Yanjun Zhang

Xiaomei Zhang, Zhaoxi Zhang, Qi Zhong et al.

Adversarial attacks are a serious threat to the reliable deployment of machine learning models in safety-critical applications. They can misguide current models to predict incorrectly by slightly modifying the inputs. Recently, substantial work has shown that adversarial examples tend to deviate from the underlying data manifold of normal examples, whereas pre-trained masked language models can fit the manifold of normal NLP data. To explore how to use the masked language model in adversarial detection, we propose a novel textual adversarial example detection method, namely Masked Language Model-based Detection (MLMD), which can produce clearly distinguishable signals between normal examples and adversarial examples by exploring the changes in manifolds induced by the masked language model. MLMD features a plug and play usage (i.e., no need to retrain the victim model) for adversarial defense and it is agnostic to classification tasks, victim model's architectures, and to-be-defended attack methods. We evaluate MLMD on various benchmark textual datasets, widely studied machine learning models, and state-of-the-art (SOTA) adversarial attacks (in total $3*4*4 = 48$ settings). Experimental results show that MLMD can achieve strong performance, with detection accuracy up to 0.984, 0.967, and 0.901 on AG-NEWS, IMDB, and SST-2 datasets, respectively. Additionally, MLMD is superior, or at least comparable to, the SOTA detection defenses in detection accuracy and F1 score. Among many defenses based on the off-manifold assumption of adversarial examples, this work offers a new angle for capturing the manifold change. The code for this work is openly accessible at \url{https://github.com/mlmddetection/MLMDdetection}.

Jiaheng Wei, Yanjun Zhang, Leo Yu Zhang et al. · tencent-ai

Federated Learning (FL) enables distributed participants (e.g., mobile devices) to train a global model without sharing data directly to a central server. Recent studies have revealed that FL is vulnerable to gradient inversion attack (GIA), which aims to reconstruct the original training samples and poses high risk against the privacy of clients in FL. However, most existing GIAs necessitate control over the server and rely on strong prior knowledge including batch normalization and data distribution information. In this work, we propose Client-side poisoning Gradient Inversion (CGI), which is a novel attack method that can be launched from clients. For the first time, we show the feasibility of a client-side adversary with limited knowledge being able to recover the training samples from the aggregated global model. We take a distinct approach in which the adversary utilizes a malicious model that amplifies the loss of a specific targeted class of interest. When honest clients employ the poisoned global model, the gradients of samples belonging to the targeted class are magnified, making them the dominant factor in the aggregated update. This enables the adversary to effectively reconstruct the private input belonging to other clients using the aggregated update. In addition, our CGI also features its ability to remain stealthy against Byzantine-robust aggregation rules (AGRs). By optimizing malicious updates and blending benign updates with a malicious replacement vector, our method remains undetected by these defense mechanisms. To evaluate the performance of CGI, we conduct experiments on various benchmark datasets, considering representative Byzantine-robust AGRs, and exploring diverse FL settings with different levels of adversary knowledge about the data. Our results demonstrate that CGI consistently and successfully extracts training input in all tested scenarios.

Yubin Qu, Yi Liu, Gelei Deng et al.

A coding agent executes a benign task as a sequence of shell, file, and network actions, any of which can quietly exceed the authorized scope while the task still completes. We call this overeager behavior: the prompt is not adversarial and the run succeeds, yet an out-of-scope step can leak credentials or delete files. Existing benchmarks miss it: task-completion suites credit any finished run, jailbreak suites probe adversarial prompts, and the one prior overeager benchmark applies a single fixed prompt set to every agent-model pair, leaving its easiest and most resistant pairs under-measured. We present SNARE (Synthesizing Non-adversarial scenarios for Adaptive Reward-guided Elicitation), a pipeline that composes benign scenarios from reusable scope and trap fragments, scores each run with a judge-free oracle flagging trap-pattern matches and unsolicited file additions or deletions, and uses Thompson sampling to steer each pair's run budget toward the scenarios that most often trigger it. Instantiating it over 24 overeager archetypes yields OverEager, which we run across a 4x5 matrix of four coding agents and five base models. Across 10,000 benign runs, 19.51% trigger overeager behavior, with per-pair rates spanning 11.9x. This variation is driven by the agent framework, not the model: the framework accounts for 56% of it against the model's 21%, so any single-framework or single-model evaluation undercounts the matrix by about a fifth.

Zirui Gong, Liyue Shen, Yanjun Zhang et al.

