ScholarPulse 日报 2026-08-23

2026-08-23 学术简报：2 篇。金融合规评估应聚焦于规则接地的行动和证据使用，而非单一合规分数。

今日速览

序号	标题	来源	日期	主题	推荐等级
1	ReguSim: Evaluating LLM Agent Rule Grounding in Financial Compliance	arXiv	2026-08-20	AI-Agent	高
2	Bringing analytic rigor to agentic AI for science: The Brain Researcher platform for neuroimaging data analysis	arXiv	2026-08-20	AI-Agent	高

重点论文与技术动态

1. ReguSim: Evaluating LLM Agent Rule Grounding in Financial Compliance

来源：arXiv
日期：2026-08-20
作者/机构：Yiyang Luo, Yihang Jiang, Qijun Xie, Liang Lan, Lin Willian Cong, Anyi Rao
主题标签：AI-Agent, arXiv
推荐等级：高
分类：cs.AI

一句话结论

金融合规评估应聚焦于规则接地的行动和证据使用，而非单一合规分数。

核心内容

引入ReguSim环境和ReguBench基准，分离金融合规中的四个关键要素：陈述推理、尝试行动、执行执行和监控证据。
交易者运行中，使用DeepSeek V4 Pro和Gemini 3.5 Flash模型，可见规则虽能减少但无法消除被拒绝行动，且激励或角色框架显著改变行为。
桥接研究证实，交易者理由可能误导独立监控者，除非提供执行证据，凸显证据展示的必要性。

方法与数据

基于ReguSim环境和ReguBench基准进行测试，使用DeepSeek V4 Pro和Gemini 3.5 Flash模型。
摘要未明确具体数据集。

价值判断

值得关注：金融合规评估需转向审计规则接地行动和证据使用，而非依赖单一分数。
可复用点：ReguSim和ReguBench框架可推广至其他金融合规场景以系统化评估。
局限/待核查：规则不能完全消除违规行为，需结合执行证据避免监控误导，摘要未深入探讨长期适用性。

摘要

LLM agents in financial markets may cite rules yet still submit orders that violate executable constraints or misread surveillance evidence. We introduce ReguSim, a controlled financial-compliance environment, and ReguBench, a target-marked monitoring benchmark, to separate four artifacts: stated reasoning, attempted action, execution enforcement, and monitor evidence. In trader runs with DeepSeek V4 Pro and Gemini 3.5 Flash, visible rules reduce but do not eliminate rejected actions, and incentive or persona framing shifts behavior. A bridge study shows that trader rationales can mislead an independent monitor unless enforcement evidence is shown. In monitoring, simple structured baselines either match or exceed prompt-only LLMs. The results frame financial compliance evaluation as an audit of rule-grounded actions and evidence use, rather than a single compliance score.

2. Bringing analytic rigor to agentic AI for science: The Brain Researcher platform for neuroimaging data analysis

来源：arXiv
日期：2026-08-20
作者/机构：Zijiao Chen, Nicholas Lu, Xinhui Li, Jocelyn A. Ricard, Ce Ju, Huan H. Wang
主题标签：AI-Agent, arXiv
推荐等级：高
分类：cs.AI, q-bio.NC

一句话结论

Brain Researcher 通过规则化分析流程显著提升神经影像数据分析的严谨性与准确性。

核心内容

Brain Researcher 在神经影像研究计算环境中运行，遵循可接受分析规则、必要检查和声明范围，有效防止选择性分析、过早成功声明及优化不完善标准。
基准测试中，工具选择准确率提升70.2个百分点（23.3%至93.6%），可验证依据从4.6%增至22.0%。
协作与自演化研究中，多宇宙分析揭示分析选择敏感性，科学评审将声明分类为接受、有条件接受、修订、阻止、拒绝或推迟，实现决策与证据的链接。

方法与数据

摘要未明确。

价值判断

值得关注：解决AI代理在科学分析中的常见缺陷（如选择性分析），通过嵌入方法论判断至工作流提升分析可辩护性。
可复用点：规则化框架可推广至其他科学数据分析任务，增强方法论严谨性与可重复性。
局限/待核查：摘要未明确。

摘要

AI agents can execute scientific analyses, but an analytic output becomes a defensible claim only after alternatives are weighed and the claim is limited to what the evidence supports. Agents may reproduce failures including selective analysis, premature declarations of success and optimization of imperfect criteria. We present Brain Researcher, an agentic research harness operating in a neuroimaging researcher's computational environment under rules for admissible analyses, required checks and claim scope. In benchmarks, Brain Researcher increased first-choice tool-selection accuracy across seven models by 70.2 percentage points (23.3% without it versus 93.6% with it) and verifiable grounding from 4.6% to 22.0%. In collaborator-led and self-evolving studies, multiverse analyses exposed analytic-choice sensitivity, and scientific review classified claims as accepted, qualified, revised, blocked, rejected or deferred. By linking decisions to evidence and provenance, Brain Researcher embeds methodological judgment within the workflow, not after it.