Guardado en:
Detalles Bibliográficos
Autores principales: Xue, Zhiyi, Chen, Xiaohong, Zhang, Min
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:https://arxiv.org/abs/2601.09762
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918289873043456
author Xue, Zhiyi
Chen, Xiaohong
Zhang, Min
author_facet Xue, Zhiyi
Chen, Xiaohong
Zhang, Min
contents Compliance testing in highly regulated domains is crucial but largely manual, requiring domain experts to translate complex regulations into executable test cases. While large language models (LLMs) show promise for automation, their susceptibility to hallucinations limits reliable application. Existing hybrid approaches mitigate this issue by constraining LLMs with formal models, but still rely on costly manual modeling. To solve this problem, this paper proposes RAFT, a framework for requirements auto-formalization and compliance test generation via explicating tacit regulatory knowledge from multiple LLMs. RAFT employs an Adaptive Purification-Aggregation strategy to explicate tacit regulatory knowledge from multiple LLMs and integrate it into three artifacts: a domain meta-model, a formal requirements representation, and testability constraints. These artifacts are then dynamically injected into prompts to guide high-precision requirement formalization and automated test generation. Experiments across financial, automotive, and power domains show that RAFT achieves expert-level performance, substantially outperforms state-of-the-art (SOTA) methods while reducing overall generation and review time.
format Preprint
id arxiv_https___arxiv_org_abs_2601_09762
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Explicating Tacit Regulatory Knowledge from LLMs to Auto-Formalize Requirements for Compliance Test Case Generation
Xue, Zhiyi
Chen, Xiaohong
Zhang, Min
Software Engineering
Artificial Intelligence
Compliance testing in highly regulated domains is crucial but largely manual, requiring domain experts to translate complex regulations into executable test cases. While large language models (LLMs) show promise for automation, their susceptibility to hallucinations limits reliable application. Existing hybrid approaches mitigate this issue by constraining LLMs with formal models, but still rely on costly manual modeling. To solve this problem, this paper proposes RAFT, a framework for requirements auto-formalization and compliance test generation via explicating tacit regulatory knowledge from multiple LLMs. RAFT employs an Adaptive Purification-Aggregation strategy to explicate tacit regulatory knowledge from multiple LLMs and integrate it into three artifacts: a domain meta-model, a formal requirements representation, and testability constraints. These artifacts are then dynamically injected into prompts to guide high-precision requirement formalization and automated test generation. Experiments across financial, automotive, and power domains show that RAFT achieves expert-level performance, substantially outperforms state-of-the-art (SOTA) methods while reducing overall generation and review time.
title Explicating Tacit Regulatory Knowledge from LLMs to Auto-Formalize Requirements for Compliance Test Case Generation
topic Software Engineering
Artificial Intelligence
url https://arxiv.org/abs/2601.09762