The collaborative nature of federated learning (FL) poses a major threat in the form of manipulation of local training data and local updates, known as the Byzantine poisoning attack. To address this issue, many Byzantine-robust aggregation rules (AGRs) have been proposed to filter out or moderate suspicious local updates uploaded by Byzantine participants. This paper introduces a novel approach called AGRAMPLIFIER, aiming to simultaneously improve the robustness, fidelity, and efficiency of the existing AGRs. The core idea of AGRAMPLIFIER is to amplify the "morality" of local updates by identifying the most repressive features of each gradient update, which provides a clearer distinction between malicious and benign updates, consequently improving the detection effect. To achieve this objective, two approaches, namely AGRMP and AGRXAI, are proposed. AGRMP organizes local updates into patches and extracts the largest value from each patch, while AGRXAI leverages explainable AI methods to extract the gradient of the most activated features. By equipping AGRAMPLIFIER with the existing Byzantine-robust mechanisms, we successfully enhance the model's robustness, maintaining its fidelity and improving overall efficiency. AGRAMPLIFIER is universally compatible with the existing Byzantine-robust mechanisms. The paper demonstrates its effectiveness by integrating it with all mainstream AGR mechanisms. Extensive evaluations conducted on seven datasets from diverse domains against seven representative poisoning attacks consistently show enhancements in robustness, fidelity, and efficiency, with average gains of 40.08%, 39.18%, and 10.68%, respectively.

Yi Liu, Zhihao Chen, Yanjun Zhang et al.

Third-party agent skills extend LLM-based agents with instruction files and executable code that run on users' machines. Skills execute with user privileges and are distributed through community registries with minimal vetting, but no ground-truth dataset exists to characterize the resulting threats. We construct the first labeled dataset of malicious agent skills by behaviorally verifying 98,380 skills from two community registries, confirming 157 malicious skills with 632 vulnerabilities. These attacks are not incidental. Malicious skills average 4.03 vulnerabilities across a median of three kill chain phases, and the ecosystem has split into two archetypes: Data Thieves that exfiltrate credentials through supply chain techniques, and Agent Hijackers that subvert agent decision-making through instruction manipulation. A single actor accounts for 54.1\% of confirmed cases through templated brand impersonation. Shadow features, capabilities absent from public documentation, appear in 0\% of basic attacks but 100\% of advanced ones; several skills go further by exploiting the AI platform's own hook system and permission flags. Responsible disclosure led to 93.6\% removal within 30 days. We release the dataset and analysis pipeline to support future work on agent skill security.

Jiaheng Wei, Yanjun Zhang, He Zhang et al.

Federated learning (FL) increasingly needs machine unlearning to comply with privacy regulations. However, existing federated unlearning approaches may overlook the overlapping information between the unlearning and remaining data, leading to ineffective unlearning and unfairness between clients. In this work, we revisit federated unlearning through the lens of memorization. We argue that unlearning should mainly remove the unique memorized information attributable to the data to be forgotten, while preserving overlapping patterns that are also supported by the remaining data. Specifically, we propose Grouped Memorization Evaluation, an example-level metric that separates memorized knowledge from overlapping knowledge. Building on this metric, we introduce Federated Memorization Pruning (FedMemPrune), a pruning-based unlearning approach that resets redundant parameters responsible for memorization. Extensive experiments show that FedMemPrune closely matches retraining-based unlearning baselines while more effectively eliminating memorization than existing federated unlearning algorithms, yielding strong unlearning performance without sacrificing the utility of retained knowledge.

Zhihao Chen, Zirui Gong, Jianting Ning et al.

Federated Rank Learning (FRL) is a promising Federated Learning (FL) paradigm designed to be resilient against model poisoning attacks due to its discrete, ranking-based update mechanism. Unlike traditional FL methods that rely on model updates, FRL leverages discrete rankings as a communication parameter between clients and the server. This approach significantly reduces communication costs and limits an adversary's ability to scale or optimize malicious updates in the continuous space, thereby enhancing its robustness. This makes FRL particularly appealing for applications where system security and data privacy are crucial, such as web-based auction and bidding platforms. While FRL substantially reduces the attack surface, we demonstrate that it remains vulnerable to a new class of local model poisoning attack, i.e., fine-grained control attacks. We introduce the Edge Control Attack (ECA), the first fine-grained control attack tailored to ranking-based FL frameworks. Unlike conventional denial-of-service (DoS) attacks that cause conspicuous disruptions, ECA enables an adversary to precisely degrade a competitor's accuracy to any target level while maintaining a normal-looking convergence trajectory, thereby avoiding detection. ECA operates in two stages: (i) identifying and manipulating Ascending and Descending Edges to align the global model with the target model, and (ii) widening the selection boundary gap to stabilize the global model at the target accuracy. Extensive experiments across seven benchmark datasets and nine Byzantine-robust aggregation rules (AGRs) show that ECA achieves fine-grained accuracy control with an average error of only 0.224%, outperforming the baseline by up to 17x. Our findings highlight the need for stronger defenses against advanced poisoning attacks. Our code is available at: https://github.com/Chenzh0205/ECA

Di Mi, Yanjun Zhang, Leo Yu Zhang et al.

