arXiv subjects

Jing Sun

Publications and source records attributed to Jing Sun.

At least 37 records · Page 2Linked to original sources

arXiv2026

Noise Reduction for Pufferfish Privacy: A Practical Noise Calibration Method

This paper introduces a relaxed noise calibration method to enhance data utility while attaining pufferfish privacy. This work builds on the existing $1$-Wasserstein (Kantorovich) mechanism by alleviating the existing overly strict condition that leads to excessive noise, and proposes a practical mechanism design algorithm as a general solution. We prove that a strict noise reduction by our approach always exists compared to $1$-Wasserstein mechanism for all privacy budgets $ε$ and prior beliefs, and the noise reduction (also represents improvement on data utility) gains increase significantly for low privacy budget situations--which are commonly seen in real-world deployments. We also analyze the variation and optimality of the noise reduction with different prior distributions. Moreover, all the properties of the noise reduction still exist in the worst-case $1$-Wasserstein mechanism we introduced, when the additive noise is largest. We further show that the worst-case $1$-Wasserstein mechanism is equivalent to the $\ell_1$-sensitivity method. Experimental results on three real-world datasets demonstrate $47\%$ to $87\%$ improvement in data utility.

cs.CR↗

arXiv2026

Battery State of Health Estimation and Incremental Capacity Analysis under Dynamic Charging Profile Using Neural Networks

Incremental capacity analysis (ICA) and differential voltage analysis (DVA) are two effective approaches for battery degradation monitoring. One limiting factor for their real-world application is that they require constant-current (CC) charging profiles. This research removes this limitation and proposes an approach that extends ICA/DVA-based degradation monitoring from CC charging to dynamic charging profiles. A novel concept of virtual incremental capacity (VIC) and virtual differential voltage (VDV) is proposed. Then, two related convolutional neural networks (CNNs), called U-Net and Conv-Net, are proposed to construct VIC/VDV curves and estimate the state of health (SOH) from dynamic charging profiles across any state-of-charge (SOC) range that satisfies some constraints. Finally, two CNNs called Mobile U-Net and Mobile-Net are proposed as replacements for the U-Net and Conv-Net, respectively, to reduce the computational footprint and memory requirements, while keeping similar performance. Using an extensive experimental dataset of battery modules, the proposed CNNs are demonstrated to provide accurate VIC/VDV curves and enable ICA/DVA-based battery degradation monitoring under various fast-charging protocols and different SOC ranges.

eess.SY↗

arXiv2025

Pure Vision Language Action (VLA) Models: A Comprehensive Survey

The emergence of Vision Language Action (VLA) models marks a paradigm shift from traditional policy-based control to generalized robotics, reframing Vision Language Models (VLMs) from passive sequence generators into active agents for manipulation and decision-making in complex, dynamic environments. This survey delves into advanced VLA methods, aiming to provide a clear taxonomy and a systematic, comprehensive review of existing research. It presents a comprehensive analysis of VLA applications across different scenarios and classifies VLA approaches into several paradigms: autoregression-based, diffusion-based, reinforcement-based, hybrid, and specialized methods; while examining their motivations, core strategies, and implementations in detail. In addition, foundational datasets, benchmarks, and simulation platforms are introduced. Building on the current VLA landscape, the review further proposes perspectives on key challenges and future directions to advance research in VLA models and generalizable robotics. By synthesizing insights from over three hundred recent studies, this survey maps the contours of this rapidly evolving field and highlights the opportunities and challenges that will shape the development of scalable, general-purpose VLA methods.

cs.RO↗

arXiv2025

FTSmartAudit: A Knowledge Distillation-Enhanced Framework for Automated Smart Contract Auditing Using Fine-Tuned LLMs

The rapid growth of blockchain technology has driven the widespread adoption of smart contracts. However, their inherent vulnerabilities have led to significant financial losses. Traditional auditing methods, while essential, struggle to keep pace with the increasing complexity and scale of smart contracts. Large Language Models (LLMs) offer promising capabilities for automating vulnerability detection, but their adoption is often limited by high computational costs. Although prior work has explored leveraging large models through agents or workflows, relatively little attention has been given to improving the performance of smaller, fine-tuned models--a critical factor for achieving both efficiency and data privacy. In this paper, we introduce HKT-SmartAudit, a framework for developing lightweight models optimized for smart contract auditing. It features a multi-stage knowledge distillation pipeline that integrates classical distillation, external domain knowledge, and reward-guided learning to transfer high-quality insights from large teacher models. A single-task learning strategy is employed to train compact student models that maintain high accuracy and robustness while significantly reducing computational overhead. Experimental results show that our distilled models outperform both commercial tools and larger models in detecting complex vulnerabilities and logical flaws, offering a practical, secure, and scalable solution for smart contract auditing. The source code is available at Github repository.

