arXiv subjects

Zhengyu Zhao

Publications and source records attributed to Zhengyu Zhao.

At least 37 records · Page 2Linked to original sources

arXiv2025

Revisiting Adversarial Patch Defenses on Object Detectors: Unified Evaluation, Large-Scale Dataset, and New Insights

Developing reliable defenses against patch attacks on object detectors has attracted increasing interest. However, we identify that existing defense evaluations lack a unified and comprehensive framework, resulting in inconsistent and incomplete assessments of current methods. To address this issue, we revisit 11 representative defenses and present the first patch defense benchmark, involving 2 attack goals, 13 patch attacks, 11 object detectors, and 4 diverse metrics. This leads to the large-scale adversarial patch dataset with 94 types of patches and 94,000 images. Our comprehensive analyses reveal new insights: (1) The difficulty in defending against naturalistic patches lies in the data distribution, rather than the commonly believed high frequencies. Our new dataset with diverse patch distributions can be used to improve existing defenses by 15.09% AP@0.5. (2) The average precision of the attacked object, rather than the commonly pursued patch detection accuracy, shows high consistency with defense performance. (3) Adaptive attacks can substantially bypass existing defenses, and defenses with complex/stochastic models or universal patch properties are relatively robust. We hope that our analyses will serve as guidance on properly evaluating patch attacks/defenses and advancing their design. Code and dataset are available at https://github.com/Gandolfczjh/APDE, where we will keep integrating new attacks/defenses.

cs.CV↗

arXiv2025

D3: Training-Free AI-Generated Video Detection Using Second-Order Features

The evolution of video generation techniques, such as Sora, has made it increasingly easy to produce high-fidelity AI-generated videos, raising public concern over the dissemination of synthetic content. However, existing detection methodologies remain limited by their insufficient exploration of temporal artifacts in synthetic videos. To bridge this gap, we establish a theoretical framework through second-order dynamical analysis under Newtonian mechanics, subsequently extending the Second-order Central Difference features tailored for temporal artifact detection. Building on this theoretical foundation, we reveal a fundamental divergence in second-order feature distributions between real and AI-generated videos. Concretely, we propose Detection by Difference of Differences (D3), a novel training-free detection method that leverages the above second-order temporal discrepancies. We validate the superiority of our D3 on 4 open-source datasets (Gen-Video, VideoPhy, EvalCrafter, VidProM), 40 subsets in total. For example, on GenVideo, D3 outperforms the previous best method by 10.39% (absolute) mean Average Precision. Additional experiments on time cost and post-processing operations demonstrate D3's exceptional computational efficiency and strong robust performance. Our code is available at https://github.com/Zig-HS/D3.

cs.CV↗

arXiv2025

Generalizable Targeted Data Poisoning against Varying Physical Objects

Targeted data poisoning (TDP) aims to compromise the model's prediction on a specific (test) target by perturbing a small subset of training data. Existing work on TDP has focused on an overly ideal threat model in which the same image sample of the target is used during both poisoning and inference stages. However, in the real world, a target object often appears in complex variations due to changes of physical settings such as viewpoint, background, and lighting conditions. In this work, we take the first step toward understanding the real-world threats of TDP by studying its generalizability across varying physical conditions. In particular, we observe that solely optimizing gradient directions, as adopted by the best previous TDP method, achieves limited generalization. To address this limitation, we propose optimizing both the gradient direction and magnitude for more generalizable gradient matching, thereby leading to higher poisoning success rates. For instance, our method outperforms the state of the art by 19.49% when poisoning CIFAR-10 images targeting multi-view cars.

cs.CV↗

arXiv2025

Prediction Inconsistency Helps Achieve Generalizable Detection of Adversarial Examples

Adversarial detection protects models from adversarial attacks by refusing suspicious test samples. However, current detection methods often suffer from weak generalization: their effectiveness tends to degrade significantly when applied to adversarially trained models rather than naturally trained ones, and they generally struggle to achieve consistent effectiveness across both white-box and black-box attack settings. In this work, we observe that an auxiliary model, differing from the primary model in training strategy or model architecture, tends to assign low confidence to the primary model's predictions on adversarial examples (AEs), while preserving high confidence on normal examples (NEs). Based on this discovery, we propose Prediction Inconsistency Detector (PID), a lightweight and generalizable detection framework to distinguish AEs from NEs by capturing the prediction inconsistency between the primal and auxiliary models. PID is compatible with both naturally and adversarially trained primal models and outperforms four detection methods across 3 white-box, 3 black-box, and 1 mixed adversarial attacks. Specifically, PID achieves average AUC scores of 99.29\% and 99.30\% on CIFAR-10 when the primal model is naturally and adversarially trained, respectively, and 98.31% and 96.81% on ImageNet under the same conditions, outperforming existing SOTAs by 4.70%$\sim$25.46%.

