用几何方法揭示如何精准窃取模型最后一层，明确边界。
The Geometry of Last-Layer Model Stealing
- 基于几何分析，给出完美复制Transformer最后一层的条件。
- 发现隐藏层无法仅凭输出结果完全逆向重构。
- 为模型安全提供清晰的可盗与不可盗边界，适合安全研究者参考。
本文通过几何视角解释一种已知的机器学习模型窃取方法。作者精确推导出完全复制Transformer网络最后一层所需的条件。深入分析隐藏层后，明确了其重构的理论极限。研究表明，仅通过观察最终输出结果，无法完全逆向工程隐藏网络结构。该研究系统地划清了模型可被窃取与不可被窃取的边界。
原文摘要 · Abstract (English)
This paper uses geometry to explain how a machine learning model can be stolen using an already existing well-known method. The author has shown the exact conditions required to perfectly copy the final layer of a transformer network. When looking deeper into the hidden layers the author has explained clear limits. The author has also demonstrated that a hidden network cannot be fully reverse engineered just by looking at the final results. The research clearly maps out what can and cannot be stolen from a model.
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。