Model extraction attacks (MEAs) enable an attacker to replicate the functionality of a victim deep neural network (DNN) model by only querying its API service remotely, posing a severe threat to the security and integrity of pay-per-query DNN-based services. Although the majority of current research on MEAs has primarily concentrated on neural classifiers, there is a growing prevalence of image-to-image translation (I2IT) tasks in our everyday activities. However, techniques developed for MEA of DNN classifiers cannot be directly transferred to the case of I2IT, rendering the vulnerability of I2IT models to MEA attacks often underestimated. This paper unveils the threat of MEA in I2IT tasks from a new perspective. Diverging from the traditional approach of bridging the distribution gap between attacker queries and victim training samples, we opt to mitigate the effect caused by the different distributions, known as the domain shift. This is achieved by introducing a new regularization term that penalizes high-frequency noise, and seeking a flatter minimum to avoid overfitting to the shifted distribution. Extensive experiments on different image translation tasks, including image super-resolution and style transfer, are performed on different backbone victim models, and the new design consistently outperforms the baseline by a large margin across all metrics. A few real-life I2IT APIs are also verified to be extremely vulnerable to our attack, emphasizing the need for enhanced defenses and potentially revised API publishing policies.

Yubin Qu, Ying Zhang, Yanjun Zhang et al.

Coding agents now run autonomously with shell, file, and network privileges. When a user issues a benign request, the agent sometimes does more than asked: it deletes unrelated files, wipes a stale credentials backup, or rewrites configuration the user never mentioned. We call these scope expansions overeager actions, an authorization problem distinct from capability failures, prompt injection, or sandbox escapes. We present OverEager-Gen, a benchmark dedicated to overeager behavior on benign tasks. Building it surfaces a measurement-validity issue: if a benchmark spells out the authorized scope inside the prompt, the agent stops inferring boundaries and starts pattern-matching declaration text. On Claude Code, stripping the consent declaration alone raises the overeager rate from 0.0% to 17.1% on paired scenarios (McNemar exact p = 2.4 x 10^-4). OverEager-Gen therefore certifies each scenario's discriminative power before admission via a behavioral-gradient validator, audits internal tool calls through a dual-channel stack (PATH-injected shim plus per-agent event streams), and ships byte-identical consent_kept and consent_stripped variants. OverEager-Bench contains 500 validated scenarios and ~7,500 runs across four agent products (Claude Code, OpenHands, Codex CLI, Gemini CLI) and six base models; a 50-sample re-annotation gives Cohen's kappa = 0.73 and rule-judge recall = 1.00. Stripping consent multiplies the overeager rate on every shared base model (Delta in [11.9, 17.2] pp). The framework axis dominates effect size: a permissive cluster (Claude Code, Codex CLI, Gemini CLI) runs at 5.4-27.7% while the ask-to-continue framework (OpenHands) sits at 0.2-4.5% (Fisher p <= 10^-5). Within-framework base-model variance reaches 15.9 pp, indicating that model-layer alignment does not fully propagate through permissive permission gating.

Lulu Xue, Shengshan Hu, Wei Lu et al.

Machine unlearning is an emerging technique that aims to remove the influence of specific data from trained models, thereby enhancing privacy protection. However, recent research has uncovered critical privacy vulnerabilities, showing that adversaries can exploit unlearning inversion to reconstruct data that was intended to be erased. Despite the severity of this threat, dedicated defenses remain lacking. To address this gap, we propose UnlearnShield, the first defense specifically tailored to counter unlearning inversion. UnlearnShield introduces directional perturbations in the cosine representation space and regulates them through a constraint module to jointly preserve model accuracy and forgetting efficacy, thereby reducing inversion risk while maintaining utility. Experiments demonstrate that it achieves a good trade-off among privacy protection, accuracy, and forgetting.

Lulu Xue, Shengshan Hu, Linqiang Qian et al.