cs.CR↗

arXiv2025

Adversarially Robust AI-Generated Image Detection for Free: An Information Theoretic Perspective

Rapid advances in Artificial Intelligence Generated Images (AIGI) have facilitated malicious use, such as forgery and misinformation. Therefore, numerous methods have been proposed to detect fake images. Although such detectors have been proven to be universally vulnerable to adversarial attacks, defenses in this field are scarce. In this paper, we first identify that adversarial training (AT), widely regarded as the most effective defense, suffers from performance collapse in AIGI detection. Through an information-theoretic lens, we further attribute the cause of collapse to feature entanglement, which disrupts the preservation of feature-label mutual information. Instead, standard detectors show clear feature separation. Motivated by this difference, we propose Training-free Robust Detection via Information-theoretic Measures (TRIM), the first training-free adversarial defense for AIGI detection. TRIM builds on standard detectors and quantifies feature shifts using prediction entropy and KL divergence. Extensive experiments across multiple datasets and attacks validate the superiority of our TRIM, e.g., outperforming the state-of-the-art defense by 33.88% (28.91%) on ProGAN (GenImage), while well maintaining original accuracy.

cs.CV↗

arXiv2025

Seeing It or Not? Interpretable Vision-aware Latent Steering to Mitigate Object Hallucinations

Large Vision-Language Models (LVLMs) have achieved remarkable success but continue to struggle with object hallucination (OH), generating outputs inconsistent with visual inputs. While previous work has proposed methods to reduce OH, the visual decision-making mechanisms that lead to hallucinations remain poorly understood. In this paper, we propose VaLSe, a Vision-aware Latent Steering framework that adopts an interpretation-then-mitigation strategy to address OH in LVLMs. By tackling dual challenges of modeling complex vision-language interactions and eliminating spurious activation artifacts, VaLSe can generate visual contribution maps that trace how specific visual inputs influence individual output tokens. These maps reveal the model's vision-aware focus regions, which are then used to perform latent space steering, realigning internal representations toward semantically relevant content and reducing hallucinated outputs. Extensive experiments demonstrate that VaLSe is a powerful interpretability tool and an effective method for enhancing model robustness against OH across multiple benchmarks. Furthermore, our analysis uncovers limitations in existing OH evaluation metrics, underscoring the need for more nuanced, interpretable, and visually grounded OH benchmarks in future work. Code is available at: https://github.com/Ziwei-Zheng/VaLSe.

cs.CV↗

arXiv2025

Adversarial Example Soups: Improving Transferability and Stealthiness for Free

Transferable adversarial examples cause practical security risks since they can mislead a target model without knowing its internal knowledge. A conventional recipe for maximizing transferability is to keep only the optimal adversarial example from all those obtained in the optimization pipeline. In this paper, for the first time, we revisit this convention and demonstrate that those discarded, sub-optimal adversarial examples can be reused to boost transferability. Specifically, we propose ``Adversarial Example Soups'' (AES), with AES-tune for averaging discarded adversarial examples in hyperparameter tuning and AES-rand for stability testing. In addition, our AES is inspired by ``model soups'', which averages weights of multiple fine-tuned models for improved accuracy without increasing inference time. Extensive experiments validate the global effectiveness of our AES, boosting 10 state-of-the-art transfer attacks and their combinations by up to 13\% against 10 diverse (defensive) target models. We also show the possibility of generalizing AES to other types, \textit{e.g.}, directly averaging multiple in-the-wild adversarial examples that yield comparable success. A promising byproduct of AES is the improved stealthiness of adversarial examples since the perturbation variances are naturally reduced.

cs.CV↗

arXiv2025

An Optimizable Suffix Is Worth A Thousand Templates: Efficient Black-box Jailbreaking without Affirmative Phrases via LLM as Optimizer

