Jinyin Chen

Haibin Zheng, Haiyang Xiong, Haonan Ma et al.

Link prediction, inferring the undiscovered or potential links of the graph, is widely applied in the real-world. By facilitating labeled links of the graph as the training data, numerous deep learning based link prediction methods have been studied, which have dominant prediction accuracy compared with non-deep methods. However,the threats of maliciously crafted training graph will leave a specific backdoor in the deep model, thus when some specific examples are fed into the model, it will make wrong prediction, defined as backdoor attack. It is an important aspect that has been overlooked in the current literature. In this paper, we prompt the concept of backdoor attack on link prediction, and propose Link-Backdoor to reveal the training vulnerability of the existing link prediction methods. Specifically, the Link-Backdoor combines the fake nodes with the nodes of the target link to form a trigger. Moreover, it optimizes the trigger by the gradient information from the target model. Consequently, the link prediction model trained on the backdoored dataset will predict the link with trigger to the target state. Extensive experiments on five benchmark datasets and five well-performing link prediction models demonstrate that the Link-Backdoor achieves the state-of-the-art attack success rate under both white-box (i.e., available of the target model parameter)and black-box (i.e., unavailable of the target model parameter) scenarios. Additionally, we testify the attack under defensive circumstance, and the results indicate that the Link-Backdoor still can construct successful attack on the well-performing link prediction methods. The code and data are available at https://github.com/Seaocn/Link-Backdoor.

Jinyin Chen, Haibin Zheng, Tao Liu et al.

With the development of deep learning processors and accelerators, deep learning models have been widely deployed on edge devices as part of the Internet of Things. Edge device models are generally considered as valuable intellectual properties that are worth for careful protection. Unfortunately, these models have a great risk of being stolen or illegally copied. The existing model protections using encryption algorithms are suffered from high computation overhead which is not practical due to the limited computing capacity on edge devices. In this work, we propose a light-weight, practical, and general Edge device model Pro tection method at neuron level, denoted as EdgePro. Specifically, we select several neurons as authorization neurons and set their activation values to locking values and scale the neuron outputs as the "asswords" during training. EdgePro protects the model by ensuring it can only work correctly when the "passwords" are met, at the cost of encrypting and storing the information of the "passwords" instead of the whole model. Extensive experimental results indicate that EdgePro can work well on the task of protecting on datasets with different modes. The inference time increase of EdgePro is only 60% of state-of-the-art methods, and the accuracy loss is less than 1%. Additionally, EdgePro is robust against adaptive attacks including fine-tuning and pruning, which makes it more practical in real-world applications. EdgePro is also open sourced to facilitate future research: https://github.com/Leon022/Edg

Haibin Zheng, Haiyang Xiong, Jinyin Chen et al.

Graph neural network (GNN) with a powerful representation capability has been widely applied to various areas, such as biological gene prediction, social recommendation, etc. Recent works have exposed that GNN is vulnerable to the backdoor attack, i.e., models trained with maliciously crafted training samples are easily fooled by patched samples. Most of the proposed studies launch the backdoor attack using a trigger that either is the randomly generated subgraph (e.g., erdős-rényi backdoor) for less computational burden, or the gradient-based generative subgraph (e.g., graph trojaning attack) to enable a more effective attack. However, the interpretation of how is the trigger structure and the effect of the backdoor attack related has been overlooked in the current literature. Motifs, recurrent and statistically significant sub-graphs in graphs, contain rich structure information. In this paper, we are rethinking the trigger from the perspective of motifs, and propose a motif-based backdoor attack, denoted as Motif-Backdoor. It contributes from three aspects. (i) Interpretation: it provides an in-depth explanation for backdoor effectiveness by the validity of the trigger structure from motifs, leading to some novel insights, e.g., using subgraphs that appear less frequently in the graph as the trigger can achieve better attack performance. (ii) Effectiveness: Motif-Backdoor reaches the state-of-the-art (SOTA) attack performance in both black-box and defensive scenarios. (iii) Efficiency: based on the graph motif distribution, Motif-Backdoor can quickly obtain an effective trigger structure without target model feedback or subgraph model generation. Extensive experimental results show that Motif-Backdoor realizes the SOTA performance on three popular models and four public datasets compared with five baselines.

Jinyin Chen, Mingjun Li, Tao Liu et al.

Federated learning (FL) is a distributed machine learning approach where multiple clients collaboratively train a joint model without exchanging their data. Despite FL's unprecedented success in data privacy-preserving, its vulnerability to free-rider attacks has attracted increasing attention. Existing defenses may be ineffective against highly camouflaged or high percentages of free riders. To address these challenges, we reconsider the defense from a novel perspective, i.e., model weight evolving frequency.Empirically, we gain a novel insight that during the FL's training, the model weight evolving frequency of free-riders and that of benign clients are significantly different. Inspired by this insight, we propose a novel defense method based on the model Weight Evolving Frequency, referred to as WEF-Defense.Specifically, we first collect the weight evolving frequency (defined as WEF-Matrix) during local training. For each client, it uploads the local model's WEF-Matrix to the server together with its model weight for each iteration. The server then separates free-riders from benign clients based on the difference in the WEF-Matrix. Finally, the server uses a personalized approach to provide different global models for corresponding clients. Comprehensive experiments conducted on five datasets and five models demonstrate that WEF-Defense achieves better defense effectiveness than the state-of-the-art baselines.

Jinyin Chen, Mingjun Li, Mingjun Li et al.