Machine unlearning is a newly popularized technique for removing specific training data from a trained model, enabling it to comply with data deletion requests. While it protects the rights of users requesting unlearning, it also introduces new privacy risks. Prior works have primarily focused on the privacy of data that has been unlearned, while the risks to retained data remain largely unexplored. To address this gap, we focus on the privacy risks of retained data and, for the first time, reveal the vulnerabilities introduced by machine unlearning under the dual-view setting, where an adversary can query both the original and the unlearned models. From an information-theoretic perspective, we introduce the concept of {privacy knowledge gain} and demonstrate that the dual-view setting allows adversaries to obtain more information than querying either model alone, thereby amplifying privacy leakage. To effectively demonstrate this threat, we propose DVIA, a Dual-View Inference Attack, which extracts membership information on retained data using black-box queries to both models. DVIA eliminates the need to train an attack model and employs a lightweight likelihood ratio inference module for efficient inference. Experiments across different datasets and model architectures validate the effectiveness of DVIA and highlight the privacy risks inherent in the dual-view setting.

Zirui Gong, Leo Yu Zhang, Yanjun Zhang et al.

Federated Learning (FL) enables collaborative model training by sharing model updates instead of raw data, aiming to protect user privacy. However, recent studies reveal that these shared updates can inadvertently leak sensitive training data through gradient inversion attacks (GIAs). Among them, active GIAs are particularly powerful, enabling high-fidelity reconstruction of individual samples even under large batch sizes. Nevertheless, existing approaches often require architectural modifications, which limit their practical applicability. In this work, we bridge this gap by introducing the Activation REcovery via Sparse inversion (ARES) attack, an active GIA designed to reconstruct training samples from large training batches without requiring architectural modifications. Specifically, we formulate the recovery problem as a noisy sparse recovery task and solve it using the generalized Least Absolute Shrinkage and Selection Operator (Lasso). To extend the attack to multi-sample recovery, ARES incorporates the imprint method to disentangle activations, enabling scalable per-sample reconstruction. We further establish the expected recovery rate and derive an upper bound on the reconstruction error, providing theoretical guarantees for the ARES attack. Extensive experiments on CNNs and MLPs demonstrate that ARES achieves high-fidelity reconstruction across diverse datasets, significantly outperforming prior GIAs under large batch sizes and realistic FL settings. Our results highlight that intermediate activations pose a serious and underestimated privacy risk in FL, underscoring the urgent need for stronger defenses.

Zhihao Chen, Ying Zhang, Yi Liu et al.

Third-party skills extend LLM agents with powerful capabilities but often handle sensitive credentials in privileged environments, making leakage risks poorly understood. We present the first large-scale empirical study of this problem, analyzing 17,022 skills (sampled from 170,226 on SkillsMP) using static analysis, sandbox testing, and manual inspection. We identify 520 vulnerable skills with 1,708 issues and derive a taxonomy of 10 leakage patterns (4 accidental and 6 adversarial). We find that (1) leakage is fundamentally cross-modal: 76.3% require joint analysis of code and natural language, while 3.1% arise purely from prompt injection; (2) debug logging is the primary vector, with print and console.log causing 73.5% of leaks due to stdout exposure to LLMs; and (3) leaked credentials are both exploitable (89.6% without privileges) and persistent, as forks retain secrets even after upstream fixes. After disclosure, all malicious skills were removed and 91.6% of hardcoded credentials were fixed. We release our dataset, taxonomy, and detection pipeline to support future research.

Hangtao Zhang, Yichen Wang, Shihui Yan et al.

Object detection models are vulnerable to backdoor attacks, where attackers poison a small subset of training samples by embedding a predefined trigger to manipulate prediction. Detecting poisoned samples (i.e., those containing triggers) at test time can prevent backdoor activation. However, unlike image classification tasks, the unique characteristics of object detection -- particularly its output of numerous objects -- pose fresh challenges for backdoor detection. The complex attack effects (e.g., "ghost" object emergence or "vanishing" object) further render current defenses fundamentally inadequate. To this end, we design TRAnsformation Consistency Evaluation (TRACE), a brand-new method for detecting poisoned samples at test time in object detection. Our journey begins with two intriguing observations: (1) poisoned samples exhibit significantly more consistent detection results than clean ones across varied backgrounds. (2) clean samples show higher detection consistency when introduced to different focal information. Based on these phenomena, TRACE applies foreground and background transformations to each test sample, then assesses transformation consistency by calculating the variance in objects confidences. TRACE achieves black-box, universal backdoor detection, with extensive experiments showing a 30% improvement in AUROC over state-of-the-art defenses and resistance to adaptive attacks.

Yechao Zhang, Yingzhe Xu, Junyu Shi et al.

