The distinguishability of Longa's atomic patterns for elliptic curve point doubling and addition operations could not be conclusively determined due to technical limitations, despite implementing the patterns on a microcontroller and analyzing the measured electromagnetic trace.
本文通過計算特定方程在有限域上的解的數量,明確地確定了在偶特徵有限域上冪函數x2m+3和x2m+5(m > 2為正整數)的二次零微分頻譜,以及在奇特徵有限域上函數xpk+1的二次零微分一致性。其中,函數F(x) = x4在奇n的F3n上是一個完美非線性(PN)和二次零微分0一致的函數。
유한체 F2n과 Fpn 상의 특정 멱함수 F(x) = x2m+3, x2m+5, xpk+1의 2차 영 미분 스펙트럼을 계산하여 이들의 Feistel 부메랑 균일성을 결정하였다. 특히 F(x) = x4는 홀수 n에 대해 완전 비선형이면서 2차 영 미분 균일성이 0인 함수임을 보였다.
本論文では、偶数特性を持つ有限体上の冪関数x2m+3およびx2m+5の2次零微分スペクトラムを明示的に決定し、奇数特性を持つ有限体上の関数xpk+1の2次零微分一様性を計算した。特に、F(x) = x4は奇数nを持つF3nの上で完全非線形(PN)かつ2次零微分的に0一様な関数であることを示した。
The paper presents explicit computations of the second-order zero differential spectra for several classes of power functions over finite fields, including x^(2m+3), x^(2m+5), and x^(pk+1), which are crucial for analyzing the resistance of Feistel network-based ciphers against differential and boomerang attacks.
This paper proposes the first functionally equivalent extraction attack against ReLU neural networks under the hard-label setting, where the adversary can only access the most likely class label instead of the raw output.
A novel adversarial attack method using Golden Ratio Search (GRS) that generates minimal power adversarial perturbations to effectively fool deep learning models in Automatic Modulation Classification (AMC).
이 논문은 기존 및 포스트양자 암호화 설정에서 NIST 표준 및 맞춤형 설계를 통한 임계값 디지털 서명에 대한 포괄적이고 체계적인 조사를 제공합니다. 또한 안전한 다자간 계산을 통한 일반적인 임계값 기법과 특수한 서명 기능(그룹, 링, 다중 서명)을 다룹니다.
This survey provides a comprehensive and systematic examination of threshold digital signatures, encompassing conventional and post-quantum cryptography, as well as exotic signature schemes, with a focus on their real-world applications.
提出一種新的輕量級抗量子數字簽章方案 TVPD-HORS,能夠顯著提高實時應用的驗證速度,同時保持與傳統方案相當的簽名生成速度。