arXiv subjects

Hongsheng Hu

Publications and source records attributed to Hongsheng Hu.

At least 19 recordsLinked to original sources

arXiv2026

When T2I Synthetic Data Backfires: Amplified Privacy Risks in Real-Synthetic Mix Training

To overcome data scarcity and privacy constraints in data collection, it has become standard practice across academia and industry to augment real training data with text-to-image (T2I)-generated synthetic data, a paradigm we term Real-Synthetic Mix-Training (RSMT). While substituting synthetic data for sensitive real samples is widely regarded as a means to mitigate privacy exposure of the substituted data, the risk to the remaining real samples that actively participate in training has remained largely unexamined. This work reveals, for the first time, that RSMT can substantially amplify privacy leakage of these real training samples. We establish a theoretical framework, RSMT Memorization Amplification, proving that incorporating synthetic data displaces real samples toward peripheral regions of the mixed feature space, in turn forcing the model to memorize them more aggressively. Guided by this foundation, we propose RSMixLeak to systematically assess this risk through membership inference attacks (MIAs). RSMixLeak comprises two variants depending on the adversary's capability. The non-adversarial variant audits a benign RSMT pipeline with an honest T2I provider, establishing a lower bound on the leakage induced by the intrinsic gap between real and T2I-generated data. The adversarial variant considers an adversary who controls the T2I model or contributes crafted data to the T2I provider, and deliberately enlarges this distributional gap on a target class via either high-level semantic attribute binding or imperceptible pixel-level coating, further amplifying leakage on real training data while improving downstream model utility. Motivated by these findings, we further propose a lightweight leakage propensity indicator computable from real data alone that reliably identifies high-risk datasets unsuitable for entering RSMT, as a self-assessable mitigation.

cs.CR↗

arXiv2026

From Multiplicity to Vulnerability: Privacy Amplification Risk from One-Dataset-Multiple-Model Exposure

To efficiently exploit a valuable data source (e.g., facial or medical images), it is frequently harnessed to fulfill multiple learning objectives (e.g., facial recognition, age estimation, and race classification). Each trained model is then deployed as an independent API service for corresponding inference. However, the privacy risk introduced by this one-dataset-multiple-model (ODMM) paradigm is completely overlooked by the community. For the first time, this work reveals that the ODMM setting substantially amplifies privacy leakage. We establish a theoretical framework that proves that privacy leakage accumulates as more ODMM models are exposed, a phenomenon we term ODMM privacy composition. Guided by this theoretical foundation, we propose PRIME (Privacy Amplification RIsk from One-Dataset-Multiple-Model Exposure) to systematically assess this risk and quantify the resulting leakage using membership inference attacks (MIAs). Under black-box access to ODMM models, we design an aggregation mechanism that collectively captures carefully identified privacy signals leaked by individual ODMM models, and construct an attack meta-classifier over the aggregated meta-information to infer the membership status of a given sample jointly. Our results provide strong evidence that dataset reuse across ODMM models strikingly jeopardizes privacy, which is consistently evident across five privacy-sensitive image and textual benchmark datasets and diverse model architectures (from ResNet and ViT to Qwen3-1.7B), spanning three domains: facial analysis, medical imaging, and textual attribution analysis. While mitigations such as differential privacy can reduce the effectiveness of PRIME with trade-offs, our attack still consistently outperforms single-task MIAs.

cs.CR↗

arXiv2026

Infinite families of constacyclic codes supporting 3-designs and their applications in coding theory

Constacyclic codes over finite fields are of theoretical importance as they are closely related to a number of areas of mathematics such as algebra, algebraic geometry, graph theory, combinatorial designs and number theory. However, the study of constacyclic codes in this context remains limited compared to classical cyclic codes. This paper provides two infinite families of $\lambda$-constacyclic codes over $\mathbb{F}_{q^2}$ that support infinite families of 3-designs, which generalize the results in [IEEE Trans. Inf. Theory 69(4): 2341-2354, 2023]. The parameters and weight distributions are determined completely. Besides, we study their subfield subcodes and applications on constructing entanglement-assisted quantum error-correcting codes (EAQECCs) and locally recoverable codes (LRCs). It is worthy to mention that two classes of maximal entanglement EAQECCs with a negative or a high positive net rate are derived. Moreover, two classes of distance-optimal and dimension-optimal LRCs are also obtained.