Despite prior safety alignment efforts, mainstream LLMs can still generate harmful and unethical content when subjected to jailbreaking attacks. Existing jailbreaking methods fall into two main categories: template-based and optimization-based methods. The former requires significant manual effort and domain knowledge, while the latter, exemplified by Greedy Coordinate Gradient (GCG), which seeks to maximize the likelihood of harmful LLM outputs through token-level optimization, also encounters several limitations: requiring white-box access, necessitating pre-constructed affirmative phrase, and suffering from low efficiency. In this paper, we present ECLIPSE, a novel and efficient black-box jailbreaking method utilizing optimizable suffixes. Drawing inspiration from LLMs' powerful generation and optimization capabilities, we employ task prompts to translate jailbreaking goals into natural language instructions. This guides the LLM to generate adversarial suffixes for malicious queries. In particular, a harmfulness scorer provides continuous feedback, enabling LLM self-reflection and iterative optimization to autonomously and efficiently produce effective suffixes. Experimental results demonstrate that ECLIPSE achieves an average attack success rate (ASR) of 0.92 across three open-source LLMs and GPT-3.5-Turbo, significantly surpassing GCG in 2.4 times. Moreover, ECLIPSE is on par with template-based methods in ASR while offering superior attack efficiency, reducing the average attack overhead by 83%.

cs.AI↗

arXiv2025

Improving Adversarial Transferability on Vision Transformers via Forward Propagation Refinement

Vision Transformers (ViTs) have been widely applied in various computer vision and vision-language tasks. To gain insights into their robustness in practical scenarios, transferable adversarial examples on ViTs have been extensively studied. A typical approach to improving adversarial transferability is by refining the surrogate model. However, existing work on ViTs has restricted their surrogate refinement to backward propagation. In this work, we instead focus on Forward Propagation Refinement (FPR) and specifically refine two key modules of ViTs: attention maps and token embeddings. For attention maps, we propose Attention Map Diversification (AMD), which diversifies certain attention maps and also implicitly imposes beneficial gradient vanishing during backward propagation. For token embeddings, we propose Momentum Token Embedding (MTE), which accumulates historical token embeddings to stabilize the forward updates in both the Attention and MLP blocks. We conduct extensive experiments with adversarial examples transferred from ViTs to various CNNs and ViTs, demonstrating that our FPR outperforms the current best (backward) surrogate refinement by up to 7.0\% on average. We also validate its superiority against popular defenses and its compatibility with other transfer methods. Codes and appendix are available at https://github.com/RYC-98/FPR.

cs.CV↗

arXiv2025

Nullu: Mitigating Object Hallucinations in Large Vision-Language Models via HalluSpace Projection

Recent studies have shown that large vision-language models (LVLMs) often suffer from the issue of object hallucinations (OH). To mitigate this issue, we introduce an efficient method that edits the model weights based on an unsafe subspace, which we call HalluSpace in this paper. With truthful and hallucinated text prompts accompanying the visual content as inputs, the HalluSpace can be identified by extracting the hallucinated embedding features and removing the truthful representations in LVLMs. By orthogonalizing the model weights, input features will be projected into the Null space of the HalluSpace to reduce OH, based on which we name our method Nullu. We reveal that HalluSpaces generally contain prior information in the large language models (LLMs) applied to build LVLMs, which have been shown as essential causes of OH in previous studies. Therefore, null space projection suppresses the LLMs' priors to filter out the hallucinated features, resulting in contextually accurate outputs. Experiments show that our method can effectively mitigate OH across different LVLM families without extra inference costs and also show strong performance in general LVLM benchmarks. Code is released at https://github.com/Ziwei-Zheng/Nullu.

cs.CV↗

arXiv2025

Systematic Categorization, Construction and Evaluation of New Attacks against Multi-modal Mobile GUI Agents

The integration of Large Language Models (LLMs) and Multi-modal Large Language Models (MLLMs) into mobile GUI agents has significantly enhanced user efficiency and experience. However, this advancement also introduces potential security vulnerabilities that have yet to be thoroughly explored. In this paper, we present a systematic security investigation of multi-modal mobile GUI agents, addressing this critical gap in the existing literature. Our contributions are twofold: (1) we propose a novel threat modeling methodology, leading to the discovery and feasibility analysis of 34 previously unreported attacks, and (2) we design an attack framework to systematically construct and evaluate these threats. Through a combination of real-world case studies and extensive dataset-driven experiments, we validate the severity and practicality of those attacks, highlighting the pressing need for robust security measures in mobile GUI systems.

