arXiv subjects

Chao Su

Publications and source records attributed to Chao Su.

8 recordsLinked to original sources

arXiv2026

Steady-State and Transient Heat Conduction Analysis Using a Polygonal Cell-Based Smoothed Finite Element Method

This paper presents a polygonal cell-based smoothed finite element method (CS-FEM) for two-dimensional steady-state and transient heat-conduction analysis. In the proposed formulation, Wachspress shape functions are employed to construct the temperature approximation over general polygonal elements, and the smoothed temperature gradient is evaluated through boundary integration over cell-based smoothing domains. The resulting formulation is implemented in ABAQUS through the user-defined element (UEL) interface, enabling heat-conduction analysis using polygonal meshes within a commercial finite element environment. Several numerical examples, including a linear patch test, steady-state benchmark problems, and transient heat-conduction problems with different boundary conditions, are investigated to verify the accuracy, convergence behavior, and robustness of the proposed method. The numerical results show good agreement with analytical or reference solutions. Compared with conventional FEM using triangular and quadrilateral elements, the proposed polygonal CS-FEM exhibits favorable accuracy and convergence performance, while providing greater flexibility in mesh generation for complex geometries. The proposed framework therefore offers an accurate and robust numerical approach for steady-state and transient heat-conduction analysis.

math.NT↗

arXiv2025

Semantic-Consistent Bidirectional Contrastive Hashing for Noisy Multi-Label Cross-Modal Retrieval

Cross-modal hashing (CMH) facilitates efficient retrieval across different modalities (e.g., image and text) by encoding data into compact binary representations. While recent methods have achieved remarkable performance, they often rely heavily on fully annotated datasets, which are costly and labor-intensive to obtain. In real-world scenarios, particularly in multi-label datasets, label noise is prevalent and severely degrades retrieval performance. Moreover, existing CMH approaches typically overlook the partial semantic overlaps inherent in multi-label data, limiting their robustness and generalization. To tackle these challenges, we propose a novel framework named Semantic-Consistent Bidirectional Contrastive Hashing (SCBCH). The framework comprises two complementary modules: (1) Cross-modal Semantic-Consistent Classification (CSCC), which leverages cross-modal semantic consistency to estimate sample reliability and reduce the impact of noisy labels; (2) Bidirectional Soft Contrastive Hashing (BSCH), which dynamically generates soft contrastive sample pairs based on multi-label semantic overlap, enabling adaptive contrastive learning between semantically similar and dissimilar samples across modalities. Extensive experiments on four widely-used cross-modal retrieval benchmarks validate the effectiveness and robustness of our method, consistently outperforming state-of-the-art approaches under noisy multi-label conditions.

cs.CV↗

arXiv2025

SWA-PF: Semantic-Weighted Adaptive Particle Filter for Memory-Efficient 4-DoF UAV Localization in GNSS-Denied Environments

Vision-based Unmanned Aerial Vehicle (UAV) localization systems have been extensively investigated for Global Navigation Satellite System (GNSS)-denied environments. However, existing retrieval-based approaches face limitations in dataset availability and persistent challenges including suboptimal real-time performance, environmental sensitivity, and limited generalization capability, particularly in dynamic or temporally varying environments. To overcome these limitations, we present a large-scale Multi-Altitude Flight Segments dataset (MAFS) for variable altitude scenarios and propose a novel Semantic-Weighted Adaptive Particle Filter (SWA-PF) method. This approach integrates robust semantic features from both UAV-captured images and satellite imagery through two key innovations: a semantic weighting mechanism and an optimized particle filtering architecture. Evaluated using our dataset, the proposed method achieves 10x computational efficiency gain over feature extraction methods, maintains global positioning errors below 10 meters, and enables rapid 4 degree of freedom (4-DoF) pose estimation within seconds using accessible low-resolution satellite maps. Code and dataset will be available at https://github.com/YuanJiayuuu/SWA-PF.

cs.CV↗

arXiv2025

Three dimensional seepage analysis using a polyhedral scaled boundary finite element method

This work presents a polyhedral scaled boundary finite element method (PSBFEM) for three dimensional seepage analysis. We first derive the scaled boundary formulation for 3D seepage problems, and subsequently incorporate Wachspress shape functions to construct shape functions over arbitrary polygonal elements, thereby establishing the foundation of the proposed polyhedral SBFEM. The method combines the semi-analytical nature of the SBFEM with the geometric flexibility of polyhedral and octree meshes, making it well-suited for complex seepage simulations. The PSBFEM is implemented within the ABAQUS UEL framework to facilitate steady-state, transient, and free-surface seepage analyses. A series of numerical examples are conducted to verify the accuracy, efficiency, and convergence properties of the proposed approach, including benchmark tests and applications with intricate geometries. The results demonstrate that the PSBFEM achieves higher accuracy and faster convergence than conventional FEM, particularly when using hybrid octree meshes with local refinement. This framework provides a robust and efficient computational tool for three-dimensional seepage analysis in geotechnical and hydraulic engineering applications.