Federated learning (FL), an effective distributed machine learning framework, implements model training and meanwhile protects local data privacy. It has been applied to a broad variety of practice areas due to its great performance and appreciable profits. Who owns the model, and how to protect the copyright has become a real problem. Intuitively, the existing property rights protection methods in centralized scenarios (e.g., watermark embedding and model fingerprints) are possible solutions for FL. But they are still challenged by the distributed nature of FL in aspects of the no data sharing, parameter aggregation, and federated training settings. For the first time, we formalize the problem of copyright protection for FL, and propose FedRight to protect model copyright based on model fingerprints, i.e., extracting model features by generating adversarial examples as model fingerprints. FedRight outperforms previous works in four key aspects: (i) Validity: it extracts model features to generate transferable fingerprints to train a detector to verify the copyright of the model. (ii) Fidelity: it is with imperceptible impact on the federated training, thus promising good main task performance. (iii) Robustness: it is empirically robust against malicious attacks on copyright protection, i.e., fine-tuning, model pruning, and adaptive attacks. (iv) Black-box: it is valid in the black-box forensic scenario where only application programming interface calls to the model are available. Extensive evaluations across 3 datasets and 9 model structures demonstrate FedRight's superior fidelity, validity, and robustness.

Jinyin Chen, Chengyu Jia, Haibin Zheng et al.

The proliferation of fake news and its serious negative social influence push fake news detection methods to become necessary tools for web managers. Meanwhile, the multi-media nature of social media makes multi-modal fake news detection popular for its ability to capture more modal features than uni-modal detection methods. However, current literature on multi-modal detection is more likely to pursue the detection accuracy but ignore the robustness of the detector. To address this problem, we propose a comprehensive robustness evaluation of multi-modal fake news detectors. In this work, we simulate the attack methods of malicious users and developers, i.e., posting fake news and injecting backdoors. Specifically, we evaluate multi-modal detectors with five adversarial and two backdoor attack methods. Experiment results imply that: (1) The detection performance of the state-of-the-art detectors degrades significantly under adversarial attacks, even worse than general detectors; (2) Most multi-modal detectors are more vulnerable when subjected to attacks on visual modality than textual modality; (3) Popular events' images will cause significant degradation to the detectors when they are subjected to backdoor attacks; (4) The performance of these detectors under multi-modal attacks is worse than under uni-modal attacks; (5) Defensive methods will improve the robustness of the multi-modal detectors.

Bin Zhu, Zhaoquan Gu, Le Wang et al.

Recent studies have found that removing the norm-bounded projection and increasing search steps in adversarial training can significantly improve robustness. However, we observe that a too large number of search steps can hurt accuracy. We aim to obtain strong robustness efficiently using fewer steps. Through a toy experiment, we find that perturbing the clean data to the decision boundary but not crossing it does not degrade the test accuracy. Inspired by this, we propose friendly adversarial data augmentation (FADA) to generate friendly adversarial data. On top of FADA, we propose geometry-aware adversarial training (GAT) to perform adversarial training on friendly adversarial data so that we can save a large number of search steps. Comprehensive experiments across two widely used datasets and three pre-trained language models demonstrate that GAT can obtain stronger robustness via fewer steps. In addition, we provide extensive empirical results and in-depth analyses on robustness to facilitate future studies.

Haibin Zheng, Jinyin Chen, Haibo Jin

Deep neural networks (DNNs) have demonstrated their outperformance in various software systems, but also exhibit misbehavior and even result in irreversible disasters. Therefore, it is crucial to identify the misbehavior of DNN-based software and improve DNNs' quality. Test input prioritization is one of the most appealing ways to guarantee DNNs' quality, which prioritizes test inputs so that more bug-revealing inputs can be identified earlier with limited time and manual labeling efforts. However, the existing prioritization methods are still limited from three aspects: certifiability, effectiveness, and generalizability. To overcome the challenges, we propose CertPri, a test input prioritization technique designed based on a movement cost perspective of test inputs in DNNs' feature space. CertPri differs from previous works in three key aspects: (1) certifiable: it provides a formal robustness guarantee for the movement cost; (2) effective: it leverages formally guaranteed movement costs to identify malicious bug-revealing inputs; and (3) generic: it can be applied to various tasks, data, models, and scenarios. Extensive evaluations across 2 tasks (i.e., classification and regression), 6 data forms, 4 model structures, and 2 scenarios (i.e., white-box and black-box) demonstrate CertPri's superior performance. For instance, it significantly improves 53.97% prioritization effectiveness on average compared with baselines. Its robustness and generalizability are 1.41~2.00 times and 1.33~3.39 times that of baselines on average, respectively.

Jinyin Chen, Shulong Hu, Haibin Zheng et al.

Penetration testing (PT) is an efficient network testing and vulnerability mining tool by simulating a hacker's attack for valuable information applied in some areas. Compared with manual PT, intelligent PT has become a dominating mainstream due to less time-consuming and lower labor costs. Unfortunately, RL-based PT is still challenged in real exploitation scenarios because the agent's action space is usually high-dimensional discrete, thus leading to algorithm convergence difficulty. Besides, most PT methods still rely on the decisions of security experts. Addressing the challenges, for the first time, we introduce expert knowledge to guide the agent to make better decisions in RL-based PT and propose a Generative Adversarial Imitation Learning-based generic intelligent Penetration testing framework, denoted as GAIL-PT, to solve the problems of higher labor costs due to the involvement of security experts and high-dimensional discrete action space. Specifically, first, we manually collect the state-action pairs to construct an expert knowledge base when the pre-trained RL / DRL model executes successful penetration testings. Second, we input the expert knowledge and the state-action pairs generated online by the different RL / DRL models into the discriminator of GAIL for training. At last, we apply the output reward of the discriminator to guide the agent to perform the action with a higher penetration success rate to improve PT's performance. Extensive experiments conducted on the real target host and simulated network scenarios show that GAIL-PT achieves the SOTA penetration performance against DeepExploit in exploiting actual target Metasploitable2 and Q-learning in optimizing penetration path, not only in small-scale with or without honey-pot network environments but also in the large-scale virtual network environment.

