arXiv subjects

Yupeng Ren

Publications and source records attributed to Yupeng Ren.

11 recordsLinked to original sources

arXiv2026

Scratchy: Visual-Scratchpad Multimodal Reasoning for Cryptographic Proof Generation in EasyCrypt

Large language models (LLMs) have recently made substantial progress in formal proof generation, yet presenting distinctive challenges in cryptographic area. Computational security arguments posit that a valid proof must coordinate probability, adversarial games, invariants, assumptions and bounds, which can be provided by a machine-checked framework named EasyCrypt. Although all objects may appear in available context, LLMs still struggle because proof-theoretic dependencies are typically implicit in a linear representation and distributed across multiple programs. So, this paper presents Scratchy, a visual-scratchpad approach that exposes these dependencies for multimodal generation. Given the natural-language security description, with formal context and target propositions, the proof objects can be normalized into a typed proof-relation graph. Then a structure-preserving visual compiler transforms the graph into the formula-rich visual proof state that guides a multimodal model in generating the EasyCrypt proof. Also, the Scratchy-eval, a 114-task dataset derived from reliable official EasyCrypt files, has been introduced. It contains 64 security-form proof generations and 50 multiple-choice knowledge tests. After a series of evaluations, covering semantic grounding, relational invariants, and game reductions, classical LLMs like GPT-5.6-Sol and Claude-Opus-5 have gained a clear advantage from Scratchy's structured visual proof states. This contrast suggests that explicit proof structure can make the improvement and multimodal proof-state representation as a promising direction for computer-aided cryptography.

cs.CR↗

arXiv2025

SAMST: A Transformer framework based on SAM pseudo label filtering for remote sensing semi-supervised semantic segmentation

Public remote sensing datasets often face limitations in universality due to resolution variability and inconsistent land cover category definitions. To harness the vast pool of unlabeled remote sensing data, we propose SAMST, a semi-supervised semantic segmentation method. SAMST leverages the strengths of the Segment Anything Model (SAM) in zero-shot generalization and boundary detection. SAMST iteratively refines pseudo-labels through two main components: supervised model self-training using both labeled and pseudo-labeled data, and a SAM-based Pseudo-label Refiner. The Pseudo-label Refiner comprises three modules: a Threshold Filter Module for preprocessing, a Prompt Generation Module for extracting connected regions and generating prompts for SAM, and a Label Refinement Module for final label stitching. By integrating the generalization power of large models with the training efficiency of small models, SAMST improves pseudo-label accuracy, thereby enhancing overall model performance. Experiments on the Potsdam dataset validate the effectiveness and feasibility of SAMST, demonstrating its potential to address the challenges posed by limited labeled data in remote sensing semantic segmentation.

cs.CV↗

arXiv2024

FATS: A Prompt Injection Attack Utilizing Feign Security Agents with Deceptive Few-shots Learning

Large Language Models (LLMs) face significant security risks despite their advanced capabilities. While techniques like Reinforcement Learning with Human Feedback (RLHF) improve ethical alignment, excessive exposure to security-related training data may cause LLMs to overtrust such information, creating new vulnerabilities. Investigating this issue, we propose a novel attack method termed FATS (Feign Agent Attack with Toxic-shots). By obfuscating preference extraction, compromising toxicity samples, and inducing malicious behavior, we can effectively mislead LLMs into generating harmful outputs. To evaluate FATS effectiveness, we introduce the FAQuery dataset and conduct experiments on various LLMs. Well-known benchmarks like Advbench were selected to assess the approach. Results demonstrate that mainstream models, including GPT-4.1 (61.6\%) and Deepseek-R1 (99.3\%) are highly susceptible. It underscored the need to rigorously analyze security-related data sources during model training, developing more secure and reliable LLMs.

cs.CR↗

arXiv2023

Nonparametric Stochastic Analysis of Dynamic Frequency in Power Systems: A Generalized Ito Process Model

