arXiv subjects

Haining Wang

Publications and source records attributed to Haining Wang.

At least 37 records · Page 2Linked to original sources

arXiv2025

Whose Journey Matters? Investigating Identity Biases in Large Language Models (LLMs) for Travel Planning Assistance

As large language models (LLMs) become increasingly integral to the hospitality and tourism industry, concerns about their fairness in serving diverse identity groups persist. Grounded in social identity theory and sociotechnical systems theory, this study examines ethnic and gender biases in travel recommendations generated by LLMs. Using fairness probing, we analyze outputs from three leading open-source LLMs. The results show that test accuracy for both ethnicity and gender classifiers exceed random chance. Analysis of the most influential features reveals the presence of stereotype bias in LLM-generated recommendations. We also found hallucinations among these features, occurring more frequently in recommendations for minority groups. These findings indicate that LLMs exhibit ethnic and gender bias when functioning as travel planning assistants. This study underscores the need for bias mitigation strategies to improve the inclusivity and reliability of generative AI-driven travel planning assistance.

cs.AI↗

arXiv2025

Dynamic Vulnerability Patching for Heterogeneous Embedded Systems Using Stack Frame Reconstruction

Existing dynamic vulnerability patching techniques are not well-suited for embedded devices, especially mission-critical ones such as medical equipment, as they have limited computational power and memory but uninterrupted service requirements. Those devices often lack sufficient idle memory for dynamic patching, and the diverse architectures of embedded systems further complicate the creation of patch triggers that are compatible across various system kernels and hardware platforms. To address these challenges, we propose a hot patching framework called StackPatch that facilitates patch development based on stack frame reconstruction. StackPatch introduces different triggering strategies to update programs stored in memory units. We leverage the exception-handling mechanisms commonly available in embedded processors to enhance StackPatch's adaptability across different processor architectures for control flow redirection. We evaluated StackPatch on embedded devices featuring three major microcontroller (MCU) architectures: ARM, RISC-V, and Xtensa. In the experiments, we used StackPatch to successfully fix 102 publicly disclosed vulnerabilities in real-time operating systems (RTOS). We applied patching to medical devices, soft programmable logic controllers (PLCs), and network services, with StackPatch consistently completing each vulnerability remediation in less than 260 MCU clock cycles.

cs.CR↗

arXiv2025

DALR: Dual-level Alignment Learning for Multimodal Sentence Representation Learning

Previous multimodal sentence representation learning methods have achieved impressive performance. However, most approaches focus on aligning images and text at a coarse level, facing two critical challenges:cross-modal misalignment bias and intra-modal semantic divergence, which significantly degrade sentence representation quality. To address these challenges, we propose DALR (Dual-level Alignment Learning for Multimodal Sentence Representation). For cross-modal alignment, we propose a consistency learning module that softens negative samples and utilizes semantic similarity from an auxiliary task to achieve fine-grained cross-modal alignment. Additionally, we contend that sentence relationships go beyond binary positive-negative labels, exhibiting a more intricate ranking structure. To better capture these relationships and enhance representation quality, we integrate ranking distillation with global intra-modal alignment learning. Comprehensive experiments on semantic textual similarity (STS) and transfer (TR) tasks validate the effectiveness of our approach, consistently demonstrating its superiority over state-of-the-art baselines.

cs.CL↗

arXiv2025

Learnable-Differentiable Finite Volume Solver for Accelerated Simulation of Flows

Simulation of fluid flows is crucial for modeling physical phenomena like meteorology, aerodynamics, and biomedicine. Classical numerical solvers often require fine spatiotemporal grids to satisfy stability, consistency, and convergence conditions, leading to substantial computational costs. Although machine learning has demonstrated better efficiency, they typically suffer from issues of interpretability, generalizability, and data dependency. Hence, we propose a learnable and differentiable finite volume solver, called LDSolver, designed for efficient and accurate simulation of fluid flows on spatiotemporal coarse grids. LDSolver comprises two key components: (1) a differentiable finite volume solver, and (2) an learnable module providing equivalent approximation for fluxes (derivatives and interpolations), and temporal error correction on coarse grids. Even with limited training data (e.g., only a few trajectories), our model could accelerate the simulation while maintaining a high accuracy with superior generalizability. Experiments on different flow systems (e.g., Burgers, decaying, forced and shear flows) show that LDSolver achieves state-of-the-art performance, surpassing baseline models with notable margins.