Jinyin Chen, Jie Ge, Shilian Zheng et al.

A wireless communications system usually consists of a transmitter which transmits the information and a receiver which recovers the original information from the received distorted signal. Deep learning (DL) has been used to improve the performance of the receiver in complicated channel environments and state-of-the-art (SOTA) performance has been achieved. However, its robustness has not been investigated. In order to evaluate the robustness of DL-based information recovery models under adversarial circumstances, we investigate adversarial attacks on the SOTA DL-based information recovery model, i.e., DeepReceiver. We formulate the problem as an optimization problem with power and peak-to-average power ratio (PAPR) constraints. We design different adversarial attack methods according to the adversary's knowledge of DeepReceiver's model and/or testing samples. Extensive experiments show that the DeepReceiver is vulnerable to the designed attack methods in all of the considered scenarios. Even in the scenario of both model and test sample restricted, the adversary can attack the DeepReceiver and increase its bit error rate (BER) above 10%. It can also be found that the DeepReceiver is vulnerable to adversarial perturbations even with very low power and limited PAPR. These results suggest that defense measures should be taken to enhance the robustness of DeepReceiver.

Ruoxi Chen, Haibo Jin, Jinyin Chen et al.

Deep neural networks (DNNs) are vulnerable to adversarial examples, which may lead to catastrophe in security-critical domains. Numerous detection methods are proposed to characterize the feature uniqueness of adversarial examples, or to distinguish DNN's behavior activated by the adversarial examples. Detections based on features cannot handle adversarial examples with large perturbations. Besides, they require a large amount of specific adversarial examples. Another mainstream, model-based detections, which characterize input properties by model behaviors, suffer from heavy computation cost. To address the issues, we introduce the concept of local gradient, and reveal that adversarial examples have a quite larger bound of local gradient than the benign ones. Inspired by the observation, we leverage local gradient for detecting adversarial examples, and propose a general framework AdvCheck. Specifically, by calculating the local gradient from a few benign examples and noise-added misclassified examples to train a detector, adversarial examples and even misclassified natural inputs can be precisely distinguished from benign ones. Through extensive experiments, we have validated the AdvCheck's superior performance to the state-of-the-art (SOTA) baselines, with detection rate ($\sim \times 1.2$) on general adversarial attacks and ($\sim \times 1.4$) on misclassified natural inputs on average, with average 1/500 time cost. We also provide interpretable results for successful detection.

Ruoxi Chen, Haibo Jin, Yixin Liu et al.

Text-to-image diffusion models have emerged as an evolutionary for producing creative content in image synthesis. Based on the impressive generation abilities of these models, instruction-guided diffusion models can edit images with simple instructions and input images. While they empower users to obtain their desired edited images with ease, they have raised concerns about unauthorized image manipulation. Prior research has delved into the unauthorized use of personalized diffusion models; however, this problem of instruction-guided diffusion models remains largely unexplored. In this paper, we first propose a protection method EditShield against unauthorized modifications from such models. Specifically, EditShield works by adding imperceptible perturbations that can shift the latent representation used in the diffusion process, tricking models into generating unrealistic images with mismatched subjects. Our extensive experiments demonstrate EditShield's effectiveness among synthetic and real-world datasets. Besides, we found that EditShield performs robustly against various manipulation settings across editing types and synonymous instruction phrases.

Yao Lu, Wen Yang, Yunzhe Zhang et al.

There are good arguments to support the claim that deep neural networks (DNNs) capture better feature representations than the previous hand-crafted feature engineering, which leads to a significant performance improvement. In this paper, we move a tiny step towards understanding the dynamics of feature representations over layers. Specifically, we model the process of class separation of intermediate representations in pre-trained DNNs as the evolution of communities in dynamic graphs. Then, we introduce modularity, a generic metric in graph theory, to quantify the evolution of communities. In the preliminary experiment, we find that modularity roughly tends to increase as the layer goes deeper and the degradation and plateau arise when the model complexity is great relative to the dataset. Through an asymptotic analysis, we prove that modularity can be broadly used for different applications. For example, modularity provides new insights to quantify the difference between feature representations. More crucially, we demonstrate that the degradation and plateau in modularity curves represent redundant layers in DNNs and can be pruned with minimal impact on performance, which provides theoretical guidance for layer pruning. Our code is available at https://github.com/yaolu-zjut/Dynamic-Graphs-Construction.

Jinyin Chen, Guohan Huang, Haibin Zheng et al.

Graph neural network (GNN) has achieved great success on graph representation learning. Challenged by large scale private data collected from user-side, GNN may not be able to reflect the excellent performance, without rich features and complete adjacent relationships. Addressing the problem, vertical federated learning (VFL) is proposed to implement local data protection through training a global model collaboratively. Consequently, for graph-structured data, it is a natural idea to construct a GNN based VFL framework, denoted as GVFL. However, GNN has been proved vulnerable to adversarial attacks. Whether the vulnerability will be brought into the GVFL has not been studied. This is the first study of adversarial attacks on GVFL. A novel adversarial attack method is proposed, named Graph-Fraudster. It generates adversarial perturbations based on the noise-added global node embeddings via the privacy leakage and the gradient of pairwise node. Specifically, first, Graph-Fraudster steals the global node embeddings and sets up a shadow model of the server for the attack generator. Second, noise is added into node embeddings to confuse the shadow model. At last, the gradient of pairwise node is used to generate attacks with the guidance of noise-added node embeddings. Extensive experiments on five benchmark datasets demonstrate that Graph-Fraudster achieves the state-of-the-art attack performance compared with baselines in different GNN based GVFLs. Furthermore, Graph-Fraudster can remain a threat to GVFL even if two possible defense mechanisms are applied. Additionally, some suggestions are put forward for the future work to improve the robustness of GVFL. The code and datasets can be downloaded at https://github.com/hgh0545/Graph-Fraudster.