Deep neural networks (DNNs) are susceptible to universal adversarial perturbations (UAPs). These perturbations are meticulously designed to fool the target model universally across all sample classes. Unlike instance-specific adversarial examples (AEs), generating UAPs is more complex because they must be generalized across a wide range of data samples and models. Our research reveals that existing universal attack methods, which optimize UAPs using DNNs with static model parameter snapshots, do not fully leverage the potential of DNNs to generate more effective UAPs. Rather than optimizing UAPs against static DNN models with a fixed training set, we suggest using dynamic model-data pairs to generate UAPs. In particular, we introduce a dynamic maximin optimization strategy, aiming to optimize the UAP across a variety of optimal model-data pairs. We term this approach DM-UAP. DM-UAP utilizes an iterative max-min-min optimization framework that refines the model-data pairs, coupled with a curriculum UAP learning algorithm to examine the combined space of model parameters and data thoroughly. Comprehensive experiments on the ImageNet dataset demonstrate that the proposed DM-UAP markedly enhances both cross-sample universality and cross-model transferability of UAPs. Using only 500 samples for UAP generation, DM-UAP outperforms the state-of-the-art approach with an average increase in fooling ratio of 12.108%.

Xiaoyan Feng, He Zhang, Yanjun Zhang et al.

Recent advances in Large Language Models (LLMs) have raised urgent concerns about LLM-generated text authenticity, prompting regulatory demands for reliable identification mechanisms. Although watermarking offers a promising solution, existing approaches struggle to simultaneously achieve three critical requirements: text quality preservation, model-agnostic detection, and message embedding capacity, which are crucial for practical implementation. To achieve these goals, the key challenge lies in balancing the trade-off between text quality preservation and message embedding capacity. To address this challenge, we propose BiMark, a novel watermarking framework that achieves these requirements through three key innovations: (1) a bit-flip unbiased reweighting mechanism enabling model-agnostic detection, (2) a multilayer architecture enhancing detectability without compromising generation quality, and (3) an information encoding approach supporting multi-bit watermarking. Through theoretical analysis and extensive experiments, we validate that, compared to state-of-the-art multi-bit watermarking methods, BiMark achieves up to 30% higher extraction rates for short texts while maintaining text quality indicated by lower perplexity, and performs comparably to non-watermarked text on downstream tasks such as summarization and translation.

Zirui Gong, Yanjun Zhang, Leo Yu Zhang et al.

Federated Ranking Learning (FRL) is a state-of-the-art FL framework that stands out for its communication efficiency and resilience to poisoning attacks. It diverges from the traditional FL framework in two ways: 1) it leverages discrete rankings instead of gradient updates, significantly reducing communication costs and limiting the potential space for malicious updates, and 2) it uses majority voting on the server side to establish the global ranking, ensuring that individual updates have minimal influence since each client contributes only a single vote. These features enhance the system's scalability and position FRL as a promising paradigm for FL training. However, our analysis reveals that FRL is not inherently robust, as certain edges are particularly vulnerable to poisoning attacks. Through a theoretical investigation, we prove the existence of these vulnerable edges and establish a lower bound and an upper bound for identifying them in each layer. Based on this finding, we introduce a novel local model poisoning attack against FRL, namely the Vulnerable Edge Manipulation (VEM) attack. The VEM attack focuses on identifying and perturbing the most vulnerable edges in each layer and leveraging an optimization-based approach to maximize the attack's impact. Through extensive experiments on benchmark datasets, we demonstrate that our attack achieves an overall 53.23% attack impact and is 3.7x more impactful than existing methods. Our findings highlight significant vulnerabilities in ranking-based FL systems and underline the urgency for the development of new robust FL frameworks.

Xiaomei Zhang, Zhaoxi Zhang, Yanjun Zhang et al.

Textual adversarial examples pose serious threats to the reliability of natural language processing systems. Recent studies suggest that adversarial examples tend to deviate from the underlying manifold of normal texts, whereas pre-trained masked language models can approximate the manifold of normal data. These findings inspire the exploration of masked language models for detecting textual adversarial attacks. We first introduce Masked Language Model-based Detection (MLMD), leveraging the mask and unmask operations of the masked language modeling (MLM) objective to induce the difference in manifold changes between normal and adversarial texts. Although MLMD achieves competitive detection performance, its exhaustive one-by-one masking strategy introduces significant computational overhead. Our posterior analysis reveals that a significant number of non-keywords in the input are not important for detection but consume resources. Building on this, we introduce Gradient-guided MLMD (GradMLMD), which leverages gradient information to identify and skip non-keywords during detection, significantly reducing resource consumption without compromising detection performance.