cs.CR↗

arXiv2025

Modeling and Dynamic Simulation of a Hybrid Wind-Wave System on a Hexagonal Semi-Submersible Platform

Offshore renewable energy systems offer promising solutions for sustainable power generation, yet most existing platforms harvest either wind or wave energy in isolation. This study presents a hybrid floating offshore platform that integrates a wind turbine with three oscillating surge wave energy converters (WECs) into a hexagonal semi-submersible structure. In this configuration, the flaps are integrated with the platform geometry to provide both energy extraction and hydrodynamic stability. A modeling and simulation framework was developed using WEC-Sim and benchmarked against the NREL 5 MW semisubmersible reference. Metacentric height analysis confirmed hydrostatic stability across a range of prescribed flap angles. Sensitivity analysis of twelve geometric variables identified flap dimensions and tower length as dominant drivers of stability, energy capture, and tower stress. Time-domain simulations revealed dependence on wave incidence angle, with variations in flap power sharing, capture width ratio (CWR), and platform response. The feasibility of using flap sweeps to modulate pitch motion was also demonstrated. Annual energy production (AEP) estimates based on site-specific data indicate 16.86 GWh from wind and 3.65 GWh from wave energy, with WECs contributing about 18% of the total. These results highlight the potential of integrated wind-wave platforms and point toward future studies on structural modeling and advanced control.

eess.SY↗

arXiv2025

Goal-oriented Backdoor Attack against Vision-Language-Action Models via Physical Objects

Recent advances in vision-language-action (VLA) models have greatly improved embodied AI, enabling robots to follow natural language instructions and perform diverse tasks. However, their reliance on uncurated training datasets raises serious security concerns. Existing backdoor attacks on VLAs mostly assume white-box access and result in task failures instead of enforcing specific actions. In this work, we reveal a more practical threat: attackers can manipulate VLAs by simply injecting physical objects as triggers into the training dataset. We propose goal-oriented backdoor attacks (GoBA), where the VLA behaves normally in the absence of physical triggers but executes predefined and goal-oriented actions in the presence of physical triggers. Specifically, based on a popular VLA benchmark LIBERO, we introduce BadLIBERO that incorporates diverse physical triggers and goal-oriented backdoor actions. In addition, we propose a three-level evaluation that categorizes the victim VLA's actions under GoBA into three states: nothing to do, try to do, and success to do. Experiments show that GoBA enables the victim VLA to successfully achieve the backdoor goal in 97 percentage of inputs when the physical trigger is present, while causing zero performance degradation on clean inputs. Finally, by investigating factors related to GoBA, we find that the action trajectory and trigger color significantly influence attack performance, while trigger size has surprisingly little effect. The code and BadLIBERO dataset are accessible via the project page at https://goba-attack.github.io/.

cs.CR↗

arXiv2025

Distilling Lightweight Language Models for C/C++ Vulnerabilities

The increasing complexity of modern software systems exacerbates the prevalence of security vulnerabilities, posing risks of severe breaches and substantial economic loss. Consequently, robust code vulnerability detection is essential for software security. While Large Language Models (LLMs) have demonstrated remarkable capabilities in natural language processing, their potential for automated code vulnerability detection remains underexplored. This paper presents FineSec, a novel framework that harnesses LLMs through knowledge distillation to enable efficient and precise vulnerability identification in C/C++ codebases. FineSec utilizes knowledge distillation to transfer expertise from large teacher models to compact student models, achieving high accuracy with minimal computational cost. By integrating data preparation, training, evaluation, and continuous learning into a unified, single-task workflow, FineSec offers a streamlined approach. Extensive evaluations on C/C++ codebases demonstrate its superiority over both base models and larger LLMs in identifying complex vulnerabilities and logical flaws, establishing FineSec as a practical and scalable solution for real-world software security. To facilitate reproducibility, the datasets, source code, and experimental results are made publicly available at: https://github.com/yangxiaoxuan123/FineSec_detect.