The large-scale integration of intermittent renewable energy has brought serious challenges to the frequency security of power systems. In this paper, a novel nonparametric stochastic analysis method of system dynamic frequency is proposed to accurately analyze the impact of renewable energy uncertainty on power system frequency security, independent of any parametric distribution assumption. The nonparametric uncertainty of renewable generation disturbance is quantified based on probabilistic forecasting. Then, a novel generalized Ito process is proposed as a linear combination of several Gaussian Ito processes, which can represent any probability distribution. Furthermore, a stochastic model of power system frequency response is constructed by considering virtual synchronization control of wind power. On basis of generalized Ito process, the complex nonlinear stochastic differential equation is transformed into a linear combination of several linear stochastic differential equations to approximate nonparametric probability distribution of the system dynamic frequency. Finally, the validity of the proposed method is verified by the single-machine system and IEEE 39-Bus system.

eess.SY↗

arXiv2022

MASNet:Improve Performance of Siamese Networks with Mutual-attention for Remote Sensing Change Detection Tasks

Siamese networks are widely used for remote sensing change detection tasks. A vanilla siamese network has two identical feature extraction branches which share weights, these two branches work independently and the feature maps are not fused until about to be sent to a decoder head. However we find that it is critical to exchange information between two feature extraction branches at early stage for change detection task. In this work we present Mutual-Attention Siamese Network (MASNet), a general siamese network with mutual-attention plug-in, so to exchange information between the two feature extraction branches. We show that our modification improve the performance of siamese networks on multi change detection datasets, and it works for both convolutional neural network and visual transformer.

cs.CV↗

arXiv2021

High order finite difference Hermite WENO fixed-point fast sweeping method for static Hamilton-Jacobi equations

In this paper, we combine the nonlinear HWENO reconstruction in \cite{newhwenozq} and the fixed-point iteration with Gauss-Seidel fast sweeping strategy, to solve the static Hamilton-Jacobi equations in a novel HWENO framework recently developed in \cite{mehweno1}. The proposed HWENO frameworks enjoys several advantages. First, compared with the traditional HWENO framework, the proposed methods do not need to introduce additional auxiliary equations to update the derivatives of the unknown function $\phi$. They are now computed from the current value of $\phi$ and the previous spatial derivatives of $\phi$. This approach saves the computational storage and CPU time, which greatly improves the computational efficiency of the traditional HWENO scheme. In addition, compared with the traditional WENO method, reconstruction stencil of the HWENO methods becomes more compact, their boundary treatment is simpler, and the numerical errors are smaller on the same mesh. Second, the fixed-point fast sweeping method is used to update the numerical approximation. It is an explicit method and does not involve the inverse operation of nonlinear Hamiltonian, therefore any Hamilton-Jacobi equations with complex Hamiltonian can be solved easily. It also resolves some known issues, including that the iterative number is very sensitive to the parameter $\varepsilon$ used in the nonlinear weights, as observed in previous studies. Finally, in order to further reduce the computational cost, a hybrid strategy is also presented. Extensive numerical experiments are performed on two-dimensional problems, which demonstrate the good performance of the proposed fixed-point fast sweeping HWENO methods.

math.NA↗

arXiv2021

High Order Residual Distribution Conservative Finite Difference HWENO Scheme for Steady State Problems

In this paper, we develop a high order residual distribution (RD) method for solving steady state conservation laws in a novel Hermite weighted essentially non-oscillatory (HWENO) framework recently developed in [24]. In particular, we design a high order HWENO integration for the integrals of source term and fluxes based on the point value of the solution and its spatial derivatives, and the principles of residual distribution schemes are adapted to obtain steady state solutions. Two advantages of the novel HWENO framework have been shown in [24]: first, compared with the traditional HWENO framework, the proposed method does not need to introduce additional auxiliary equations to update the derivatives of the unknown variable, and just compute them from the current point value of the solution and its old spatial derivatives, which saves the computational storage and CPU time, and thereby improve the computational efficiency of the traditional HWENO framework. Second, compared with the traditional WENO method, reconstruction stencil of the HWENO methods becomes more compact, their boundary treatment is simpler, and the numerical errors are smaller at the same grid. Thus, it is also a compact scheme when we design the higher order accuracy, compared with that in [11] Chou and Shu proposed. Extensive numerical experiments for one- and two-dimensional scalar and systems problems confirm the high order accuracy and good quality of our scheme.

math.NA↗

arXiv2021

High order asymptotic preserving Hermite WENO fast sweeping method for the steady-state $S_{N}$ transport equation