Jinyin Chen, Dunjie Zhang, Zhaoyan Ming et al.

Graph neural networks (GNNs) have been successfully exploited in graph analysis tasks in many real-world applications. The competition between attack and defense methods also enhances the robustness of GNNs. In this competition, the development of adversarial training methods put forward higher requirement for the diversity of attack examples. By contrast, most attack methods with specific attack strategies are difficult to satisfy such a requirement. To address this problem, we propose GraphAttacker, a novel generic graph attack framework that can flexibly adjust the structures and the attack strategies according to the graph analysis tasks. GraphAttacker generates adversarial examples through alternate training on three key components: the multi-strategy attack generator (MAG), the similarity discriminator (SD), and the attack discriminator (AD), based on the generative adversarial network (GAN). Furthermore, we introduce a novel similarity modification rate SMR to conduct a stealthier attack considering the change of node similarity distribution. Experiments on various benchmark datasets demonstrate that GraphAttacker can achieve state-of-the-art attack performance on graph analysis tasks of node classification, graph classification, and link prediction, no matter the adversarial training is conducted or not. Moreover, we also analyze the unique characteristics of each task and their specific response in the unified attack framework. The project code is available at https://github.com/honoluluuuu/GraphAttacker.

Jinyin Chen, Zhen Wang, Haibin Zheng et al.

With the successful application of deep learning models in many real-world tasks, the model robustness becomes more and more critical. Often, we evaluate the robustness of the deep models by attacking them with purposely generated adversarial samples, which is computationally costly and dependent on the specific attackers and the model types. This work proposes a generic evaluation metric ROBY, a novel attack-independent robustness measure based on the model's decision boundaries. Independent of adversarial samples, ROBY uses the inter-class and intra-class statistic features to capture the features of the model's decision boundaries. We experimented on ten state-of-the-art deep models and showed that ROBY matches the robustness gold standard of attack success rate (ASR) by a strong first-order generic attacker. with only 1% of time cost. To the best of our knowledge, ROBY is the first lightweight attack-independent robustness evaluation metric that can be applied to a wide range of deep models. The code of ROBY is open sourced at https://github.com/baaaad/ROBY-Evaluating-the-Robustness-of-a-Deep-Model-by-its-Decision-Boundaries.

Mingqi Lv, HongZhe Gao, Xuebo Qiu et al.

APT (Advanced Persistent Threat) with the characteristics of persistence, stealth, and diversity is one of the greatest threats against cyber-infrastructure. As a countermeasure, existing studies leverage provenance graphs to capture the complex relations between system entities in a host for effective APT detection. In addition to detecting single attack events as most existing work does, understanding the tactics / techniques (e.g., Kill-Chain, ATT&CK) applied to organize and accomplish the APT attack campaign is more important for security operations. Existing studies try to manually design a set of rules to map low-level system events to high-level APT tactics / techniques. However, the rule based methods are coarse-grained and lack generalization ability, thus they can only recognize APT tactics and cannot identify fine-grained APT techniques and mutant APT attacks. In this paper, we propose TREC, the first attempt to recognize APT tactics / techniques from provenance graphs by exploiting deep learning techniques. To address the "needle in a haystack" problem, TREC segments small and compact subgraphs covering individual APT technique instances from a large provenance graph based on a malicious node detection model and a subgraph sampling algorithm. To address the "training sample scarcity" problem, TREC trains the APT tactic / technique recognition model in a few-shot learning manner by adopting a Siamese neural network. We evaluate TREC based on a customized dataset collected and made public by our team. The experiment results show that TREC significantly outperforms state-of-the-art systems in APT tactic recognition and TREC can also effectively identify APT techniques.

Jia Chengyu, AprilPyone MaungMaung, Huy H. Nguyen et al.

Recent advances in vision-language models (VLMs) trained on web-scale image-text pairs have enabled impressive zero-shot transfer across a diverse range of visual tasks. However, comprehensive and independent evaluation beyond standard benchmarks is essential to understand their robustness, limitations, and real-world applicability. This paper presents a systematic evaluation framework for VLMs under natural adversarial scenarios for diverse downstream tasks, which has been overlooked in previous evaluation works. We evaluate a wide range of VLMs (CLIP, robust CLIP, BLIP2, and SigLIP2) on curated adversarial datasets (typographic attacks, ImageNet-A, and natural language-induced adversarial examples). We measure the natural adversarial performance of selected VLMs for zero-shot image classification, semantic segmentation, and visual question answering. Our analysis reveals that robust CLIP models can amplify natural adversarial vulnerabilities, and CLIP models significantly reduce performance for natural language-induced adversarial examples. Additionally, we provide interpretable analyses to identify failure modes. We hope our findings inspire future research in robust and fair multimodal pattern recognition.

