发表机构
University of Bergen; Université Paris 8; University of Zagreb; Radboud University; KU Leuven(卑尔根大学; 巴黎第八大学; 萨格勒布大学; 拉德堡德大学; 荷语鲁汶大学)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
该研究从几何代数角度研究ReLU网络的密码分析提取,通过分析相邻线性区域的向量值行为得到雅可比差的秩一表征,补充列侧信息,提升了不同精度下的提取估计效果,并将分析拓展至低精度场景。
AI 中文摘要
我们从几何和代数视角重新研究ReLU网络的密码分析提取问题。我们并未将研究局限于单个输出分量,而是研究了相邻线性区域上的完整向量值行为。这推导出了雅可比差的秩一表征,该表征既能恢复常规的行特征信息,还能揭示互补的列侧信息。我们的实验展示了如何将这种额外结构应用于提取任务,并在不同数值精度 regime(float64、float32和float16)下提升了数值估计效果。我们将分析范围从文献中常见的高精度和输出舍入设置,拓展至许多实际场景中遇到的低精度设置。
英文摘要
We revisit cryptanalytic extraction of ReLU networks from a geometric and algebraic perspective. Rather than restricting attention to a single output component, we study the full vector-valued behavior across adjacent linear regions. This leads to a rank-one characterization of Jacobian differences that recovers the usual row-signature information while also revealing complementary column-side information. Our experiments show how this additional structure can be used in ex- traction and improves numerical estimation under different numerical- precision regimes (float64, float32 and float16). We extend the analysis beyond the high-precision and output-rounding settings commonly con- sidered in the literature towards the low-precision settings encountered in many practical settings.