math.NA↗

arXiv2025

Improving the Transferability of Adversarial Examples by Inverse Knowledge Distillation

In recent years, the rapid development of deep neural networks has brought increased attention to the security and robustness of these models. While existing adversarial attack algorithms have demonstrated success in improving adversarial transferability, their performance remains suboptimal due to a lack of consideration for the discrepancies between target and source models. To address this limitation, we propose a novel method, Inverse Knowledge Distillation (IKD), designed to enhance adversarial transferability effectively. IKD introduces a distillation-inspired loss function that seamlessly integrates with gradient-based attack methods, promoting diversity in attack gradients and mitigating overfitting to specific model architectures. By diversifying gradients, IKD enables the generation of adversarial samples with superior generalization capabilities across different models, significantly enhancing their effectiveness in black-box attack scenarios. Extensive experiments on the ImageNet dataset validate the effectiveness of our approach, demonstrating substantial improvements in the transferability and attack success rates of adversarial samples across a wide range of models.

cs.LG↗

arXiv2024

SABER: Model-agnostic Backdoor Attack on Chain-of-Thought in Neural Code Generation

Recent studies have proposed integrating Chain-of-Thought (CoT) reasoning to further enhance the reliability of Code Language Models (CLMs) in generating code, a step-by-step approach that breaks down complex programming tasks into manageable sub-problems. Advances in this area have introduced CoT models, specifically designed to integrate CoT reasoning effectively into language models, achieving notable improvements in code generation. Despite these advancements, the security of CoT models has not been systematically studied. In this study, we aim to fill this gap by investigating the vulnerability of CoT models to backdoor injection in code generation tasks. To address this, we propose a model-agnostic backdoor attack method SABER (Self-Attention-BasEd backdooR) based on the self-attention mechanism. SABER begins by selecting a malicious output as the backdoor using code mutation operations. It then identifies the tokens most relevant to poisoned content by analyzing self-attention scores in the CodeBERT model. Finally, it mimicks user behavior to generate adaptive and natural triggers. Our experiments on HumanEval-CoT and OpenEval-CoT test sets demonstrate that CoT models are susceptible to backdoor attacks via data poisoning. Taking the HumanEval-CoT dataset as an example, SABER achieves an ASR of 80.95%, representing an improvement of 33.33% over RIPPLe and a substantial 4.76% enhancement compared to BadPre. Further evaluations using ONION for automated detection and human studies reveal that SABER is stealthier and harder to detect, bypassing 61.90% of automated detection, with a human detection rate of just 3.17%. Our findings reveal that backdoors can be injected into CoT models to manipulate downstream code generation tasks. This highlights the urgent need for further research to understand and mitigate the security vulnerabilities in CoT models.

cs.SE↗

arXiv2024

A polyhedral scaled boundary finite element method solving three-dimensional heat conduction problems

In this study, we derived a three-dimensional scaled boundary finite element formulation for heat conduction problems. By incorporating Wachspress shape functions, a polyhedral scaled boundary finite element method (PSBFEM) was proposed to address heat conduction challenges in complex geometries. To address the complexity of traditional methods, this work introduced polygonal discretization techniques that simplified the topological structure of the polyhedral mesh and effectively integrated polyhedral and octree meshes, thereby reducing the number of element faces and enhancing mesh efficiency to accommodate intricate shapes. The developed formulation supported both steady-state and transient heat conduction analyses and was implemented in ABAQUS through a user-defined element (UEL). Through a series of numerical examples, the accuracy and convergence of the proposed method were validated. The results indicated that the PSBFEM consistently achieved higher accuracy than the FEM as the mesh was refined. The polyhedral elements offered a computationally efficient solution for complex simulations, significantly reducing computational costs.Additionally, by utilizing the octree mesh parent element acceleration technique, the computational efficiency of PSBFEM surpassed that of the FEM.

math.NA↗

arXiv2017

A Workload-Specific Memory Capacity Configuration Approach for In-Memory Data Analytic Platforms

We propose WSMC, a workload-specific memory capacity configuration approach for the Spark workloads, which guides users on the memory capacity configuration with the accurate prediction of the workload's memory requirement under various input data size and parameter settings.First, WSMC classifies the in-memory computing workloads into four categories according to the workloads' Data Expansion Ratio. Second, WSMC establishes a memory requirement prediction model with the consideration of the input data size, the shuffle data size, the parallelism of the workloads and the data block size. Finally, for each workload category, WSMC calculates the shuffle data size in the prediction model in a workload-specific way. For the ad-hoc workload, WSMC can profile its Data Expansion Ratio with small-sized input data and decide the category that the workload falls into. Users can then determine the accurate configuration in accordance with the corresponding memory requirement prediction.Through the comprehensive evaluations with SparkBench workloads, we found that, contrasting with the default configuration, configuration with the guide of WSMC can save over 40% memory capacity with the workload performance slight degradation (only 5%), and compared to the proper configuration found out manually, the configuration with the guide of WSMC leads to only 7% increase in the memory waste with the workload's performance slight improvement (about 1%)

cs.DC↗