cs.IT↗

arXiv2026

SoK: Robustness in Large Language Models against Jailbreak Attacks

Large Language Models (LLMs) have achieved remarkable success but remain highly susceptible to jailbreak attacks, in which adversarial prompts coerce models into generating harmful, unethical, or policy-violating outputs. Such attacks pose real-world risks, eroding safety, trust, and regulatory compliance in high-stakes applications. Although a variety of attack and defense methods have been proposed, existing evaluation practices are inadequate, often relying on narrow metrics like attack success rate that fail to capture the multidimensional nature of LLM security. In this paper, we present a systematic taxonomy of jailbreak attacks and defenses and introduce Security Cube, a unified, multi-dimensional framework for comprehensive evaluation of these techniques. We provide detailed comparison tables of existing attacks and defenses, highlighting key insights and open challenges across the literature. Leveraging Security Cube, we conduct benchmark studies on 13 representative attacks and 5 defenses, establishing a clear view of the current landscape encompassing jailbreak attacks, defenses, automated judges, and LLM vulnerabilities. Based on these evaluations, we distill critical findings, identify unresolved problems, and outline promising research directions for enhancing LLM robustness against jailbreak attacks. Our analysis aims to pave the way towards more robust, interpretable, and trustworthy LLM systems. Our code is available at Code.

cs.CR↗

arXiv2025

Reference Recommendation based Membership Inference Attack against Hybrid-based Recommender Systems

Recommender systems have been widely deployed across various domains such as e-commerce and social media, and intelligently suggest items like products and potential friends to users based on their preferences and interaction history, which are often privacy-sensitive. Recent studies have revealed that recommender systems are prone to membership inference attacks (MIAs), where an attacker aims to infer whether or not a user's data has been used for training a target recommender system. However, existing MIAs fail to exploit the unique characteristic of recommender systems, and therefore are only applicable to mixed recommender systems consisting of two recommendation algorithms. This leaves a gap in investigating MIAs against hybrid-based recommender systems where the same algorithm utilizing user-item historical interactions and attributes of users and items serves and produces personalised recommendations. To investigate how the personalisation in hybrid-based recommender systems influences MIA, we propose a novel metric-based MIA. Specifically, we leverage the characteristic of personalisation to obtain reference recommendation for any target users. Then, a relative membership metric is proposed to exploit a target user's historical interactions, target recommendation, and reference recommendation to infer the membership of the target user's data. Finally, we theoretically and empirically demonstrate the efficacy of the proposed metric-based MIA on hybrid-based recommender systems.

cs.CR↗

arXiv2025

Malicious Image Analysis via Vision-Language Segmentation Fusion: Detection, Element, and Location in One-shot

Detecting illicit visual content demands more than image-level NSFW flags; moderators must also know what objects make an image illegal and where those objects occur. We introduce a zero-shot pipeline that simultaneously (i) detects if an image contains harmful content, (ii) identifies each critical element involved, and (iii) localizes those elements with pixel-accurate masks - all in one pass. The system first applies foundation segmentation model (SAM) to generate candidate object masks and refines them into larger independent regions. Each region is scored for malicious relevance by a vision-language model using open-vocabulary prompts; these scores weight a fusion step that produces a consolidated malicious object map. An ensemble across multiple segmenters hardens the pipeline against adaptive attacks that target any single segmentation method. Evaluated on a newly-annotated 790-image dataset spanning drug, sexual, violent and extremist content, our method attains 85.8% element-level recall, 78.1% precision and a 92.1% segment-success rate - exceeding direct zero-shot VLM localization by 27.4% recall at comparable precision. Against PGD adversarial perturbations crafted to break SAM and VLM, our method's precision and recall decreased by no more than 10%, demonstrating high robustness against attacks. The full pipeline processes an image in seconds, plugs seamlessly into existing VLM workflows, and constitutes the first practical tool for fine-grained, explainable malicious-image moderation.

