arXiv subjects

Jin Wen

Publications and source records attributed to Jin Wen.

10 recordsLinked to original sources

arXiv2026

VisTCP: A Visualization Framework to Construct Knowledge-Graph-Based Representation for Traditional Chinese Painting

Structured representation can characterize semantic objects and relationships in images. It provides a possible effective way for the semantic understanding of Traditional Chinese Paintings (TCPs) to better support archaeology and art history research. However, most image-oriented structured representation methods perform poorly on TCPs, due to two major challenges: 1) the objects and events of TCPs exhibit substantial differences from modern natural images, which results in semantic misunderstandings of TCPs; and 2) it is difficult to achieve accurate identification of ancient objects and events in TCPs, even for domain experts.In this paper, we propose VisTCP, a visualization framework that combines a TCP-oriented intelligent model and expert knowledge, which enables art historians to achieve trustworthy structured representations of TCPs in a human-in-the-loop manner. Firstly, we conduct a pilot study with three domain experts to build a semantic taxonomy of TCPs. Then, expert-annotated data are used to train a TCP-oriented structured representation model, which can automatically extract meaningful objects and their relationships in TCPs. To inform users of the model uncertainty, we design a joint embedding visualization view to show the differences between expert annotations and model predictions. This allows users to refine the structured representation based on their domain knowledge, enabling iterative optimization of the model. Finally, we conduct a case study, a usage scenario, and expert interviews on a real dataset to demonstrate the effectiveness of VisTCP in supporting the structured representation and semantic understanding of TCPs.

cs.HC↗

arXiv2026

Robustness and Trade-offs for Code LLMs on Protected Code

Code large language models (LLMs) are increasingly used on software artifacts that may be intentionally obfuscated for intellectual-property protection, reverse-engineering resistance, or controlled access. In reverse engineering and security analysis, deobfuscation is commonly treated as the preprocessing step before downstream analysis or inference, yet its utility for code LLM pipelines has not been systematically validated across models and protection methods. We present an execution-based study of seven code LLMs on protected-code translation and completion, spanning source programs from C++, Go, Java, and JavaScript, five obfuscation methods, and three inference protocols: plain, obfuscated, and deobfuscated. Our results show that direct inference on obfuscated code often matches or exceeds inference on restored code. In this controlled benchmark, higher-capability models such as GPT-4.1 and Qwen3-Coder-30B retain about 90% Pass@1 on obfuscated translation inputs, indicating that explicit restoration is often unnecessary. Same-model restoration does recover some obfuscation-induced failures, but it also degrades many cases that already succeed, with lower-capability models showing the largest net losses. Across settings, model capability is the primary factor, while source language and obfuscation method have secondary but consistent effects. Overall, our findings support model-aware pipeline design and indicate that protected-code workflows should be evaluated primarily with execution-based metrics rather than static similarity alone.

cs.SE↗

arXiv2026

Towards Demystifying and Repairing LLM-in-the-Loop Vulnerabilities

Large Language Models (LLMs) have been actively integrated into modern software systems as critical components, introducing a new type of software vulnerability, LLM-in-the-Loop (LiL) vulnerability, in which threats are caused by LLMs. Although some studies have attempted to investigate the impact of LiL vulnerabilities, they have unfortunately failed to clearly distinguish LiL vulnerabilities from conventional ones, leaving the understanding of real-world LiL vulnerabilities an open problem. To address this gap, we first clearly define the scope of LiL vulnerability, and discuss the differences between LiL vulnerabilities and vulnerabilities that exist in LLM systems but are not really caused by LLMs (i.e., LLM-ecosystem vulnerabilities). Then, we construct the first LiL vulnerability dataset, LiLCVE, covering 41 LiL vulnerabilities and 75 LLM-ecosystem vulnerabilities, to facilitate the risk analysis of LLM-integrated software. The analysis of LiLCVE reveals that LiL vulnerabilities have higher severity than LLM-ecosystem vulnerabilities and conventional software vulnerabilities, with 15.5% and 30.3% more critical vulnerabilities, respectively. Furthermore, given the high severity of LiL vulnerabilities and the potential of LLM-based vulnerability repair methods in patching conventional software vulnerabilities. We explore the capabilities of existing widely-used LLM-based methods in repairing vulnerabilities in LiLCVE. Experimental results on 20 agent-model configuration demonstrate that LiL vulnerabilities are far more challenging to fix, with an average decrease of 10.8% Pass@1 rate compared to other types of vulnerabilities. More critically, three categories, Generated Query Execution, Agent Action, and Model Output Rendering, frequently receive 0% repair success rates.

