine-tuning or unpredictable inference budgets. Our code is available at https://github.com/XanderZhou2022/FINCARDS.