In this paper, we propose to combine the fifth order Hermite weighted essentially non-oscillatory (HWENO) scheme and fast sweeping method (FSM) for the solution of the steady-state $S_{N}$ transport equation in the finite volume framework. It is well-known that the $S_{N}$ transport equation asymptotically converges to a macroscopic diffusion equation in the limit of optically thick systems with small absorption and sources. Numerical methods which can preserve the asymptotic limit are referred to as asymptotic preserving methods. In the one-dimensional case, we provide the analysis to demonstrate the asymptotic preserving property of the high order finite volume HWENO method, by showing that its cell-edge and cell-average fluxes possess the thick diffusion limit. Numerical results in both one- and two- dimensions are presented to validate its asymptotic preserving property. A hybrid strategy to compute the nonlinear weights in the HWENO reconstruction is introduced to save computational cost. Extensive one- and two-dimensional numerical experiments are performed to verify the accuracy, asymptotic preserving property and positivity of the proposed HWENO FSM.

math.NA↗

arXiv2020

High order finite difference Hermite WENO fast sweeping methods for static Hamilton-Jacobi equations

In this paper, we propose a novel Hermite weighted essentially non-oscillatory (HWENO) fast sweeping method to solve the static Hamilton-Jacobi equations efficiently. During the HWENO reconstruction procedure, the proposed method is built upon a new finite difference fifth order HWENO scheme involving one big stencil and two small stencils. However, one major novelty and difference from the traditional HWENO framework lies in the fact that, we do not need to introduce and solve any additional equations to update the derivatives of the unknown function $\phi$. Instead, we use the current $\phi$ and the old spatial derivative of $\phi$ to update them. The traditional HWENO fast sweeping method is also introduced in this paper for comparison, where additional equations governing the spatial derivatives of $\phi$ are introduced. The novel HWENO fast sweeping methods are shown to yield great savings in both computational time and storage, which improves the computational efficiency of the traditional HWENO scheme. In addition, a hybrid strategy is also introduced to further reduce computational costs. Extensive numerical experiments are provided to validate the accuracy and efficiency of the proposed approaches.

math.NA↗

arXiv2017

A region-growing approach for automatic outcrop fracture extraction from a three-dimensional point cloud

Conventional manual surveys of rock mass fractures usually require large amounts of time and labor; yet, they provide a relatively small set of data that cannot be considered representative of the study region. Terrestrial laser scanners are increasingly used for fracture surveys because they can efficiently acquire large area, high-resolution, three-dimensional (3D) point clouds from outcrops. However, extracting fractures and other planar surfaces from 3D outcrop point clouds is still a challenging task. No method has been reported that can be used to automatically extract the full extent of every individual fracture from a 3D outcrop point cloud. In this study, we propose a method using a region-growing approach to address this problem; the method also estimates the orientation of each fracture. In this method, criteria based on the local surface normal and curvature of the point cloud are used to initiate and control the growth of the fracture region. In tests using outcrop point cloud data, the proposed method identified and extracted the full extent of individual fractures with high accuracy. Compared with manually acquired field survey data, our method obtained better-quality fracture data, thereby demonstrating the high potential utility of the proposed method.

cs.CV↗

arXiv2017

Outcrop fracture characterization on suppositional planes cutting through digital outcrop models (DOMs)

Conventional fracture data collection methods are usually implemented on planar surfaces or assuming they are planar; these methods may introduce sampling errors on uneven outcrop surfaces. Consequently, data collected on limited types of outcrop surfaces (mainly bedding surfaces) may not be a sufficient representation of fracture network characteristic in outcrops. Recent development of techniques that obtain DOMs from outcrops and extract the full extent of individual fractures offers the opportunity to address the problem of performing the conventional sampling methods on uneven outcrop surfaces. In this study, we propose a new method that performs outcrop fracture characterization on suppositional planes cutting through DOMs. The suppositional plane is the best fit plane of the outcrop surface, and the fracture trace map is extracted on the suppositional plane so that the fracture network can be further characterized. The amount of sampling errors introduced by the conventional methods and avoided by the new method on 16 uneven outcrop surfaces with different roughnesses are estimated. The results show that the conventional sampling methods don't apply to outcrops other than bedding surfaces or outcrops whose roughness > 0.04 m, and that the proposed method can greatly extend the types of outcrop surfaces for outcrop fracture characterization with the suppositional plane cutting through DOMs.

physics.geo-ph↗