Jinyin Chen, Haonan Ma, Haibin Zheng

Graph has become increasingly integral to the advancement of recommendation systems, particularly with the fast development of graph neural network(GNN). By exploring the virtue of rich node features and link information, GNN is designed to provide personalized and accurate suggestions. Meanwhile, the privacy leakage of GNN in such contexts has also captured special attention. Prior work has revealed that a malicious user can utilize auxiliary knowledge to extract sensitive link data of the target graph, integral to recommendation systems, via the decision made by the target GNN model. This poses a significant risk to the integrity and confidentiality of data used in recommendation system. Though important, previous works on GNN's privacy leakage are still challenged in three aspects, i.e., limited stealing attack scenarios, sub-optimal attack performance, and adaptation against defense. To address these issues, we propose a diffusion model based link stealing attack, named DM4Steal. It differs previous work from three critical aspects. (i) Generality: aiming at six attack scenarios with limited auxiliary knowledge, we propose a novel training strategy for diffusion models so that DM4Steal is transferable to diverse attack scenarios. (ii) Effectiveness: benefiting from the retention of semantic structure in the diffusion model during the training process, DM4Steal is capable to learn the precise topology of the target graph through the GNN decision process. (iii) Adaptation: when GNN is defensive (e.g., DP, Dropout), DM4Steal relies on the stability that comes from sampling the score model multiple times to keep performance degradation to a minimum, thus DM4Steal implements successful adaptive attack on defensive GNN.

Jinyin Chen, Wenbo Mu, Luxin Zhang et al.

Graph neural network (GNN) has captured wide attention due to its capability of graph representation learning for graph-structured data. However, the distributed data silos limit the performance of GNN. Vertical federated learning (VFL), an emerging technique to process distributed data, successfully makes GNN possible to handle the distributed graph-structured data. Despite the prosperous development of vertical federated graph learning (VFGL), the robustness of VFGL against the adversarial attack has not been explored yet. Although numerous adversarial attacks against centralized GNNs are proposed, their attack performance is challenged in the VFGL scenario. To the best of our knowledge, this is the first work to explore the adversarial attack against VFGL. A query-efficient hybrid adversarial attack framework is proposed to significantly improve the centralized adversarial attacks against VFGL, denoted as NA2, short for Neuron-based Adversarial Attack. Specifically, a malicious client manipulates its local training data to improve its contribution in a stealthy fashion. Then a shadow model is established based on the manipulated data to simulate the behavior of the server model in VFGL. As a result, the shadow model can improve the attack success rate of various centralized attacks with a few queries. Extensive experiments on five real-world benchmarks demonstrate that NA2 improves the performance of the centralized adversarial attacks against VFGL, achieving state-of-the-art performance even under potential adaptive defense where the defender knows the attack method. Additionally, we provide interpretable experiments of the effectiveness of NA2 via sensitive neurons identification and visualization of t-SNE.

Jinyin Chen, Danxin Liao, Sheng Xiang et al.

Since DNN is vulnerable to carefully crafted adversarial examples, adversarial attack on LiDAR sensors have been extensively studied. We introduce a robust black-box attack dubbed LiDAttack. It utilizes a genetic algorithm with a simulated annealing strategy to strictly limit the location and number of perturbation points, achieving a stealthy and effective attack. And it simulates scanning deviations, allowing it to adapt to dynamic changes in real world scenario variations. Extensive experiments are conducted on 3 datasets (i.e., KITTI, nuScenes, and self-constructed data) with 3 dominant object detection models (i.e., PointRCNN, PointPillar, and PV-RCNN++). The results reveal the efficiency of the LiDAttack when targeting a wide range of object detection models, with an attack success rate (ASR) up to 90%.

Jinyin Chen, Xiaoming Zhao, Haibin Zheng et al.

Benefiting from well-trained deep neural networks (DNNs), model compression have captured special attention for computing resource limited equipment, especially edge devices. Knowledge distillation (KD) is one of the widely used compression techniques for edge deployment, by obtaining a lightweight student model from a well-trained teacher model released on public platforms. However, it has been empirically noticed that the backdoor in the teacher model will be transferred to the student model during the process of KD. Although numerous KD methods have been proposed, most of them focus on the distillation of a high-performing student model without robustness consideration. Besides, some research adopts KD techniques as effective backdoor mitigation tools, but they fail to perform model compression at the same time. Consequently, it is still an open problem to well achieve two objectives of robust KD, i.e., student model's performance and backdoor mitigation. To address these issues, we propose RobustKD, a robust knowledge distillation that compresses the model while mitigating backdoor based on feature variance. Specifically, RobustKD distinguishes the previous works in three key aspects: (1) effectiveness: by distilling the feature map of the teacher model after detoxification, the main task performance of the student model is comparable to that of the teacher model; (2) robustness: by reducing the characteristic variance between the teacher model and the student model, it mitigates the backdoor of the student model under backdoored teacher model scenario; (3) generic: RobustKD still has good performance in the face of multiple data models (e.g., WRN 28-4, Pyramid-200) and diverse DNNs (e.g., ResNet50, MobileNet).

Haibo Jin, Ruoxi Chen, Haibin Zheng et al.

Despite impressive capabilities and outstanding performance, deep neural networks (DNNs) have captured increasing public concern about their security problems, due to their frequently occurred erroneous behaviors. Therefore, it is necessary to conduct a systematical testing for DNNs before they are deployed to real-world applications. Existing testing methods have provided fine-grained metrics based on neuron coverage and proposed various approaches to improve such metrics. However, it has been gradually realized that a higher neuron coverage does \textit{not} necessarily represent better capabilities in identifying defects that lead to errors. Besides, coverage-guided methods cannot hunt errors due to faulty training procedure. So the robustness improvement of DNNs via retraining by these testing examples are unsatisfactory. To address this challenge, we introduce the concept of excitable neurons based on Shapley value and design a novel white-box testing framework for DNNs, namely DeepSensor. It is motivated by our observation that neurons with larger responsibility towards model loss changes due to small perturbations are more likely related to incorrect corner cases due to potential defects. By maximizing the number of excitable neurons concerning various wrong behaviors of models, DeepSensor can generate testing examples that effectively trigger more errors due to adversarial inputs, polluted data and incomplete training. Extensive experiments implemented on both image classification models and speaker recognition models have demonstrated the superiority of DeepSensor.

