用基准测试评估大模型在差分隐私推理中的能力
DPrivBench: Benchmarking LLMs' Reasoning for Differential Privacy
- 构建专门的差分隐私推理基准，测试大模型判断算法是否满足隐私保证
- 强模型能处理基础机制，但对高级算法普遍表现不佳
- 适合研究自动化隐私验证和大模型推理能力的学者
差分隐私（DP）在数据隐私保护中有广泛应用，但设计和验证DP算法需要专家级推理能力，对非专家构成高门槛。以往工作要么依赖需大量领域知识的专业验证语言，要么仍为半自动化且需人工干预。本文探究大语言模型（LLMs）能否自动化完成DP推理。我们提出DPrivBench，一个基准测试，每个实例询问在特定假设下某个函数或算法是否满足给定的DP保证。该基准覆盖广泛的DP主题，涵盖不同难度层级，并通过避免简单模式匹配来防止捷径推理。实验表明，尽管最强模型能处理教科书级机制，所有模型在高级算法上均表现不佳，暴露出当前DP推理能力的显著差距。通过进一步分析与失败模式研究，我们识别出若干改进自动化DP推理的潜在方向。本基准为开发和评估此类方法提供了坚实基础，补充了现有的数学推理基准。
原文摘要 · Abstract (English)
Differential privacy (DP) has a wide range of applications for protecting data privacy, but designing and verifying DP algorithms requires expert-level reasoning, creating a high barrier for non-expert practitioners. Prior works either rely on specialized verification languages that demand substantial domain expertise or remain semi-automated and require human-in-the-loop guidance. In this work, we investigate whether large language models (LLMs) can automate DP reasoning. We introduce DPrivBench, a benchmark in which each instance asks whether a function or algorithm satisfies a stated DP guarantee under specified assumptions. The benchmark is carefully designed to cover a broad range of DP topics, span diverse difficulty levels, and resist shortcut reasoning through trivial pattern matching. Experiments show that while the strongest models handle textbook mechanisms well, all models struggle with advanced algorithms, revealing substantial gaps in current DP reasoning capabilities. Through further analytic study and failure-mode analysis, we identify several promising directions for improving automated DP reasoning. Our benchmark provides a solid foundation for developing and evaluating such methods, and complements existing benchmarks for mathematical reasoning.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。