cs.SE↗

arXiv2024

Evaluation and Improvement of Fault Detection for Large Language Models

Large language models (LLMs) have recently achieved significant success across various application domains, garnering substantial attention from different communities. Unfortunately, even for the best LLM, many \textit{faults} still exist that LLM cannot properly predict. Such faults will harm the usability of LLMs in general and could introduce safety issues in reliability-critical systems such as autonomous driving systems. How to quickly reveal these faults in real-world datasets that LLM could face is important, but challenging. The major reason is that the ground truth is necessary but the data labeling process is heavy considering the time and human effort. To handle this problem, in the conventional deep learning testing field, test selection methods have been proposed for efficiently evaluating deep learning models by prioritizing faults. However, despite their importance, the usefulness of these methods on LLMs is unclear, and lack of exploration. In this paper, we conduct the first empirical study to investigate the effectiveness of existing fault detection methods for LLMs. Experimental results on four different tasks~(including both code tasks and natural language processing tasks) and four LLMs~(e.g., LLaMA3 and GPT4) demonstrated that simple methods such as Margin perform well on LLMs but there is still a big room for improvement. Based on the study, we further propose \textbf{MuCS}, a prompt \textbf{Mu}tation-based prediction \textbf{C}onfidence \textbf{S}moothing framework to boost the fault detection capability of existing methods. Concretely, multiple prompt mutation techniques have been proposed to help collect more diverse outputs for confidence smoothing. The results show that our proposed framework significantly enhances existing methods with the improvement of test relative coverage by up to 70.53\%.

cs.SE↗

arXiv2023

A critical review of cyber-physical security for building automation systems

Modern Building Automation Systems (BASs), as the brain that enables the smartness of a smart building, often require increased connectivity both among system components as well as with outside entities, such as optimized automation via outsourced cloud analytics and increased building-grid integrations. However, increased connectivity and accessibility come with increased cyber security threats. BASs were historically developed as closed environments with limited cyber-security considerations. As a result, BASs in many buildings are vulnerable to cyber-attacks that may cause adverse consequences, such as occupant discomfort, excessive energy usage, and unexpected equipment downtime. Therefore, there is a strong need to advance the state-of-the-art in cyber-physical security for BASs and provide practical solutions for attack mitigation in buildings. However, an inclusive and systematic review of BAS vulnerabilities, potential cyber-attacks with impact assessment, detection & defense approaches, and cyber-secure resilient control strategies is currently lacking in the literature. This review paper fills the gap by providing a comprehensive up-to-date review of cyber-physical security for BASs at three levels in commercial buildings: management level, automation level, and field level. The general BASs vulnerabilities and protocol-specific vulnerabilities for the four dominant BAS protocols are reviewed, followed by a discussion on four attack targets and seven potential attack scenarios. The impact of cyber-attacks on BASs is summarized as signal corruption, signal delaying, and signal blocking. The typical cyber-attack detection and defense approaches are identified at the three levels. Cyber-secure resilient control strategies for BASs under attack are categorized into passive and active resilient control schemes. Open challenges and future opportunities are finally discussed.

cs.CR↗

arXiv2023

Development of a Hardware-in-the-loop Testbed for Laboratory Performance Verification of Flexible Building Equipment in Typical Commercial Buildings