cs.CV↗

arXiv2025

NonTextual Target Attack

Existing gradient-based jailbreak attacks on Large Language Models (LLMs) typically optimize adversarial suffixes to align the LLM output with predefined target responses. However, restricting the objective as inducing fixed targets inherently constrains the adversarial search space, limiting the overall attack efficacy. Furthermore, existing methods typically require numerous optimization iterations to fulfill the large gap between the fixed target and the original LLM output, resulting in low attack efficiency. To overcome these limitations, we propose NonTextual Target Attack (NTA), the first gradient-based attack that relies on a non-textual constrained objective to maximize the unsafety probability of the LLM output, without enforcing any response patterns. For tractable optimization, we further decompose this objective into two constrained sub-objectives, which can be approximated by two differentiable unconstrained losses, to iteratively optimize the response and the adversarial prompt in the neighborhood of the original prompt, with a theoretical analysis to validate the decomposition. In contrast to existing attacks, NTA first realizes gradient-based prompt optimization on a non-textual target and significantly expands the attack space, enabling more flexible and efficient exploration of LLM vulnerabilities. Extensive evaluations show that \textsc{NTA} achieves an average attack success rate of 96.8\% against recent safety-aligned LLMs with only 100 optimization iterations on AdvBench, outperforming state-of-the-art gradient-based attacks by over 40\%.

cs.CR↗

arXiv2025

BadFU: Backdoor Federated Learning through Adversarial Machine Unlearning

Federated learning (FL) has been widely adopted as a decentralized training paradigm that enables multiple clients to collaboratively learn a shared model without exposing their local data. As concerns over data privacy and regulatory compliance grow, machine unlearning, which aims to remove the influence of specific data from trained models, has become increasingly important in the federated setting to meet legal, ethical, or user-driven demands. However, integrating unlearning into FL introduces new challenges and raises largely unexplored security risks. In particular, adversaries may exploit the unlearning process to compromise the integrity of the global model. In this paper, we present the first backdoor attack in the context of federated unlearning, demonstrating that an adversary can inject backdoors into the global model through seemingly legitimate unlearning requests. Specifically, we propose BadFU, an attack strategy where a malicious client uses both backdoor and camouflage samples to train the global model normally during the federated training process. Once the client requests unlearning of the camouflage samples, the global model transitions into a backdoored state. Extensive experiments under various FL frameworks and unlearning strategies validate the effectiveness of BadFU, revealing a critical vulnerability in current federated unlearning practices and underscoring the urgent need for more secure and robust federated unlearning mechanisms.

cs.CR↗

arXiv2025

CompLeak: Deep Learning Model Compression Exacerbates Privacy Leakage

Model compression is crucial for minimizing memory storage and accelerating inference in deep learning (DL) models, including recent foundation models like large language models (LLMs). Users can access different compressed model versions according to their resources and budget. However, while existing compression operations primarily focus on optimizing the trade-off between resource efficiency and model performance, the privacy risks introduced by compression remain overlooked and insufficiently understood. In this work, through the lens of membership inference attack (MIA), we propose CompLeak, the first privacy risk evaluation framework examining three widely used compression configurations that are pruning, quantization, and weight clustering supported by the commercial model compression framework of Google's TensorFlow-Lite (TF-Lite) and Facebook's PyTorch Mobile. CompLeak has three variants, given available access to the number of compressed models and original model. CompLeakNR starts by adopting existing MIA methods to attack a single compressed model, and identifies that different compressed models influence members and non-members differently. When the original model and one compressed model are available, CompLeakSR leverages the compressed model as a reference to the original model and uncovers more privacy by combining meta information (e.g., confidence vector) from both models. When multiple compressed models are available with/without accessing the original model, CompLeakMR innovatively exploits privacy leakage info from multiple compressed versions to substantially signify the overall privacy leakage. We conduct extensive experiments on seven diverse model architectures (from ResNet to foundation models of BERT and GPT-2), and six image and textual benchmark datasets.