cs.LG↗

arXiv2025

SlotPi: Physics-informed Object-centric Reasoning Models

Understanding and reasoning about dynamics governed by physical laws through visual observation, akin to human capabilities in the real world, poses significant challenges. Currently, object-centric dynamic simulation methods, which emulate human behavior, have achieved notable progress but overlook two critical aspects: 1) the integration of physical knowledge into models. Humans gain physical insights by observing the world and apply this knowledge to accurately reason about various dynamic scenarios; 2) the validation of model adaptability across diverse scenarios. Real-world dynamics, especially those involving fluids and objects, demand models that not only capture object interactions but also simulate fluid flow characteristics. To address these gaps, we introduce SlotPi, a slot-based physics-informed object-centric reasoning model. SlotPi integrates a physical module based on Hamiltonian principles with a spatio-temporal prediction module for dynamic forecasting. Our experiments highlight the model's strengths in tasks such as prediction and Visual Question Answering (VQA) on benchmark and fluid datasets. Furthermore, we have created a real-world dataset encompassing object interactions, fluid dynamics, and fluid-object interactions, on which we validated our model's capabilities. The model's robust performance across all datasets underscores its strong adaptability, laying a foundation for developing more advanced world models.

cs.CV↗

arXiv2025

DGIQA: Depth-guided Feature Attention and Refinement for Generalizable Image Quality Assessment

A long-held challenge in no-reference image quality assessment (NR-IQA) learning from human subjective perception is the lack of objective generalization to unseen natural distortions. To address this, we integrate a novel Depth-Guided cross-attention and refinement (Depth-CAR) mechanism, which distills scene depth and spatial features into a structure-aware representation for improved NR-IQA. This brings in the knowledge of object saliency and relative contrast of the scene for more discriminative feature learning. Additionally, we introduce the idea of TCB (Transformer-CNN Bridge) to fuse high-level global contextual dependencies from a transformer backbone with local spatial features captured by a set of hierarchical CNN (convolutional neural network) layers. We implement TCB and Depth-CAR as multimodal attention-based projection functions to select the most informative features, which also improve training time and inference efficiency. Experimental results demonstrate that our proposed DGIQA model achieves state-of-the-art (SOTA) performance on both synthetic and authentic benchmark datasets. More importantly, DGIQA outperforms SOTA models on cross-dataset evaluations as well as in assessing natural image distortions such as low-light effects, hazy conditions, and lens flares.

cs.CV↗

arXiv2025

An Analysis of Malicious Packages in Open-Source Software in the Wild

The open-source software (OSS) ecosystem suffers from security threats caused by malware.However, OSS malware research has three limitations: a lack of high-quality datasets, a lack of malware diversity, and a lack of attack campaign contexts. In this paper, we first build the largest dataset of 24,356 malicious packages from online sources, then propose a knowledge graph to represent the OSS malware corpus and conduct malware analysis in the wild.Our main findings include (1) it is essential to collect malicious packages from various online sources because their data overlapping degrees are small;(2) despite the sheer volume of malicious packages, many reuse similar code, leading to a low diversity of malware;(3) only 28 malicious packages were repeatedly hidden via dependency libraries of 1,354 malicious packages, and dependency-hidden malware has a shorter active time;(4) security reports are the only reliable source for disclosing the malware-based context. Index Terms: Malicious Packages, Software Analysis

cs.CR↗

arXiv2025

Science Out of Its Ivory Tower: Improving Accessibility with Reinforcement Learning

A vast amount of scholarly work is published daily, yet much of it remains inaccessible to the general public due to dense jargon and complex language. To address this challenge in science communication, we introduce a reinforcement learning framework that fine-tunes a language model to rewrite scholarly abstracts into more comprehensible versions. Guided by a carefully balanced combination of word- and sentence-level accessibility rewards, our language model effectively substitutes technical terms with more accessible alternatives, a task which models supervised fine-tuned or guided by conventional readability measures struggle to accomplish. Our best model adjusts the readability level of scholarly abstracts by approximately six U.S. grade levels -- in other words, from a postgraduate to a high school level. This translates to roughly a 90% relative boost over the supervised fine-tuning baseline, all while maintaining factual accuracy and high-quality language. An in-depth analysis of our approach shows that balanced rewards lead to systematic modifications in the base model, likely contributing to smoother optimization and superior performance. We envision this work as a step toward bridging the gap between scholarly research and the general public, particularly younger readers and those without a college degree.