The goals of reducing energy costs, shifting electricity peaks, increasing the use of renewable energy, and enhancing the stability of the electric grid can be met in part by fully exploiting the energy flexibility potential of buildings and building equipment. The development of strategies that exploit these flexibilities could be facilitated by publicly available high-resolution datasets illustrating how control of HVAC systems in commercial buildings can be used in different climate zones to shape the energy use profile of a building for grid needs. This article presents the development and integration of a Hardware-In-the-Loop Flexible load Testbed (HILFT) that integrates physical HVAC systems with a simulated building model and simulated occupants with the goal of generating datasets to verify load flexibility of typical commercial buildings. Compared to simulation-only experiments, the hardware-in-the-loop approach captures the dynamics of the physical systems while also allowing efficient testing of various boundary conditions. The HILFT integration in this article is achieved through the co-simulation among various software environments including LabVIEW, MATLAB, and EnergyPlus. Although theoretically viable, such integration has encountered many real-world challenges, such as: 1) how to design the overall data infrastructure to ensure effective, robust, and efficient integration; 2) how to avoid closed-loop hunting between simulated and emulated variables; 3) how to quantify system response times and minimize system delays; and 4) how to assess the overall integration quality. Lessons-learned using the examples of an AHU-VAV system, an air-source heat pump system, and a water-source heat pump system are presented.

eess.SY↗

arXiv2020

Performance evaluation of on-chip wavelength conversion based on InP/In$_{1-x}$Ga$_x$As$_y$P$_{1-y}$ semiconductor waveguide platforms

We propose and design the high confinement InP/In1-xGaxAsyP1-y semiconductor waveguides and report the results of effective wavelength conversion based on this platform. Efficient confinement and mode field area fluctuation at different wavelength is analyzed to achieve the high nonlinear coefficient. The numerical results show that nearly zero phase-mismatch condition can be satisfied through dispersion tailoring of InP/In1-xGaxAsyP1-y waveguides, and the wavelength conversion ranging over 40 nm with the maximum conversion efficiency -26.3 dB is achieved for fixing pump power 100 mW. Meanwhile, the influences of the doping parameter y and pumping wavelength on the bandwidth and conversion efficiency are also discussed and optimized. It is indicated the excellent optical properties of the InP/In1-xGaxAsyP1-y waveguides and pave the way towards direct integration telecom band devices on stand semiconductor platforms.

physics.optics↗

arXiv2019

Reconstruction of the time-dependent source term in a stochastic fractional diffusion equation

In this work, an inverse problem in the fractional diffusion equation with random source is considered. Statistical moments are used of the realizations of single point observation $u(x_0,t,ω).$ We build the representation of the solution $u$ in integral sense, then prove some theoretical results as uniqueness and stability. After that, we establish a numerical algorithm to solve the unknowns, where the mollification method is used.

math.AP↗

arXiv2019

Laser-Induced Dissociative Recombination of Carbon Dioxide

We experimentally investigate laser-induced dissociative recombination of CO$_2$ in linearly polarized strong laser fields with coincidence measurements. Our results show laser-induced dissociation processes originate from an electron recombination process after laser-induced double ionization. After double ionization of CO$_2$, one electron is recaptured by the CO$_2^{2+}$ and localized to O$^+$ or CO$^+$ in the following dissociation process. We found that the probability of electron localization to O$^{+}$ is much higher than that to CO$^+$. Further, our measurements reveal that the recombination probability of the first ionized electron is three times as high as that of the second ionized electron. Our work may trigger further experimental and theoretical studies on involved nuclear and electron dynamics in laser-induced dissociative recombination of molecules and their applications in controlling molecular dissociation with ultrashort laser pulses.

physics.chem-ph↗

arXiv2015

An inexact Picard iteration method for absolute value equation

Recently, a class of inexact Picard iteration method for solving the absolute value equation: $Ax-|x~|=b$ have been proposed in [Optim Lett 8:2191-2202,2014]. To further improve the performance of Picard iteration method, a new inexact Picard iteration method is proposed to solve the absolute value equation. The sufficient conditions for the convergence of the proposed method for the absolute value equation is given. Some numerical experiments are given to demonstrate the effectiveness of the new method.

math.NA↗