cs.CR↗

arXiv2025

Revisiting Training-Inference Trigger Intensity in Backdoor Attacks

Backdoor attacks typically place a specific trigger on certain training data, such that the model makes prediction errors on inputs with that trigger during inference. Despite the core role of the trigger, existing studies have commonly believed a perfect match between training-inference triggers is optimal. In this paper, for the first time, we systematically explore the training-inference trigger relation, particularly focusing on their mismatch, based on a Training-Inference Trigger Intensity Manipulation (TITIM) workflow. TITIM specifically investigates the training-inference trigger intensity, such as the size or the opacity of a trigger, and reveals new insights into trigger generalization and overfitting. These new insights challenge the above common belief by demonstrating that the training-inference trigger mismatch can facilitate attacks in two practical scenarios, posing more significant security threats than previously thought. First, when the inference trigger is fixed, using training triggers with mixed intensities leads to stronger attacks than using any single intensity. For example, on CIFAR-10 with ResNet-18, mixing training triggers with 1.0 and 0.1 opacities improves the worst-case attack success rate (ASR) (over different testing opacities) of the best single-opacity attack from 10.61\% to 92.77\%. Second, intentionally using certain mismatched training-inference triggers can improve the attack stealthiness, i.e., better bypassing defenses. For example, compared to the training/inference intensity of 1.0/1.0, using 1.0/0.7 decreases the area under the curve (AUC) of the Scale-Up defense from 0.96 to 0.62, while maintaining a high attack ASR (99.65\% vs. 91.62\%). The above new insights are validated to be generalizable across different backdoor attacks, models, datasets, tasks, and (digital/physical) domains.

cs.CR↗

arXiv2025

CLIP is Strong Enough to Fight Back: Test-time Counterattacks towards Zero-shot Adversarial Robustness of CLIP

Despite its prevalent use in image-text matching tasks in a zero-shot manner, CLIP has been shown to be highly vulnerable to adversarial perturbations added onto images. Recent studies propose to finetune the vision encoder of CLIP with adversarial samples generated on the fly, and show improved robustness against adversarial attacks on a spectrum of downstream datasets, a property termed as zero-shot robustness. In this paper, we show that malicious perturbations that seek to maximise the classification loss lead to `falsely stable' images, and propose to leverage the pre-trained vision encoder of CLIP to counterattack such adversarial images during inference to achieve robustness. Our paradigm is simple and training-free, providing the first method to defend CLIP from adversarial attacks at test time, which is orthogonal to existing methods aiming to boost zero-shot adversarial robustness of CLIP. We conduct experiments across 16 classification datasets, and demonstrate stable and consistent gains compared to test-time defence methods adapted from existing adversarial robustness studies that do not rely on external networks, without noticeably impairing performance on clean images. We also show that our paradigm can be employed on CLIP models that have been adversarially finetuned to further enhance their robustness at test time. Our code is available \href{https://github.com/Sxing2/CLIP-Test-time-Counterattacks}{here}.

cs.CV↗

arXiv2025

Typographic Attacks in a Multi-Image Setting

Large Vision-Language Models (LVLMs) are susceptible to typographic attacks, which are misclassifications caused by an attack text that is added to an image. In this paper, we introduce a multi-image setting for studying typographic attacks, broadening the current emphasis of the literature on attacking individual images. Specifically, our focus is on attacking image sets without repeating the attack query. Such non-repeating attacks are stealthier, as they are more likely to evade a gatekeeper than attacks that repeat the same attack text. We introduce two attack strategies for the multi-image setting, leveraging the difficulty of the target image, the strength of the attack text, and text-image similarity. Our text-image similarity approach improves attack success rates by 21% over random, non-specific methods on the CLIP model using ImageNet while maintaining stealth in a multi-image scenario. An additional experiment demonstrates transferability, i.e., text-image similarity calculated using CLIP transfers when attacking InstructBLIP.

cs.CR↗

arXiv2024

Improving Integrated Gradient-based Transferable Adversarial Examples by Refining the Integration Path

Transferable adversarial examples are known to cause threats in practical, black-box attack scenarios. A notable approach to improving transferability is using integrated gradients (IG), originally developed for model interpretability. In this paper, we find that existing IG-based attacks have limited transferability due to their naive adoption of IG in model interpretability. To address this limitation, we focus on the IG integration path and refine it in three aspects: multiplicity, monotonicity, and diversity, supported by theoretical analyses. We propose the Multiple Monotonic Diversified Integrated Gradients (MuMoDIG) attack, which can generate highly transferable adversarial examples on different CNN and ViT models and defenses. Experiments validate that MuMoDIG outperforms the latest IG-based attack by up to 37.3\% and other state-of-the-art attacks by 8.4\%. In general, our study reveals that migrating established techniques to improve transferability may require non-trivial efforts. Code is available at \url{https://github.com/RYC-98/MuMoDIG}.