cs.CL↗

arXiv2025

Partner in Crime: Boosting Targeted Poisoning Attacks against Federated Learning

Federated Learning (FL) exposes vulnerabilities to targeted poisoning attacks that aim to cause misclassification specifically from the source class to the target class. However, using well-established defense frameworks, the poisoning impact of these attacks can be greatly mitigated. We introduce a generalized pre-training stage approach to Boost Targeted Poisoning Attacks against FL, called BoTPA. Its design rationale is to leverage the model update contributions of all data points, including ones outside of the source and target classes, to construct an Amplifier set, in which we falsify the data labels before the FL training process, as a means to boost attacks. We comprehensively evaluate the effectiveness and compatibility of BoTPA on various targeted poisoning attacks. Under data poisoning attacks, our evaluations reveal that BoTPA can achieve a median Relative Increase in Attack Success Rate (RI-ASR) between 15.3% and 36.9% across all possible source-target class combinations, with varying percentages of malicious clients, compared to its baseline. In the context of model poisoning, BoTPA attains RI-ASRs ranging from 13.3% to 94.7% in the presence of the Krum and Multi-Krum defenses, from 2.6% to 49.2% under the Median defense, and from 2.9% to 63.5% under the Flame defense.

cs.CR↗

arXiv2025

HoneyGPT: Breaking the Trilemma in Terminal Honeypots with Large Language Model

Honeypots, as a strategic cyber-deception mechanism designed to emulate authentic interactions and bait unauthorized entities, often struggle with balancing flexibility, interaction depth, and deception. They typically fail to adapt to evolving attacker tactics, with limited engagement and information gathering. Fortunately, the emergent capabilities of large language models and innovative prompt-based engineering offer a transformative shift in honeypot technologies. This paper introduces HoneyGPT, a pioneering shell honeypot architecture based on ChatGPT, characterized by its cost-effectiveness and proactive engagement. In particular, we propose a structured prompt engineering framework that incorporates chain-of-thought tactics to improve long-term memory and robust security analytics, enhancing deception and engagement. Our evaluation of HoneyGPT comprises a baseline comparison based on a collected dataset and a three-month field evaluation. The baseline comparison demonstrates HoneyGPT's remarkable ability to strike a balance among flexibility, interaction depth, and deceptive capability. The field evaluation further validates HoneyGPT's superior performance in engaging attackers more deeply and capturing a wider array of novel attack vectors.

cs.CR↗

arXiv2025

UCFE: A User-Centric Financial Expertise Benchmark for Large Language Models

This paper introduces the UCFE: User-Centric Financial Expertise benchmark, an innovative framework designed to evaluate the ability of large language models (LLMs) to handle complex real-world financial tasks. UCFE benchmark adopts a hybrid approach that combines human expert evaluations with dynamic, task-specific interactions to simulate the complexities of evolving financial scenarios. Firstly, we conducted a user study involving 804 participants, collecting their feedback on financial tasks. Secondly, based on this feedback, we created our dataset that encompasses a wide range of user intents and interactions. This dataset serves as the foundation for benchmarking 11 LLMs services using the LLM-as-Judge methodology. Our results show a significant alignment between benchmark scores and human preferences, with a Pearson correlation coefficient of 0.78, confirming the effectiveness of the UCFE dataset and our evaluation approach. UCFE benchmark not only reveals the potential of LLMs in the financial domain but also provides a robust framework for assessing their performance and user satisfaction.

q-fin.CP↗

arXiv2024

ViTGuard: Attention-aware Detection against Adversarial Examples for Vision Transformer