Haibin Zheng, Zhiqing Chen, Tianyu Du et al.

Deep neural networks (DNNs) have demonstrated their outperformance in various domains. However, it raises a social concern whether DNNs can produce reliable and fair decisions especially when they are applied to sensitive domains involving valuable resource allocation, such as education, loan, and employment. It is crucial to conduct fairness testing before DNNs are reliably deployed to such sensitive domains, i.e., generating as many instances as possible to uncover fairness violations. However, the existing testing methods are still limited from three aspects: interpretability, performance, and generalizability. To overcome the challenges, we propose NeuronFair, a new DNN fairness testing framework that differs from previous work in several key aspects: (1) interpretable - it quantitatively interprets DNNs' fairness violations for the biased decision; (2) effective - it uses the interpretation results to guide the generation of more diverse instances in less time; (3) generic - it can handle both structured and unstructured data. Extensive evaluations across 7 datasets and the corresponding DNNs demonstrate NeuronFair's superior performance. For instance, on structured datasets, it generates much more instances (~x5.84) and saves more time (with an average speedup of 534.56%) compared with the state-of-the-art methods. Besides, the instances of NeuronFair can also be leveraged to improve the fairness of the biased DNNs, which helps build more fair and trustworthy deep learning systems.

Haibo Jin, Ruoxi Chen, Jinyin Chen et al.

The success of deep neural networks (DNNs) in real-world applications has benefited from abundant pre-trained models. However, the backdoored pre-trained models can pose a significant trojan threat to the deployment of downstream DNNs. Numerous backdoor detection methods have been proposed but are limited to two aspects: (1) high sensitivity on trigger size, especially on stealthy attacks (i.e., blending attacks and defense adaptive attacks); (2) rely heavily on benign examples for reverse engineering. To address these challenges, we empirically observed that trojaned behaviors triggered by various trojan attacks can be attributed to the trojan path, composed of top-$k$ critical neurons with more significant contributions to model prediction changes. Motivated by it, we propose CatchBackdoor, a detection method against trojan attacks. Based on the close connection between trojaned behaviors and trojan path to trigger errors, CatchBackdoor starts from the benign path and gradually approximates the trojan path through differential fuzzing. We then reverse triggers from the trojan path, to trigger errors caused by diverse trojaned attacks. Extensive experiments on MINST, CIFAR-10, and a-ImageNet datasets and 7 models (LeNet, ResNet, and VGG) demonstrate the superiority of CatchBackdoor over the state-of-the-art methods, in terms of (1) \emph{effective} - it shows better detection performance, especially on stealthy attacks ($\sim$ $\times$ 2 on average); (2) \emph{extensible} - it is robust to trigger size and can conduct detection without benign examples.

Ruoxi Chen, Haibo Jin, Haibin Zheng et al.

The vulnerabilities of deep learning models towards adversarial attacks have attracted increasing attention, especially when models are deployed in security-critical domains. Numerous defense methods, including reactive and proactive ones, have been proposed for model robustness improvement. Reactive defenses, such as conducting transformations to remove perturbations, usually fail to handle large perturbations. The proactive defenses that involve retraining, suffer from the attack dependency and high computation cost. In this paper, we consider defense methods from the general effect of adversarial attacks that take on neurons inside the model. We introduce the concept of neuron influence, which can quantitatively measure neurons' contribution to correct classification. Then, we observe that almost all attacks fool the model by suppressing neurons with larger influence and enhancing those with smaller influence. Based on this, we propose \emph{Neuron-level Inverse Perturbation} (NIP), a novel defense against general adversarial attacks. It calculates neuron influence from benign examples and then modifies input examples by generating inverse perturbations that can in turn strengthen neurons with larger influence and weaken those with smaller influence.

Jinyin Chen, Haiyang Xiong, Dunjie Zhang et al.

Recently, phishing scams have posed a significant threat to blockchains. Phishing detectors direct their efforts in hunting phishing addresses. Most of the detectors extract target addresses' transaction behavior features by random walking or constructing static subgraphs. The random walking methods,unfortunately, usually miss structural information due to limited sampling sequence length, while the static subgraph methods tend to ignore temporal features lying in the evolving transaction behaviors. More importantly, their performance undergoes severe degradation when the malicious users intentionally hide phishing behaviors. To address these challenges, we propose TEGDetector, a dynamic graph classifier that learns the evolving behavior features from transaction evolution graphs (TEGs). First, we cast the transaction series into multiple time slices, capturing the target address's transaction behaviors in different periods. Then, we provide a fast non-parametric phishing detector to narrow down the search space of suspicious addresses. Finally, TEGDetector considers both the spatial and temporal evolutions towards a complete characterization of the evolving transaction behaviors. Moreover, TEGDetector utilizes adaptively learnt time coefficient to pay distinct attention to different periods, which provides several novel insights. Extensive experiments on the large-scale Ethereum transaction dataset demonstrate that the proposed method achieves state-of-the-art detection performance.

Jinyin Chen, Haiyang Xiong, Haibin Zheng et al.