cs.CR↗

arXiv2025

When Better Features Mean Greater Risks: The Performance-Privacy Trade-Off in Contrastive Learning

With the rapid advancement of deep learning technology, pre-trained encoder models have demonstrated exceptional feature extraction capabilities, playing a pivotal role in the research and application of deep learning. However, their widespread use has raised significant concerns about the risk of training data privacy leakage. This paper systematically investigates the privacy threats posed by membership inference attacks (MIAs) targeting encoder models, focusing on contrastive learning frameworks. Through experimental analysis, we reveal the significant impact of model architecture complexity on membership privacy leakage: As more advanced encoder frameworks improve feature-extraction performance, they simultaneously exacerbate privacy-leakage risks. Furthermore, this paper proposes a novel membership inference attack method based on the p-norm of feature vectors, termed the Embedding Lp-Norm Likelihood Attack (LpLA). This method infers membership status, by leveraging the statistical distribution characteristics of the p-norm of feature vectors. Experimental results across multiple datasets and model architectures demonstrate that LpLA outperforms existing methods in attack performance and robustness, particularly under limited attack knowledge and query volumes. This study not only uncovers the potential risks of privacy leakage in contrastive learning frameworks, but also provides a practical basis for privacy protection research in encoder models. We hope that this work will draw greater attention to the privacy risks associated with self-supervised learning models and shed light on the importance of a balance between model utility and training data privacy. Our code is publicly available at: https://github.com/SeroneySun/LpLA_code.

cs.CR↗

arXiv2025

From Pixels to Trajectory: Universal Adversarial Example Detection via Temporal Imprints

For the first time, we unveil discernible temporal (or historical) trajectory imprints resulting from adversarial example (AE) attacks. Standing in contrast to existing studies all focusing on spatial (or static) imprints within the targeted underlying victim models, we present a fresh temporal paradigm for understanding these attacks. Of paramount discovery is that these imprints are encapsulated within a single loss metric, spanning universally across diverse tasks such as classification and regression, and modalities including image, text, and audio. Recognizing the distinct nature of loss between adversarial and clean examples, we exploit this temporal imprint for AE detection by proposing TRAIT (TRaceable Adversarial temporal trajectory ImprinTs). TRAIT operates under minimal assumptions without prior knowledge of attacks, thereby framing the detection challenge as a one-class classification problem. However, detecting AEs is still challenged by significant overlaps between the constructed synthetic losses of adversarial and clean examples due to the absence of ground truth for incoming inputs. TRAIT addresses this challenge by converting the synthetic loss into a spectrum signature, using the technique of Fast Fourier Transform to highlight the discrepancies, drawing inspiration from the temporal nature of the imprints, analogous to time-series signals. Across 12 AE attacks including SMACK (USENIX Sec'2023), TRAIT demonstrates consistent outstanding performance across comprehensively evaluated modalities, tasks, datasets, and model architectures. In all scenarios, TRAIT achieves an AE detection accuracy exceeding 97%, often around 99%, while maintaining a false rejection rate of 1%. TRAIT remains effective under the formulated strong adaptive attacks.

cs.CR↗

arXiv2024

ExpShield: Safeguarding Web Text from Unauthorized Crawling and LLM Exploitation

As large language models increasingly memorize web-scraped training content, they risk exposing copyrighted or private information. Existing protections require compliance from crawlers or model developers, fundamentally limiting their effectiveness. We propose ExpShield, a proactive self-guard that mitigates memorization while maintaining readability via invisible perturbations, and we formulate it as a constrained optimization problem. Due to the lack of an individual-level risk metric for natural text, we first propose instance exploitation, a metric that measures how much training on a specific text increases the chance of guessing that text from a set of candidates-with zero indicating perfect defense. Directly solving the problem is infeasible for defenders without sufficient knowledge, thus we develop two effective proxy solutions: single-level optimization and synthetic perturbation. To enhance the defense, we reveal and verify the memorization trigger hypothesis, which can help to identify key tokens for memorization. Leveraging this insight, we design targeted perturbations that (i) neutralize inherent trigger tokens to reduce memorization and (ii) introduce artificial trigger tokens to misdirect model memorization. Experiments validate our defense across attacks, model scales, and tasks in language and vision-to-language modeling. Even with privacy backdoor, the Membership Inference Attack (MIA) AUC drops from 0.95 to 0.55 under the defense, and the instance exploitation approaches zero. This suggests that compared to the ideal no-misuse scenario, the risk of exposing a text instance remains nearly unchanged despite its inclusion in the training data.