cs.CR↗

arXiv2024

Level Up with ML Vulnerability Identification: Leveraging Domain Constraints in Feature Space for Robust Android Malware Detection

Machine Learning (ML) promises to enhance the efficacy of Android Malware Detection (AMD); however, ML models are vulnerable to realistic evasion attacks--crafting realizable Adversarial Examples (AEs) that satisfy Android malware domain constraints. To eliminate ML vulnerabilities, defenders aim to identify susceptible regions in the feature space where ML models are prone to deception. The primary approach to identifying vulnerable regions involves investigating realizable AEs, but generating these feasible apps poses a challenge. For instance, previous work has relied on generating either feature-space norm-bounded AEs or problem-space realizable AEs in adversarial hardening. The former is efficient but lacks full coverage of vulnerable regions while the latter can uncover these regions by satisfying domain constraints but is known to be time-consuming. To address these limitations, we propose an approach to facilitate the identification of vulnerable regions. Specifically, we introduce a new interpretation of Android domain constraints in the feature space, followed by a novel technique that learns them. Our empirical evaluations across various evasion attacks indicate effective detection of AEs using learned domain constraints, with an average of 89.6%. Furthermore, extensive experiments on different Android malware detectors demonstrate that utilizing our learned domain constraints in Adversarial Training (AT) outperforms other AT-based defenses that rely on norm-bounded AEs or state-of-the-art non-uniform perturbations. Finally, we show that retraining a malware detector with a wide variety of feature-space realizable AEs results in a 77.9% robustness improvement against realizable AEs generated by unknown problem-space transformations, with up to 70x faster training than using problem-space realizable AEs.

cs.LG↗

arXiv2024

A Survey on Adversarial Machine Learning for Code Data: Realistic Threats, Countermeasures, and Interpretations

Code Language Models (CLMs) have achieved tremendous progress in source code understanding and generation, leading to a significant increase in research interests focused on applying CLMs to real-world software engineering tasks in recent years. However, in realistic scenarios, CLMs are exposed to potential malicious adversaries, bringing risks to the confidentiality, integrity, and availability of CLM systems. Despite these risks, a comprehensive analysis of the security vulnerabilities of CLMs in the extremely adversarial environment has been lacking. To close this research gap, we categorize existing attack techniques into three types based on the CIA triad: poisoning attacks (integrity \& availability infringement), evasion attacks (integrity infringement), and privacy attacks (confidentiality infringement). We have collected so far the most comprehensive (79) papers related to adversarial machine learning for CLM from the research fields of artificial intelligence, computer security, and software engineering. Our analysis covers each type of risk, examining threat model categorization, attack techniques, and countermeasures, while also introducing novel perspectives on eXplainable AI (XAI) and exploring the interconnections between different risks. Finally, we identify current challenges and future research opportunities. This study aims to provide a comprehensive roadmap for both researchers and practitioners and pave the way towards more reliable CLMs for practical applications.

cs.CR↗

arXiv2024

Improving Adversarial Robustness in Android Malware Detection by Reducing the Impact of Spurious Correlations

Machine learning (ML) has demonstrated significant advancements in Android malware detection (AMD); however, the resilience of ML against realistic evasion attacks remains a major obstacle for AMD. One of the primary factors contributing to this challenge is the scarcity of reliable generalizations. Malware classifiers with limited generalizability tend to overfit spurious correlations derived from biased features. Consequently, adversarial examples (AEs), generated by evasion attacks, can modify these features to evade detection. In this study, we propose a domain adaptation technique to improve the generalizability of AMD by aligning the distribution of malware samples and AEs. Specifically, we utilize meaningful feature dependencies, reflecting domain constraints in the feature space, to establish a robust feature space. Training on the proposed robust feature space enables malware classifiers to learn from predefined patterns associated with app functionality rather than from individual features. This approach helps mitigate spurious correlations inherent in the initial feature space. Our experiments conducted on DREBIN, a renowned Android malware detector, demonstrate that our approach surpasses the state-of-the-art defense, Sec-SVM, when facing realistic evasion attacks. In particular, our defense can improve adversarial robustness by up to 55% against realistic evasion attacks compared to Sec-SVM.

cs.CR↗