Certified Uncertainty Propagation in One-Shot Federated Bayesian Models via Posterior Event Transport
本文提出一种通过后验事件传播方法，在一次性联邦贝叶斯模型中确保模型安全性的框架，解决了局部证书无法直接保证聚合模型安全性的问题。
0 citationsRead paper
Selected work
Representative Papers
PG-KINN: A Physics-Informed Petrov-Galerkin Kolmogorov-Arnold Network for Solving Forward and Inverse PDEs
This work addresses the limitations of conventional physics-informed neural networks in solving non-self-adjoint, nonlinear, and inverse problems—namely spectral bias, parameter redundancy, and poor generalization. It introduces, for the first time, Kolmogorov–Arnold Networks (KANs) within a Petrov–Galerkin weak-formulation framework, employing KANs as trial functions and locally compact piecewise polynomials as test functions, with Gauss–Legendre quadrature enabling low-order derivative evaluation and well-conditioned weak residual computation. This approach overcomes the constraints of strong-form and energy-based methods in handling complex operators and parameter inversion, significantly enhancing accuracy, stability, and interpretability. Benchmark tests involving crack singularities, stress concentrations, hyperelasticity, heterogeneous medium inversion, and complex geometries consistently demonstrate superior performance over both multilayer perceptrons and existing KAN-based approaches such as PIKAN.
0 citationsRead paper
Differential Addition on Twisted Edwards Curves
This work addresses the inefficiency of differential addition in Montgomery scalar multiplication on twisted Edwards curves by proposing novel differential addition and doubling formulas. When the difference point is given in affine coordinates, the new formulas require only 5M + 4S + 1D for differential addition and 3M + 7S + 1D for doubling, where M denotes field multiplication, S squaring, and D multiplication by a constant. These operation counts represent a significant reduction in computational cost compared to existing methods. By optimizing the core operations of scalar multiplication, the proposed approach achieves the most efficient differential arithmetic known to date on twisted Edwards curves, thereby substantially enhancing overall performance.
0 citationsRead paper
Differential addition on Jacobi quartic Curves
This work addresses the inefficiency of differential addition and doubling operations on Jacobi quartic curves by proposing three novel, highly efficient formulae. Assuming input points are given in affine coordinates with their difference known, the proposed methods optimize field operations—specifically multiplications (M), squarings (S), and multiplications by constants (D)—achieving computational costs of 5M+4S+1D, 3M+7S+1D, and 3M+6S+3D, respectively. These formulae significantly reduce arithmetic complexity while preserving correctness, thereby enhancing the performance of relevant operations in elliptic curve cryptography. The contributions offer both practical utility and theoretical innovation, advancing the state of the art in efficient elliptic curve arithmetic.
0 citationsRead paper
ITS-Mina: A Harris Hawks Optimization-Based All-MLP Framework with Iterative Refinement and External Attention for Multivariate Time Series Forecasting
This work addresses the high computational cost of Transformer-based models in multivariate time series forecasting by proposing a pure MLP architecture that progressively refines predictions through iterative residual mixers. To efficiently capture cross-variable dependencies, the model incorporates an external attention mechanism with linear complexity, built upon learnable memory units. Additionally, the Harris Hawks Optimization (HHO) algorithm is employed to automatically tune critical hyperparameters, such as the dropout rate. Evaluated on six benchmark datasets, the proposed method significantly outperforms eleven state-of-the-art baseline models, achieving both superior prediction accuracy and computational efficiency.
0 citationsRead paper
Recent publications
Latest Papers
Certified Uncertainty Propagation in One-Shot Federated Bayesian Models via Posterior Event Transport
本文提出一种通过后验事件传播方法，在一次性联邦贝叶斯模型中确保模型安全性的框架，解决了局部证书无法直接保证聚合模型安全性的问题。
0 citationsRead paper
PG-KINN: A Physics-Informed Petrov-Galerkin Kolmogorov-Arnold Network for Solving Forward and Inverse PDEs
This work addresses the limitations of conventional physics-informed neural networks in solving non-self-adjoint, nonlinear, and inverse problems—namely spectral bias, parameter redundancy, and poor generalization. It introduces, for the first time, Kolmogorov–Arnold Networks (KANs) within a Petrov–Galerkin weak-formulation framework, employing KANs as trial functions and locally compact piecewise polynomials as test functions, with Gauss–Legendre quadrature enabling low-order derivative evaluation and well-conditioned weak residual computation. This approach overcomes the constraints of strong-form and energy-based methods in handling complex operators and parameter inversion, significantly enhancing accuracy, stability, and interpretability. Benchmark tests involving crack singularities, stress concentrations, hyperelasticity, heterogeneous medium inversion, and complex geometries consistently demonstrate superior performance over both multilayer perceptrons and existing KAN-based approaches such as PIKAN.
0 citationsRead paper
Differential Addition on Twisted Edwards Curves
This work addresses the inefficiency of differential addition in Montgomery scalar multiplication on twisted Edwards curves by proposing novel differential addition and doubling formulas. When the difference point is given in affine coordinates, the new formulas require only 5M + 4S + 1D for differential addition and 3M + 7S + 1D for doubling, where M denotes field multiplication, S squaring, and D multiplication by a constant. These operation counts represent a significant reduction in computational cost compared to existing methods. By optimizing the core operations of scalar multiplication, the proposed approach achieves the most efficient differential arithmetic known to date on twisted Edwards curves, thereby substantially enhancing overall performance.
0 citationsRead paper
Differential addition on Jacobi quartic Curves
This work addresses the inefficiency of differential addition and doubling operations on Jacobi quartic curves by proposing three novel, highly efficient formulae. Assuming input points are given in affine coordinates with their difference known, the proposed methods optimize field operations—specifically multiplications (M), squarings (S), and multiplications by constants (D)—achieving computational costs of 5M+4S+1D, 3M+7S+1D, and 3M+6S+3D, respectively. These formulae significantly reduce arithmetic complexity while preserving correctness, thereby enhancing the performance of relevant operations in elliptic curve cryptography. The contributions offer both practical utility and theoretical innovation, advancing the state of the art in efficient elliptic curve arithmetic.
0 citationsRead paper
ITS-Mina: A Harris Hawks Optimization-Based All-MLP Framework with Iterative Refinement and External Attention for Multivariate Time Series Forecasting
This work addresses the high computational cost of Transformer-based models in multivariate time series forecasting by proposing a pure MLP architecture that progressively refines predictions through iterative residual mixers. To efficiently capture cross-variable dependencies, the model incorporates an external attention mechanism with linear complexity, built upon learnable memory units. Additionally, the Harris Hawks Optimization (HHO) algorithm is employed to automatically tune critical hyperparameters, such as the dropout rate. Evaluated on six benchmark datasets, the proposed method significantly outperforms eleven state-of-the-art baseline models, achieving both superior prediction accuracy and computational efficiency.
0 citationsRead paper