Dynamic link prediction (DLP) makes graph prediction based on historical information. Since most DLP methods are highly dependent on the training data to achieve satisfying prediction performance, the quality of the training data is crucial. Backdoor attacks induce the DLP methods to make wrong prediction by the malicious training data, i.e., generating a subgraph sequence as the trigger and embedding it to the training data. However, the vulnerability of DLP toward backdoor attacks has not been studied yet. To address the issue, we propose a novel backdoor attack framework on DLP, denoted as Dyn-Backdoor. Specifically, Dyn-Backdoor generates diverse initial-triggers by a generative adversarial network (GAN). Then partial links of the initial-triggers are selected to form a trigger set, according to the gradient information of the attack discriminator in the GAN, so as to reduce the size of triggers and improve the concealment of the attack. Experimental results show that Dyn-Backdoor launches successful backdoor attacks on the state-of-the-art DLP models with success rate more than 90%. Additionally, we conduct a possible defense against Dyn-Backdoor to testify its resistance in defensive settings, highlighting the needs of defenses for backdoor attacks on DLP.

Dunjie Zhang, Jinyin Chen

With the popularity of blockchain technology, the financial security issues of blockchain transaction networks have become increasingly serious. Phishing scam detection methods will protect possible victims and build a healthier blockchain ecosystem. Usually, the existing works define phishing scam detection as a node classification task by learning the potential features of users through graph embedding methods such as random walk or graph neural network (GNN). However, these detection methods are suffered from high complexity due to the large scale of the blockchain transaction network, ignoring temporal information of the transaction. Addressing this problem, we defined the transaction pattern graphs for users and transformed the phishing scam detection into a graph classification task. To extract richer information from the input graph, we proposed a multi-channel graph classification model (MCGC) with multiple feature extraction channels for GNN. The transaction pattern graphs and MCGC are more able to detect potential phishing scammers by extracting the transaction pattern features of the target users. Extensive experiments on seven benchmark and Ethereum datasets demonstrate that the proposed MCGC can not only achieve state-of-the-art performance in the graph classification task but also achieve effective phishing scam detection based on the target users' transaction pattern graphs.

Jinyin Chen, Haiyang Xiong, Haibin Zhenga et al.

Graph classification is crucial in network analyses. Networks face potential security threats, such as adversarial attacks. Some defense methods may trade off the algorithm complexity for robustness, such as adversarial training, whereas others may trade off clean example performance, such as smoothingbased defense. Most suffer from high complexity or low transferability. To address this problem, we proposed EGC2, an enhanced graph classification model with easy graph compression. EGC2 captures the relationship between the features of different nodes by constructing feature graphs and improving the aggregation of the node-level representations. To achieve lower-complexity defense applied to graph classification models, EGC2 utilizes a centrality-based edge-importance index to compress the graphs, filtering out trivial structures and adversarial perturbations in the input graphs, thus improving the model's robustness. Experiments on ten benchmark datasets demonstrate that the proposed feature read-out and graph compression mechanisms enhance the robustness of multiple basic models, resulting in a state-of-the-art performance in terms of accuracy and robustness against various adversarial attacks.

Jinyin Chen, Ruoxi Chen, Haibin Zheng et al.

Recent years have witnessed unprecedented success achieved by deep learning models in the field of computer vision. However, their vulnerability towards carefully crafted adversarial examples has also attracted the increasing attention of researchers. Motivated by the observation that adversarial examples are due to the non-robust feature learned from the original dataset by models, we propose the concepts of salient feature(SF) and trivial feature(TF). The former represents the class-related feature, while the latter is usually adopted to mislead the model. We extract these two features with coupled generative adversarial network model and put forward a novel detection and defense method named salient feature extractor (SFE) to defend against adversarial attacks. Concretely, detection is realized by separating and comparing the difference between SF and TF of the input. At the same time, correct labels are obtained by re-identifying SF to reach the purpose of defense. Extensive experiments are carried out on MNIST, CIFAR-10, and ImageNet datasets where SFE shows state-of-the-art results in effectiveness and efficiency compared with baselines. Furthermore, we provide an interpretable understanding of the defense and detection process.

Jinyin Chen, Linhui Ye, Zhaoyan Ming

Text-to-Speech (TTS) synthesis plays an important role in human-computer interaction. Currently, most TTS technologies focus on the naturalness of speech, namely,making the speeches sound like humans. However, the key tasks of the expression of emotion and the speaker identity are ignored, which limits the application scenarios of TTS synthesis technology. To make the synthesized speech more realistic and expand the application scenarios, we propose a multi-task anthropomorphic speech synthesis framework (MASS), which can synthesize speeches from text with specified emotion and speaker identity. The MASS framework consists of a base TTS module and two novel voice conversion modules: the emotional voice conversion module and the speaker voice conversion module. We propose deep emotion voice conversion model (DEVC) and deep speaker voice conversion model (DSVC) based on convolution residual networks. It solves the problem of feature loss during voice conversion. The model trainings are independent of parallel datasets, and are capable of many-to-many voice conversion. In the emotional voice conversion, speaker voice conversion experiments, as well as the multi-task speech synthesis experiments, experimental results show DEVC and DSVC convert speech effectively. The quantitative and qualitative evaluation results of multi-task speech synthesis experiments show MASS can effectively synthesis speech with specified text, emotion and speaker identity.

Jinyin Chen, Xiang Lin, Dunjie Zhang et al.