cs.CR↗

arXiv2025

PAT-Agent: Autoformalization for Model Checking

Recent advances in large language models (LLMs) offer promising potential for automating formal methods. However, applying them to formal verification remains challenging due to the complexity of specification languages, the risk of hallucinated output, and the semantic gap between natural language and formal logic. We introduce PAT-Agent, an end-to-end framework for natural language autoformalization and formal model repair that combines the generative capabilities of LLMs with the rigor of formal verification to automate the construction of verifiable formal models. In PAT-Agent, a Planning LLM first extracts key modeling elements and generates a detailed plan using semantic prompts, which then guides a Code Generation LLM to synthesize syntactically correct and semantically faithful formal models. The resulting code is verified using the Process Analysis Toolkit (PAT) model checker against user-specified properties, and when discrepancies occur, a Repair Loop is triggered to iteratively correct the model using counterexamples. To improve flexibility, we built a web-based interface that enables users, particularly non-FM-experts, to describe, customize, and verify system behaviors through user-LLM interactions. Experimental results on 40 systems show that PAT-Agent consistently outperforms baselines, achieving high verification success with superior efficiency. The ablation studies confirm the importance of both planning and repair components, and the user study demonstrates that our interface is accessible and supports effective formal modeling, even for users with limited formal methods experience.

cs.SE↗

arXiv2025

Adaptive Plan-Execute Framework for Smart Contract Security Auditing

Large Language Models (LLMs) have shown great promise in code analysis and auditing; however, they still struggle with hallucinations and limited context-aware reasoning. We introduce SmartAuditFlow, a novel Plan-Execute framework that enhances smart contract security analysis through dynamic audit planning and structured execution. Unlike conventional LLM-based auditing approaches that follow fixed workflows and predefined steps, SmartAuditFlow dynamically generates and refines audit plans based on the unique characteristics of each smart contract. It continuously adjusts its auditing strategy in response to intermediate LLM outputs and newly detected vulnerabilities, ensuring a more adaptive and precise security assessment. The framework then executes these plans step by step, applying a structured reasoning process to enhance vulnerability detection accuracy while minimizing hallucinations and false positives. To further improve audit precision, SmartAuditFlow integrates iterative prompt optimization and external knowledge sources, such as static analysis tools and Retrieval-Augmented Generation (RAG). This ensures audit decisions are contextually informed and backed by real-world security knowledge, producing comprehensive security reports. Extensive evaluations across multiple benchmarks demonstrate that SmartAuditFlow outperforms existing methods, achieving 100 percent accuracy on common and critical vulnerabilities, 41.2 percent accuracy for comprehensive coverage of known smart contract weaknesses in real-world projects, and successfully identifying all 13 tested CVEs. These results highlight SmartAuditFlow's scalability, cost-effectiveness, and superior adaptability over traditional static analysis tools and contemporary LLM-based approaches, establishing it as a robust solution for automated smart contract auditing.

cs.CR↗

arXiv2025

Streamlining Biomedical Research with Specialized LLMs

In this paper, we propose a novel system that integrates state-of-the-art, domain-specific large language models with advanced information retrieval techniques to deliver comprehensive and context-aware responses. Our approach facilitates seamless interaction among diverse components, enabling cross-validation of outputs to produce accurate, high-quality responses enriched with relevant data, images, tables, and other modalities. We demonstrate the system's capability to enhance response precision by leveraging a robust question-answering model, significantly improving the quality of dialogue generation. The system provides an accessible platform for real-time, high-fidelity interactions, allowing users to benefit from efficient human-computer interaction, precise retrieval, and simultaneous access to a wide range of literature and data. This dramatically improves the research efficiency of professionals in the biomedical and pharmaceutical domains and facilitates faster, more informed decision-making throughout the R\&D process. Furthermore, the system proposed in this paper is available at https://synapse-chat.patsnap.com.

cs.CL↗

arXiv2025

Aligning First, Then Fusing: A Novel Weakly Supervised Multimodal Violence Detection Method

Weakly supervised violence detection refers to the technique of training models to identify violent segments in videos using only video-level labels. Among these approaches, multimodal violence detection, which integrates modalities such as audio and optical flow, holds great potential. Existing methods in this domain primarily focus on designing multimodal fusion models to address modality discrepancies. In contrast, we take a different approach; leveraging the inherent discrepancies across modalities in violence event representation to propose a novel multimodal semantic feature alignment method. This method sparsely maps the semantic features of local, transient, and less informative modalities ( such as audio and optical flow ) into the more informative RGB semantic feature space. Through an iterative process, the method identifies the suitable no-zero feature matching subspace and aligns the modality-specific event representations based on this subspace, enabling the full exploitation of information from all modalities during the subsequent modality fusion stage. Building on this, we design a new weakly supervised violence detection framework that consists of unimodal multiple-instance learning for extracting unimodal semantic features, multimodal alignment, multimodal fusion, and final detection. Experimental results on benchmark datasets demonstrate the effectiveness of our method, achieving an average precision (AP) of 86.07% on the XD-Violence dataset. Our code is available at https://github.com/xjpp2016/MAVD.