The use of transformers for vision tasks has challenged the traditional dominant role of convolutional neural networks (CNN) in computer vision (CV). For image classification tasks, Vision Transformer (ViT) effectively establishes spatial relationships between patches within images, directing attention to important areas for accurate predictions. However, similar to CNNs, ViTs are vulnerable to adversarial attacks, which mislead the image classifier into making incorrect decisions on images with carefully designed perturbations. Moreover, adversarial patch attacks, which introduce arbitrary perturbations within a small area, pose a more serious threat to ViTs. Even worse, traditional detection methods, originally designed for CNN models, are impractical or suffer significant performance degradation when applied to ViTs, and they generally overlook patch attacks. In this paper, we propose ViTGuard as a general detection method for defending ViT models against adversarial attacks, including typical attacks where perturbations spread over the entire input and patch attacks. ViTGuard uses a Masked Autoencoder (MAE) model to recover randomly masked patches from the unmasked regions, providing a flexible image reconstruction strategy. Then, threshold-based detectors leverage distinctive ViT features, including attention maps and classification (CLS) token representations, to distinguish between normal and adversarial samples. The MAE model does not involve any adversarial samples during training, ensuring the effectiveness of our detectors against unseen attacks. ViTGuard is compared with seven existing detection methods under nine attacks across three datasets. The evaluation results show the superiority of ViTGuard over existing detectors. Finally, considering the potential detection evasion, we further demonstrate ViTGuard's robustness against adaptive attacks for evasion.

cs.CV↗

arXiv2024

FedMADE: Robust Federated Learning for Intrusion Detection in IoT Networks Using a Dynamic Aggregation Method

The rapid proliferation of Internet of Things (IoT) devices across multiple sectors has escalated serious network security concerns. This has prompted ongoing research in Machine Learning (ML)-based Intrusion Detection Systems (IDSs) for cyber-attack classification. Traditional ML models require data transmission from IoT devices to a centralized server for traffic analysis, raising severe privacy concerns. To address this issue, researchers have studied Federated Learning (FL)-based IDSs that train models across IoT devices while keeping their data localized. However, the heterogeneity of data, stemming from distinct vulnerabilities of devices and complexity of attack vectors, poses a significant challenge to the effectiveness of FL models. While current research focuses on adapting various ML models within the FL framework, they fail to effectively address the issue of attack class imbalance among devices, which significantly degrades the classification accuracy of minority attacks. To overcome this challenge, we introduce FedMADE, a novel dynamic aggregation method, which clusters devices by their traffic patterns and aggregates local models based on their contributions towards overall performance. We evaluate FedMADE against other FL algorithms designed for non-IID data and observe up to 71.07% improvement in minority attack classification accuracy. We further show that FedMADE is robust to poisoning attacks and incurs only a 4.7% (5.03 seconds) latency overhead in each communication round compared to FedAvg, without increasing the computational load of IoT devices.

cs.CR↗

arXiv2024

Simplifying Scholarly Abstracts for Accessible Digital Libraries

Standing at the forefront of knowledge dissemination, digital libraries curate vast collections of scientific literature. However, these scholarly writings are often laden with jargon and tailored for domain experts rather than the general public. As librarians, we strive to offer services to a diverse audience, including those with lower reading levels. To extend our services beyond mere access, we propose fine-tuning a language model to rewrite scholarly abstracts into more comprehensible versions, thereby making scholarly literature more accessible when requested. We began by introducing a corpus specifically designed for training models to simplify scholarly abstracts. This corpus consists of over three thousand pairs of abstracts and significance statements from diverse disciplines. We then fine-tuned four language models using this corpus. The outputs from the models were subsequently examined both quantitatively for accessibility and semantic coherence, and qualitatively for language quality, faithfulness, and completeness. Our findings show that the resulting models can improve readability by over three grade levels, while maintaining fidelity to the original content. Although commercial state-of-the-art models still hold an edge, our models are much more compact, can be deployed locally in an affordable manner, and alleviate the privacy concerns associated with using commercial models. We envision this work as a step toward more inclusive and accessible libraries, improving our services for young readers and those without a college degree.

cs.CL↗

arXiv2024

Tactics, Techniques, and Procedures (TTPs) in Interpreted Malware: A Zero-Shot Generation with Large Language Models

Nowadays, the open-source software (OSS) ecosystem suffers from security threats of software supply chain (SSC) attacks. Interpreted OSS malware plays a vital role in SSC attacks, as criminals have an arsenal of attack vectors to deceive users into installing malware and executing malicious activities. In this paper, we introduce tactics, techniques, and procedures (TTPs) proposed by MITRE ATT\&CK into the interpreted malware analysis to characterize different phases of an attack lifecycle. Specifically, we propose GENTTP, a zero-shot approach to extracting a TTP of an interpreted malware package. GENTTP leverages large language models (LLMs) to automatically generate a TTP, where the input is a malicious package, and the output is a deceptive tactic and an execution tactic of attack vectors. To validate the effectiveness of GENTTP, we collect two datasets for evaluation: a dataset with ground truth labels and a large dataset in the wild. Experimental results show that GENTTP can generate TTPs with high accuracy and efficiency. To demonstrate GENTTP's benefits, we build an LLM-based Chatbot from 3,700+ PyPI malware's TTPs. We further conduct a quantitative analysis of malware's TTPs at a large scale. Our main findings include: (1) many OSS malicious packages share a relatively stable TTP, even with the increasing emergence of malware and attack campaigns, (2) a TTP reflects characteristics of a malware-based attack, and (3) an attacker's intent behind the malware is linked to a TTP.