Zhaoxi Zhang, Xiaomei Zhang, Yanjun Zhang et al.

Large Language Model (LLM) watermarking embeds detectable signals into generated text for copyright protection, misuse prevention, and content detection. While prior studies evaluate robustness using watermark removal attacks, these methods are often suboptimal, creating the misconception that effective removal requires large perturbations or powerful adversaries. To bridge the gap, we first formalize the system model for LLM watermark, and characterize two realistic threat models constrained on limited access to the watermark detector. We then analyze how different types of perturbation vary in their attack range, i.e., the number of tokens they can affect with a single edit. We observe that character-level perturbations (e.g., typos, swaps, deletions, homoglyphs) can influence multiple tokens simultaneously by disrupting the tokenization process. We demonstrate that character-level perturbations are significantly more effective for watermark removal under the most restrictive threat model. We further propose guided removal attacks based on the Genetic Algorithm (GA) that uses a reference detector for optimization. Under a practical threat model with limited black-box queries to the watermark detector, our method demonstrates strong removal performance. Experiments confirm the superiority of character-level perturbations and the effectiveness of the GA in removing watermarks under realistic constraints. Additionally, we argue there is an adversarial dilemma when considering potential defenses: any fixed defense can be bypassed by a suitable perturbation strategy. Motivated by this principle, we propose an adaptive compound character-level attack. Experimental results show that this approach can effectively defeat the defenses. Our findings highlight significant vulnerabilities in existing LLM watermark schemes and underline the urgency for the development of new robust mechanisms.

Bilal Hussain Abbasi, Zirui Gong, Yanjun Zhang et al.

Despite significant advancements in computer vision, semantic segmentation models may be susceptible to backdoor attacks. These attacks, involving hidden triggers, aim to cause the models to misclassify instances of the victim class as the target class when triggers are present, posing serious threats to the reliability of these models. To further explore the field of backdoor attacks against semantic segmentation, in this paper, we propose a simple yet effective backdoor attack called Contextual Segmentation Backdoor Attack (ConSeg). ConSeg leverages the contextual information inherent in semantic segmentation models to enhance backdoor performance. Our method is motivated by an intriguing observation, i.e., when the target class is set as the `co-occurring' class of the victim class, the victim class can be more easily `mis-segmented'. Building upon this insight, ConSeg mimics the contextual information of the target class and rebuilds it in the victim region to establish the contextual relationship between the target class and the victim class, making the attack easier. Our experiments reveal that ConSeg achieves improvements in Attack Success Rate (ASR) with increases of 15.55\%, compared to existing methods, while exhibiting resilience against state-of-the-art backdoor defenses.

Lulu Xue, Shengshan Hu, Wei Lu et al.

With growing demands for privacy protection, security, and legal compliance (e.g., GDPR), machine unlearning has emerged as a critical technique for ensuring the controllability and regulatory alignment of machine learning models. However, a fundamental challenge in this field lies in effectively verifying whether unlearning operations have been successfully and thoroughly executed. Despite a growing body of work on unlearning techniques, verification methodologies remain comparatively underexplored and often fragmented. Existing approaches lack a unified taxonomy and a systematic framework for evaluation. To bridge this gap, this paper presents the first structured survey of machine unlearning verification methods. We propose a taxonomy that organizes current techniques into two principal categories -- behavioral verification and parametric verification -- based on the type of evidence used to assess unlearning fidelity. We examine representative methods within each category, analyze their underlying assumptions, strengths, and limitations, and identify potential vulnerabilities in practical deployment. In closing, we articulate a set of open problems in current verification research, aiming to provide a foundation for developing more robust, efficient, and theoretically grounded unlearning verification mechanisms.

Huazi Pan, Yanjun Zhang, Leo Yu Zhang et al.

Manipulation of local training data and local updates, i.e., the poisoning attack, is the main threat arising from the collaborative nature of the federated learning (FL) paradigm. Most existing poisoning attacks aim to manipulate local data/models in a way that causes denial-of-service (DoS) issues. In this paper, we introduce a novel attack method, named Federated Learning Sliding Attack (FedSA) scheme, aiming at precisely introducing the extent of poisoning in a subtle controlled manner. It operates with a predefined objective, such as reducing global model's prediction accuracy by 10%. FedSA integrates robust nonlinear control-Sliding Mode Control (SMC) theory with model poisoning attacks. It can manipulate the updates from malicious clients to drive the global model towards a compromised state, achieving this at a controlled and inconspicuous rate. Additionally, leveraging the robust control properties of FedSA allows precise control over the convergence bounds, enabling the attacker to set the global accuracy of the poisoned model to any desired level. Experimental results demonstrate that FedSA can accurately achieve a predefined global accuracy with fewer malicious clients while maintaining a high level of stealth and adjustable learning rates.