cs.CV↗

arXiv2025

Insights into dendritic growth mechanisms in batteries: A combined machine learning and computational study

In recent years, researchers have increasingly sought batteries as an efficient and cost-effective solution for energy storage and supply, owing to their high energy density, low cost, and environmental resilience. However, the issue of dendrite growth has emerged as a significant obstacle in battery development. Excessive dendrite growth during charging and discharging processes can lead to battery short-circuiting, degradation of electrochemical performance, reduced cycle life, and abnormal exothermic events. Consequently, understanding the dendrite growth process has become a key challenge for researchers. In this study, we investigated dendrite growth mechanisms in batteries using a combined machine learning approach, specifically a two-dimensional artificial convolutional neural network (CNN) model, along with computational methods. We developed two distinct computer models to predict dendrite growth in batteries. The CNN-1 model employs standard convolutional neural network techniques for dendritic growth prediction, while CNN-2 integrates additional physical parameters to enhance model robustness. Our results demonstrate that CNN-2 significantly enhances prediction accuracy, offering deeper insights into the impact of physical factors on dendritic growth. This improved model effectively captures the dynamic nature of dendrite formation, exhibiting high accuracy and sensitivity. These findings contribute to the advancement of safer and more reliable energy storage systems.

physics.comp-ph↗

arXiv2025

Artificial and Eddy Viscosity in Large Eddy Simulation Part 1: Temporal and Spatial Schemes

We propose a novel method to quantify artificial dissipation in large eddy simulation. Here, artificial dissipation is defined as the residual of the discrete turbulent kinetic energy (TKE) equation. This method is applied to turbulent channel flow at Reynolds number 180 and 550 using a minimum-dissipation model within the OpenFOAM framework, employing both symmetry-preserving and standard OpenFOAM discretizations. Our analysis reveals that artificial viscosity can both produce and dissipate TKE. To quantify the interaction between sub-grid-scale model contributions and numerical error, we introduce viscous activity parameters, which also allow for evaluating the balance between molecular and non-molecular viscosity. Examining temporal discretization schemes, we find that all methods can deliver accurate flow predictions if an appropriate time step is used. For very small time steps, Forward-Euler is more efficient and accurate than Crank-Nicolson, though a better balance of accuracy and computational efficiency for larger time steps, making it preferable for large-scale applications. On collocated meshes, Van Kan's pressure correction method predicts better results than Chorin's. Our spatial discretization study shows that the symmetry-preserving scheme outperforms standard OpenFOAM schemes, as excessive artificial errors from spatial discretization can suppress the LES model. For the symmetry-preserving scheme, both eddy and artificial viscosities vary similarly, with eddy viscosity being dominant. The role of artificial viscosity changes with mesh refinement: on coarse meshes, it mainly dissipates TKE, while on finer meshes, it can produce TKE, counteracting eddy viscosity. The most accurate Reynolds shear stress predictions occur when non-molecular dissipation is nearly zero. For accurate mean flow kinetic energy, larger non-molecular dissipation is needed on finer meshes.

physics.flu-dyn↗

arXiv2025

Artificial and Eddy Viscosity in Large Eddy Simulation Part 2: Turbulence Models

This is the second part to our companion paper. The novel method to quantify artificial dissipation proposed in Part 1 is further applied in turbulent channel flow at $\mathrm{Re_τ}=180$ using various subgrid-scale models, with an emphasis on minimum-dissipation models (QR and AMD). We found that the amount of artificial viscosity is comparable to the eddy viscosity, but their distributions are essentially opposite. The artificial viscosity can either produce turbulent kinetic energy (TKE) in the near wall region or dissipate TKE in the bulk region. The eddy viscosity is almost uniformly distributed across the wall-normal direction and actively damps TKE at the channel center. An eddy viscosity level of $ν_e/ν< 22\%$ leads to accurate predictions of flow quantities at the current mesh resolution. The optimal coefficients for the QR model combined with symmetry-preserving discretization were found to be C=0.092 for turbulent kinetic energy and C=0.012 for Reynolds stress terms and large scale motions. Adjusting the QR model coefficient changes the artificial viscosity and significantly impact turbulence characteristics: small-scale motions remain unaffected, while larger scales are less accurately captured with larger coefficients. For error quantification, the SGS activity parameter $ν_{rel} = \frac{ν_e}{ν_e + ν}$ helps relate and reflect the error magnitude with the SGS model. Optimal coefficients for the AMD model were also identified. The trends of varying the coefficient are similar to the QR model but differ in the amplitude of changes. The AMD model is more sensitive to the model coefficients and introduces more eddy viscosity in the near-wall region compared to the QR model. This research highlights the importance of balancing numerical and eddy viscosities for accurate LES predictions.