cs.CR↗

arXiv2024

DapperFL: Domain Adaptive Federated Learning with Model Fusion Pruning for Edge Devices

Federated learning (FL) has emerged as a prominent machine learning paradigm in edge computing environments, enabling edge devices to collaboratively optimize a global model without sharing their private data. However, existing FL frameworks suffer from efficacy deterioration due to the system heterogeneity inherent in edge computing, especially in the presence of domain shifts across local data. In this paper, we propose a heterogeneous FL framework DapperFL, to enhance model performance across multiple domains. In DapperFL, we introduce a dedicated Model Fusion Pruning (MFP) module to produce personalized compact local models for clients to address the system heterogeneity challenges. The MFP module prunes local models with fused knowledge obtained from both local and remaining domains, ensuring robustness to domain shifts. Additionally, we design a Domain Adaptive Regularization (DAR) module to further improve the overall performance of DapperFL. The DAR module employs regularization generated by the pruned model, aiming to learn robust representations across domains. Furthermore, we introduce a specific aggregation algorithm for aggregating heterogeneous local models with tailored architectures and weights. We implement DapperFL on a realworld FL platform with heterogeneous clients. Experimental results on benchmark datasets with multiple domains demonstrate that DapperFL outperforms several state-of-the-art FL frameworks by up to 2.28%, while significantly achieving model volume reductions ranging from 20% to 80%. Our code is available at: https://github.com/jyzgh/DapperFL.

cs.LG↗

arXiv2024

AI-Compass: A Comprehensive and Effective Multi-module Testing Tool for AI Systems

AI systems, in particular with deep learning techniques, have demonstrated superior performance for various real-world applications. Given the need for tailored optimization in specific scenarios, as well as the concerns related to the exploits of subsurface vulnerabilities, a more comprehensive and in-depth testing AI system becomes a pivotal topic. We have seen the emergence of testing tools in real-world applications that aim to expand testing capabilities. However, they often concentrate on ad-hoc tasks, rendering them unsuitable for simultaneously testing multiple aspects or components. Furthermore, trustworthiness issues arising from adversarial attacks and the challenge of interpreting deep learning models pose new challenges for developing more comprehensive and in-depth AI system testing tools. In this study, we design and implement a testing tool, \tool, to comprehensively and effectively evaluate AI systems. The tool extensively assesses multiple measurements towards adversarial robustness, model interpretability, and performs neuron analysis. The feasibility of the proposed testing tool is thoroughly validated across various modalities, including image classification, object detection, and text classification. Extensive experiments demonstrate that \tool is the state-of-the-art tool for a comprehensive assessment of the robustness and trustworthiness of AI systems. Our research sheds light on a general solution for AI systems testing landscape.

cs.AI↗

arXiv2024

Intellectual Property Protection for Deep Learning Model and Dataset Intelligence

With the growing applications of Deep Learning (DL), especially recent spectacular achievements of Large Language Models (LLMs) such as ChatGPT and LLaMA, the commercial significance of these remarkable models has soared. However, acquiring well-trained models is costly and resource-intensive. It requires a considerable high-quality dataset, substantial investment in dedicated architecture design, expensive computational resources, and efforts to develop technical expertise. Consequently, safeguarding the Intellectual Property (IP) of well-trained models is attracting increasing attention. In contrast to existing surveys overwhelmingly focusing on model IPP mainly, this survey not only encompasses the protection on model level intelligence but also valuable dataset intelligence. Firstly, according to the requirements for effective IPP design, this work systematically summarizes the general and scheme-specific performance evaluation metrics. Secondly, from proactive IP infringement prevention and reactive IP ownership verification perspectives, it comprehensively investigates and analyzes the existing IPP methods for both dataset and model intelligence. Additionally, from the standpoint of training settings, it delves into the unique challenges that distributed settings pose to IPP compared to centralized settings. Furthermore, this work examines various attacks faced by deep IPP techniques. Finally, we outline prospects for promising future directions that may act as a guide for innovative research.