Jiaheng Wei, Yanjun Zhang, Leo Yu Zhang et al.

Deep Learning (DL) powered by Deep Neural Networks (DNNs) has revolutionized various domains, yet understanding the intricacies of DNN decision-making and learning processes remains a significant challenge. Recent investigations have uncovered an interesting memorization phenomenon in which DNNs tend to memorize specific details from examples rather than learning general patterns, affecting model generalization, security, and privacy. This raises critical questions about the nature of generalization in DNNs and their susceptibility to security breaches. In this survey, we present a systematic framework to organize memorization definitions based on the generalization and security/privacy domains and summarize memorization evaluation methods at both the example and model levels. Through a comprehensive literature review, we explore DNN memorization behaviors and their impacts on security and privacy. We also introduce privacy vulnerabilities caused by memorization and the phenomenon of forgetting and explore its connection with memorization. Furthermore, we spotlight various applications leveraging memorization and forgetting mechanisms, including noisy label learning, privacy preservation, and model enhancement. This survey offers the first-in-kind understanding of memorization in DNNs, providing insights into its challenges and opportunities for enhancing AI development while addressing critical ethical concerns.

Hangtao Zhang, Shengshan Hu, Yichen Wang et al.

Object detection tasks, crucial in safety-critical systems like autonomous driving, focus on pinpointing object locations. These detectors are known to be susceptible to backdoor attacks. However, existing backdoor techniques have primarily been adapted from classification tasks, overlooking deeper vulnerabilities specific to object detection. This paper is dedicated to bridging this gap by introducing Detector Collapse} (DC), a brand-new backdoor attack paradigm tailored for object detection. DC is designed to instantly incapacitate detectors (i.e., severely impairing detector's performance and culminating in a denial-of-service). To this end, we develop two innovative attack schemes: Sponge for triggering widespread misidentifications and Blinding for rendering objects invisible. Remarkably, we introduce a novel poisoning strategy exploiting natural objects, enabling DC to act as a practical backdoor in real-world environments. Our experiments on different detectors across several benchmarks show a significant improvement ($\sim$10\%-60\% absolute and $\sim$2-7$\times$ relative) in attack efficacy over state-of-the-art attacks.

Yanjun Zhang, Guangdong Bai, Xue Li et al.

Federated learning enables multiple participants to collaboratively train a model without aggregating the training data. Although the training data are kept within each participant and the local gradients can be securely synthesized, recent studies have shown that such privacy protection is insufficient. The global model parameters that have to be shared for optimization are susceptible to leak information about training data. In this work, we propose Confined Gradient Descent (CGD) that enhances privacy of federated learning by eliminating the sharing of global model parameters. CGD exploits the fact that a gradient descent optimization can start with a set of discrete points and converges to another set at the neighborhood of the global minimum of the objective function. It lets the participants independently train on their local data, and securely share the sum of local gradients to benefit each other. We formally demonstrate CGD's privacy enhancement over traditional FL. We prove that less information is exposed in CGD compared to that of traditional FL. CGD also guarantees desired model accuracy. We theoretically establish a convergence rate for CGD. We prove that the loss of the proprietary models learned for each participant against a model learned by aggregated training data is bounded. Extensive experimental results on two real-world datasets demonstrate the performance of CGD is comparable with the centralized learning, with marginal differences on validation loss (mostly within 0.05) and accuracy (mostly within 1%).

Tian Gao, Zishen Wan, Yuyang Zhang et al.

Stereo matching is a critical task for robot navigation and autonomous vehicles, providing the depth estimation of surroundings. Among all stereo matching algorithms, Efficient Large-scale Stereo (ELAS) offers one of the best tradeoffs between efficiency and accuracy. However, due to the inherent iterative process and unpredictable memory access pattern, ELAS can only run at 1.5-3 fps on high-end CPUs and difficult to achieve real-time performance on low-power platforms. In this paper, we propose an energy-efficient architecture for real-time ELAS-based stereo matching on FPGA platform. Moreover, the original computational-intensive and irregular triangulation module is reformed in a regular manner with points interpolation, which is much more hardware-friendly. Optimizations, including memory management, parallelism, and pipelining, are further utilized to reduce memory footprint and improve throughput. Compared with Intel i7 CPU and the state-of-the-art CPU+FPGA implementation, our FPGA realization achieves up to 38.4x and 3.32x frame rate improvement, and up to 27.1x and 1.13x energy efficiency improvement, respectively.