physics.flu-dyn↗

arXiv2025

Minimum-dissipation model and symmetry-preserving discretization for scalar transport in a turbulent flow

This work extends the minimum-dissipation model of large-eddy simulation and symmetry-preserving discretization to account for active or passive scalar transport and complex physical mechanisms.This novel scalar-minimum-dissipation model exhibits several desirable properties. It includes the effect of scalar transport, in addition to shear, on the suppression and production of turbulence. It switches off at no-slip walls, ensuring accurate capture of near-wall flow behavior without needing a wall model. It also switches off in laminar and transitional flows so that it is capable of predicting laminar-turbulent transition. The new scalar-minimum-dissipation model combined with the symmetry-preserving discretization is successfully tested in a differentially heated cavity in OpenFOAM. The results show that the symmetry-preserving discretization significantly improves predictions of flow quantity and heat transfer on highly stretched meshes.

physics.flu-dyn↗

arXiv2025

Nonlinear Magnetics Model for Permanent Magnet Synchronous Machines Capturing Saturation and Temperature Effects

This paper proposes a nonlinear magnetics model for Permanent Magnet Synchronous Machines (PMSMs) that accurately captures the effects of magnetic saturation in the machine iron and variations in rotor temperature on the permanent magnet excitation. The proposed model considers the permanent magnet as a current source rather than the more commonly used flux-linkage source. A comparison of the two modelling approaches is conducted using Finite Element Analysis (FEA) for different machine designs as well as experimental validation, where it is shown that the proposed model has substantially better accuracy. The proposed model decouples magnetic saturation and rotor temperature effects in the current/flux-linkage relationship, allowing for adaptive estimation of the PM excitation.

eess.SY↗

arXiv2025

Physics-Driven Self-Supervised Deep Learning for Free-Surface Multiple Elimination

In recent years, deep learning (DL) has emerged as a promising alternative approach for various seismic processing tasks, including primary estimation (or multiple elimination), a crucial step for accurate subsurface imaging. In geophysics, DL methods are commonly based on supervised learning from large amounts of high-quality labelled data. Instead of relying on traditional supervised learning, in the context of free-surface multiple elimination, we propose a method in which the DL model learns to effectively parameterize the free-surface multiple-free wavefield from the full wavefield by incorporating the underlying physics into the loss computation. This, in turn, yields high-quality estimates without ever being shown any ground truth data. Currently, the network reparameterization is performed independently for each dataset. We demonstrate its effectiveness through tests on both synthetic and field data. We employ industry-standard Surface-Related Multiple Elimination (SRME) using, respectively, global least-squares adaptive subtraction and local least-squares adaptive subtraction as benchmarks. The comparison shows that the proposed method outperforms the benchmarks in estimation accuracy, achieving the most complete primary estimation and the least multiple energy leakage, but at the cost of a higher computational burden.

physics.geo-ph↗

arXiv2025

Physics-Trained Neural Network as Inverse Problem Solver for Potential Fields: An Example of Downward Continuation between Arbitrary Surfaces

Downward continuation is a critical task in potential field processing, including gravity and magnetic fields, which aims to transfer data from one observation surface to another that is closer to the source of the field. Its effectiveness directly impacts the success of detecting and highlighting subsurface anomalous sources. We treat downward continuation as an inverse problem that relies on solving a forward problem defined by the formula for upward continuation, and we propose a new physics-trained deep neural network (DNN)-based solution for this task. We hard-code the upward continuation process into the DNN's learning framework, where the DNN itself learns to act as the inverse problem solver and can perform downward continuation without ever being shown any ground truth data. We test the proposed method on both synthetic magnetic data and real-world magnetic data from West Antarctica. The preliminary results demonstrate its effectiveness through comparison with selected benchmarks, opening future avenues for the combined use of DNNs and established geophysical theories to address broader potential field inverse problems, such as density and geometry modelling.

physics.geo-ph↗