Saved in:
Bibliographic Details
Main Authors: Zhou, Ruizhe, Liu, Xiaoyang, Du, Gaoyuan, Zheng, Yi, Ren, Shouxi, Chakrabarti, Deepayan, Jiang, Dengdu
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.23955
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910266166345728
author Zhou, Ruizhe
Liu, Xiaoyang
Du, Gaoyuan
Zheng, Yi
Ren, Shouxi
Chakrabarti, Deepayan
Jiang, Dengdu
author_facet Zhou, Ruizhe
Liu, Xiaoyang
Du, Gaoyuan
Zheng, Yi
Ren, Shouxi
Chakrabarti, Deepayan
Jiang, Dengdu
contents Deploying machine learning in regulated financial environments -- credit risk, fraud detection, and anti-money laundering -- exposes critical vulnerabilities in algorithmic reproducibility. While early financial ML addressed statistical challenges such as backtest overfitting, deep neural networks and Generative AI have introduced mechanical nondeterminism rooted in hardware and architecture. This survey provides a systems perspective on reproducibility failures across three modalities now dominant in financial AI: tabular models (post-hoc explanation variance), graph networks (stochastic sampling and temporal asynchrony), and LLM-based agentic workflows (batch-dependent divergence and trajectory drift). We supplement the literature analysis with first-party experiments on public financial datasets -- quantifying explanation rank instability in credit scoring, prediction flip rates in GNN-based fraud detection, and tensor-parallel-induced output divergence in LLM entity extraction. We propose a layered evaluation framework linking modality-specific metrics (RBO, D_cos, TDI, PSD) to audit readiness, and empirically validate the complementarity of logit-level and semantic-level determinism measures.
format Preprint
id arxiv_https___arxiv_org_abs_2605_23955
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle From Accuracy to Auditability: A Survey of Determinism in Financial AI Systems
Zhou, Ruizhe
Liu, Xiaoyang
Du, Gaoyuan
Zheng, Yi
Ren, Shouxi
Chakrabarti, Deepayan
Jiang, Dengdu
Artificial Intelligence
Distributed, Parallel, and Cluster Computing
Machine Learning
Social and Information Networks
Computational Finance
Deploying machine learning in regulated financial environments -- credit risk, fraud detection, and anti-money laundering -- exposes critical vulnerabilities in algorithmic reproducibility. While early financial ML addressed statistical challenges such as backtest overfitting, deep neural networks and Generative AI have introduced mechanical nondeterminism rooted in hardware and architecture. This survey provides a systems perspective on reproducibility failures across three modalities now dominant in financial AI: tabular models (post-hoc explanation variance), graph networks (stochastic sampling and temporal asynchrony), and LLM-based agentic workflows (batch-dependent divergence and trajectory drift). We supplement the literature analysis with first-party experiments on public financial datasets -- quantifying explanation rank instability in credit scoring, prediction flip rates in GNN-based fraud detection, and tensor-parallel-induced output divergence in LLM entity extraction. We propose a layered evaluation framework linking modality-specific metrics (RBO, D_cos, TDI, PSD) to audit readiness, and empirically validate the complementarity of logit-level and semantic-level determinism measures.
title From Accuracy to Auditability: A Survey of Determinism in Financial AI Systems
topic Artificial Intelligence
Distributed, Parallel, and Cluster Computing
Machine Learning
Social and Information Networks
Computational Finance
url https://arxiv.org/abs/2605.23955