用于认知诊断的隐私保护异构多大语言模型联邦推理
Privacy-Preserving Heterogeneous Multi-LLM Federated Inference for Cognitive Diagnosis
- University of Cincinnati（辛辛那提大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
该研究针对AI教育系统中隐私与认知诊断精度难以平衡的问题，提出异构多LLM联邦推理框架，结合差分隐私与残差聚合，兼具强隐私保障与低精度损失，经多基准验证实用且可跨域推广。
AI中文摘要：
在AI驱动的教育系统中，平衡隐私保护与准确认知诊断仍面临重大挑战。为解决该问题，我们提出一种联邦推理框架，该框架允许多个商用大语言模型（LLM）API协作，无需访问学生原始数据或专有模型内部结构。我们的框架基于异构多LLM架构，采用LLaMA-3.3-70B、GPT-4o-mini、Claude-3-Haiku等多个联邦实体，结合ε-局部差分隐私，在聚合前向每个实体的预测输出中添加拉普拉斯噪声，同时采用基于残差的聚合方法缓解模型异质性。该方法基于“诚实但好奇”的信任范式，假设API提供者不会滥用提交的查询，且差分隐私机制可保护发布的诊断结果免受外部推理攻击。我们开展了严格的隐私-效用分析，证明该方法具有强隐私保障且精度损失极小；在三个教育基准上进行的广泛实际评估，证实了该框架的实用性和跨域泛化能力。
英文摘要：
Significant challenges remain in AI-driven educational systems in balancing privacy preservation with accurate cognitive diagnosis. To overcome this, we propose a federated inference framework in which several commercial LLM APIs collaborate without requiring access to raw student data or proprietary model internals. Using multiple federated entities, such as LLaMA-3.3-70B, GPT-4o-mini, and Claude-3-Haiku, our framework builds upon a heterogeneous multi-LLM architecture. The predictions generated by these entities are combined with epsilon-local differential privacy by adding Laplace noise locally to each entity's prediction output before aggregation, while residual-based aggregation mitigates model heterogeneity. Our approach is predicated on an honest-but-curious trust paradigm in which API providers are presumed not to abuse submitted queries, and our differential privacy mechanism shields the published diagnostic results from external inference. We conduct rigorous privacy-utility analysis showing strong privacy guarantees with minimal accuracy loss, and extensive real-world evaluations across three educational benchmarks confirm the framework's practical usability and cross-domain generalizability.