Zishen Wan, Yuyang Zhang, Arijit Raychowdhury et al.

In our past few years' of commercial deployment experiences, we identify localization as a critical task in autonomous machine applications, and a great acceleration target. In this paper, based on the observation that the visual frontend is a major performance and energy consumption bottleneck, we present our design and implementation of an energy-efficient hardware architecture for ORB (Oriented-Fast and Rotated- BRIEF) based localization system on FPGAs. To support our multi-sensor autonomous machine localization system, we present hardware synchronization, frame-multiplexing, and parallelization techniques, which are integrated in our design. Compared to Nvidia TX1 and Intel i7, our FPGA-based implementation achieves 5.6x and 3.4x speedup, as well as 3.0x and 34.6x power reduction, respectively.

Yanjun Zhang, Guangdong Bai, Xue Li et al.

Collaborative learning enables two or more participants, each with their own training dataset, to collaboratively learn a joint model. It is desirable that the collaboration should not cause the disclosure of either the raw datasets of each individual owner or the local model parameters trained on them. This privacy-preservation requirement has been approached through differential privacy mechanisms, homomorphic encryption (HE) and secure multiparty computation (MPC), but existing attempts may either introduce the loss of model accuracy or imply significant computational and/or communicational overhead. In this work, we address this problem with the lightweight additive secret sharing technique. We propose PrivColl, a framework for protecting local data and local models while ensuring the correctness of training processes. PrivColl employs secret sharing technique for securely evaluating addition operations in a multiparty computation environment, and achieves practicability by employing only the homomorphic addition operations. We formally prove that it guarantees privacy preservation even though the majority (n-2 out of n) of participants are corrupted. With experiments on real-world datasets, we further demonstrate that PrivColl retains high efficiency. It achieves a speedup of more than 45X over the state-of-the-art MPC/HE based schemes for training linear/logistic regression, and 216X faster for training neural network.

Lvchen Cao, Huiqi Li, Yanjun Zhang et al.

Cataracts, which are lenticular opacities that may occur at different lens locations, are the leading cause of visual impairment worldwide. Accurate and timely diagnosis can improve the quality of life of cataract patients. In this paper, a feature extraction-based method for grading cataract severity using retinal images is proposed. To obtain more appropriate features for the automatic grading, the Haar wavelet is improved according to the characteristics of retinal images. Retinal images of non-cataract, as well as mild, moderate, and severe cataracts, are automatically recognized using the improved Haar wavelet. A hierarchical strategy is used to transform the four-class classification problem into three adjacent two-class classification problems. Three sets of two-class classifiers based on a neural network are trained individually and integrated together to establish a complete classification system. The accuracies of the two-class classification (cataract and non-cataract) and four-class classification are 94.83% and 85.98%, respectively. The performance analysis demonstrates that the improved Haar wavelet feature achieves higher accuracy than the original Haar wavelet feature, and the fusion of three sets of two-class classifiers is superior to a simple four-class classifier. The discussion indicates that the retinal image-based method offers significant potential for cataract detection.

Yanjun Zhang, Tie Li, Guangyu Na et al.

A new optimized extreme learning machine- (ELM-) based method for power system transient stability prediction (TSP) using synchrophasors is presented in this paper. First, the input features symbolizing the transient stability of power systems are extracted from synchronized measurements. Then, an ELM classifier is employed to build the TSP model. And finally, the optimal parameters of the model are optimized by using the improved particle swarm optimization (IPSO) algorithm. The novelty of the proposal is in the fact that it improves the prediction performance of the ELM-based TSP model by using IPSO to optimize the parameters of the model with synchrophasors. And finally, based on the test results on both IEEE 39-bus system and a large-scale real power system, the correctness and validity of the presented approach are verified.

Weikang Fang, Yanjun Zhang, Bo Yu et al.

Simultaneous Localization And Mapping (SLAM) is the problem of constructing or updating a map of an unknown environment while simultaneously keeping track of an agent's location within it. How to enable SLAM robustly and durably on mobile, or even IoT grade devices, is the main challenge faced by the industry today. The main problems we need to address are: 1.) how to accelerate the SLAM pipeline to meet real-time requirements; and 2.) how to reduce SLAM energy consumption to extend battery life. After delving into the problem, we found out that feature extraction is indeed the bottleneck of performance and energy consumption. Hence, in this paper, we design, implement, and evaluate a hardware ORB feature extractor and prove that our design is a great balance between performance and energy consumption compared with ARM Krait and Intel Core i5.