arXiv q-fin· Zijie Zhao, Mingjun Sun, Shenbo Xu·· 5 hr agoAI Score51
Survey on Uncertainty Quantification for Financial Foundation Models: Proposes a Source-Aware Framework with Six Categories of Uncertainty Sources
Uncertainty Quantification for Financial Foundation Models: A Survey
AI summary
This survey paper accepted by ICAIF 2026 proposes a source-aware framework that divides uncertainty in financial foundation model workflows into three categories and six sources: evidence and context, model and reasoning, and outcomes and actions. It treats agent workflows as an execution layer where uncertainty propagates, transforms, or is reduced.
Source:arXiv q-fin · arxiv.org