cs.CR↗

arXiv2024

TimeTravel: Real-time Timing Drift Attack on System Time Using Acoustic Waves

Real-time Clock (RTC) has been widely used in various real-time systems to provide precise system time. In this paper, we reveal a new security vulnerability of the RTC circuit, where the internal storage time or timestamp can be arbitrarily modified forward or backward. The security threat of dynamic modifications of system time caused by this vulnerability is called TimeTravel. Based on acoustic resonance and piezoelectric effects, TimeTravel applies acoustic guide waves to the quartz crystal, thereby adjusting the characteristics of the oscillating signal transmitted into the RTC circuit. By manipulating the parameters of acoustic waves, TimeTravel can accelerate or decelerate the timing speed of system time at an adjustable rate, resulting in the relative drift of the timing, which can pose serious safety threats. To assess the severity of TimeTravel, we examine nine modules and two commercial devices under the RTC circuit. The experimental results show that TimeTravel can drift system time forward and backward at a chosen speed with a maximum 93% accuracy. Our analysis further shows that TimeTravel can maintain an attack success rate of no less than 77% under environments with typical obstacle items.

cs.CR↗

arXiv2024

Pathfinder: Exploring Path Diversity for Assessing Internet Censorship Inconsistency

Internet censorship is typically enforced by authorities to achieve information control for a certain group of Internet users. So far existing censorship studies have primarily focused on country-level characterization because (1) in many cases, censorship is enabled by governments with nationwide policies and (2) it is usually hard to control how the probing packets are routed to trigger censorship in different networks inside a country. However, the deployment and implementation of censorship could be highly diverse at the ISP level. In this paper, we investigate Internet censorship from a different perspective by scrutinizing the diverse censorship deployment inside a country. Specifically, by leveraging an end-to-end measurement framework, we deploy multiple geo-distributed back-end control servers to explore various paths from one single vantage point. The generated traffic with the same domain but different control servers' IPs could be forced to traverse different transit networks, thereby being examined by different censorship devices if present. Through our large-scale experiments and in-depth investigation, we reveal that the diversity of Internet censorship caused by different routing paths inside a country is prevalent, implying that (1) the implementations of centralized censorship are commonly incomplete or flawed and (2) decentralized censorship is also common. Moreover, we identify that different hosting platforms also result in inconsistent censorship activities due to different peering relationships with the ISPs in a country. Finally, we present extensive case studies in detail to illustrate the configurations that lead to censorship inconsistency and explore the causes.

cs.CR↗

arXiv2024

AutoFirm: Automatically Identifying Reused Libraries inside IoT Firmware at Large-Scale

The Internet of Things (IoT) has become indispensable to our daily lives and work. Unfortunately, developers often reuse software libraries in the IoT firmware, leading to a major security concern. If vulnerabilities or insecure versions of these libraries go unpatched, a massive number of IoT devices can be impacted. In this paper, we propose the AutoFirm, an automated tool for detecting reused libraries in IoT firmware at a large scale. Specifically, AutoFirm leverages the syntax information (library name and version) to determine whether IoT firmware reuses the libraries. We conduct a large-scale empirical study of reused libraries of IoT firmware, investigating more than 6,900+ firmware and 2,700+ distinct vulnerabilities affecting 11,300+ vulnerable versions from 349 open-source software libraries. Leveraging this diverse information set, we conduct a qualitative assessment of vulnerable library versions to understand security gaps and the misplaced trust of libraries in IoT firmware. Our research reveals that: manufacturers neglected to update outdated libraries for IoT firmware in 67.3\% of cases; on average, outdated libraries persisted for over 1.34 years prior to remediation; vulnerabilities of software libraries have posed server threats to widespread IoT devices.

cs.CR↗