cs.CR↗

arXiv2024

Weyl group symmetries of the toric variety associated with Weyl chambers

For any crystallographic root system, let $W$ be the associated Weyl group, and let $\mathit{WP}$ be the weight polytope (also known as the $W$-permutohedron) associated with an arbitrary strongly dominant weight. The action of $W$ on $\mathit{WP}$ induces an action on the toric variety $X(\mathit{WP})$ associated with the normal fan of $\mathit{WP}$, and hence an action on the rational cohomology ring $H^*\left(X(\mathit{WP})\right)$. Let $P$ be the corresponding dominant weight polytope, which is a fundamental region of the $W$-action on $\mathit{WP}$. We give a type uniform algebraic proof that the fixed subring $H^*\left(X(\mathit{WP})\right)^{W}$ is isomorphic to the cohomology ring $H^*\left(X(P)\right)$ of the toric variety $X(P)$ associated with the normal fan of $P$. Notably, our proof applies to all finite (not necessarily crystallographic) Coxeter groups, answering a question of Horiguchi--Masuda--Shareshian--Song about non-crystallographic root systems.

math.AT↗

arXiv2024

Releasing Malevolence from Benevolence: The Menace of Benign Data on Machine Unlearning

Machine learning models trained on vast amounts of real or synthetic data often achieve outstanding predictive performance across various domains. However, this utility comes with increasing concerns about privacy, as the training data may include sensitive information. To address these concerns, machine unlearning has been proposed to erase specific data samples from models. While some unlearning techniques efficiently remove data at low costs, recent research highlights vulnerabilities where malicious users could request unlearning on manipulated data to compromise the model. Despite these attacks' effectiveness, perturbed data differs from original training data, failing hash verification. Existing attacks on machine unlearning also suffer from practical limitations and require substantial additional knowledge and resources. To fill the gaps in current unlearning attacks, we introduce the Unlearning Usability Attack. This model-agnostic, unlearning-agnostic, and budget-friendly attack distills data distribution information into a small set of benign data. These data are identified as benign by automatic poisoning detection tools due to their positive impact on model training. While benign for machine learning, unlearning these data significantly degrades model information. Our evaluation demonstrates that unlearning this benign data, comprising no more than 1% of the total training data, can reduce model accuracy by up to 50%. Furthermore, our findings show that well-prepared benign data poses challenges for recent unlearning techniques, as erasing these synthetic instances demands higher resources than regular data. These insights underscore the need for future research to reconsider "data poisoning" in the context of machine unlearning.

cs.CR↗

arXiv2024

Shadow-Free Membership Inference Attacks: Recommender Systems Are More Vulnerable Than You Thought

Recommender systems have been successfully applied in many applications. Nonetheless, recent studies demonstrate that recommender systems are vulnerable to membership inference attacks (MIAs), leading to the leakage of users' membership privacy. However, existing MIAs relying on shadow training suffer a large performance drop when the attacker lacks knowledge of the training data distribution and the model architecture of the target recommender system. To better understand the privacy risks of recommender systems, we propose shadow-free MIAs that directly leverage a user's recommendations for membership inference. Without shadow training, the proposed attack can conduct MIAs efficiently and effectively under a practice scenario where the attacker is given only black-box access to the target recommender system. The proposed attack leverages an intuition that the recommender system personalizes a user's recommendations if his historical interactions are used by it. Thus, an attacker can infer membership privacy by determining whether the recommendations are more similar to the interactions or the general popular items. We conduct extensive experiments on benchmark datasets across various recommender systems. Remarkably, our attack achieves far better attack accuracy with low false positive rates than baselines while with a much lower computational cost.

cs.CR↗