Deep learning is effective in graph analysis. It is widely applied in many related areas, such as link prediction, node classification, community detection, and graph classification etc. Graph embedding, which learns low-dimensional representations for vertices or edges in the graph, usually employs deep models to derive the embedding vector. However, these models are vulnerable. We envision that graph embedding methods based on deep models can be easily attacked using adversarial examples. Thus, in this paper, we propose Graphfool, a novel targeted label adversarial attack on graph embedding. It can generate adversarial graph to attack graph embedding methods via classifying boundary and gradient information in graph convolutional network (GCN). Specifically, we perform the following steps: 1),We first estimate the classification boundaries of different classes. 2), We calculate the minimal perturbation matrix to misclassify the attacked vertex according to the target classification boundary. 3), We modify the adjacency matrix according to the maximal absolute value of the disturbance matrix. This process is implemented iteratively. To the best of our knowledge, this is the first targeted label attack technique. The experiments on real-world graph networks demonstrate that Graphfool can derive better performance than state-of-art techniques. Compared with the second best algorithm, Graphfool can achieve an average improvement of 11.44% in attack success rate.

Jinyin Chen, Longyuan Zhang, Haibin Zheng et al.

Deep neural networks are susceptible to poisoning attacks by purposely polluted training data with specific triggers. As existing episodes mainly focused on attack success rate with patch-based samples, defense algorithms can easily detect these poisoning samples. We propose DeepPoison, a novel adversarial network of one generator and two discriminators, to address this problem. Specifically, the generator automatically extracts the target class' hidden features and embeds them into benign training samples. One discriminator controls the ratio of the poisoning perturbation. The other discriminator works as the target model to testify the poisoning effects. The novelty of DeepPoison lies in that the generated poisoned training samples are indistinguishable from the benign ones by both defensive methods and manual visual inspection, and even benign test samples can achieve the attack. Extensive experiments have shown that DeepPoison can achieve a state-of-the-art attack success rate, as high as 91.74%, with only 7% poisoned samples on publicly available datasets LFW and CASIA. Furthermore, we have experimented with high-performance defense algorithms such as autodecoder defense and DBSCAN cluster detection and showed the resilience of DeepPoison.

Liang Huang, You Zhang, Weijian Pan et al.

Deep learning has recently been successfully applied in automatic modulation classification by extracting and classifying radio features in an end-to-end way. However, deep learning-based radio modulation classifiers are lack of interpretability, and there is little explanation or visibility into what kinds of radio features are extracted and chosen for classification. In this paper, we visualize different deep learning-based radio modulation classifiers by introducing a class activation vector. Specifically, both convolutional neural networks (CNN) based classifier and long short-term memory (LSTM) based classifier are separately studied, and their extracted radio features are visualized. Extensive numerical results show both the CNN-based classifier and LSTM-based classifier extract similar radio features relating to modulation reference points. In particular, for the LSTM-based classifier, its obtained radio features are similar to the knowledge of human experts. Our numerical results indicate the radio features extracted by deep learning-based classifiers greatly depend on the contents carried by radio signals, and a short radio sample may lead to misclassification.

Jinyin Chen, Jian Zhang, Zhi Chen et al.

In network link prediction, it is possible to hide a target link from being predicted with a small perturbation on network structure. This observation may be exploited in many real world scenarios, for example, to preserve privacy, or to exploit financial security. There have been many recent studies to generate adversarial examples to mislead deep learning models on graph data. However, none of the previous work has considered the dynamic nature of real-world systems. In this work, we present the first study of adversarial attack on dynamic network link prediction (DNLP). The proposed attack method, namely time-aware gradient attack (TGA), utilizes the gradient information generated by deep dynamic network embedding (DDNE) across different snapshots to rewire a few links, so as to make DDNE fail to predict target links. We implement TGA in two ways: one is based on traversal search, namely TGA-Tra; and the other is simplified with greedy search for efficiency, namely TGA-Gre. We conduct comprehensive experiments which show the outstanding performance of TGA in attacking DNLP algorithms.

Yun Xiang, Zhuangzhi Chen, Zuohui Chen et al.

Deep neural networks are becoming popular and important assets of many AI companies. However, recent studies indicate that they are also vulnerable to adversarial attacks. Adversarial attacks can be either white-box or black-box. The white-box attacks assume full knowledge of the models while the black-box ones assume none. In general, revealing more internal information can enable much more powerful and efficient attacks. However, in most real-world applications, the internal information of embedded AI devices is unavailable, i.e., they are black-box. Therefore, in this work, we propose a side-channel information based technique to reveal the internal information of black-box models. Specifically, we have made the following contributions: (1) we are the first to use side-channel information to reveal internal network architecture in embedded devices; (2) we are the first to construct models for internal parameter estimation; and (3) we validate our methods on real-world devices and applications. The experimental results show that our method can achieve 96.50\% accuracy on average. Such results suggest that we should pay strong attention to the security problem of many AI applications, and further propose corresponding defensive strategies in the future.

Jinyin Chen, Mengmeng Su, Shijing Shen et al.

Most deep learning models are easily vulnerable to adversarial attacks. Various adversarial attacks are designed to evaluate the robustness of models and develop defense model. Currently, adversarial attacks are brought up to attack their own target model with their own evaluation metrics. And most of the black-box adversarial attack algorithms cannot achieve the expected success rate compared with white-box attacks. In this paper, comprehensive evaluation metrics are brought up for different adversarial attack methods. A novel perturbation optimized black-box adversarial attack based on genetic algorithm (POBA-GA) is proposed for achieving white-box comparable attack performances. Approximate optimal adversarial examples are evolved through evolutionary operations including initialization, selection, crossover and mutation. Fitness function is specifically designed to evaluate the example individual in both aspects of attack ability and perturbation control. Population diversity strategy is brought up in evolutionary process to promise the approximate optimal perturbations obtained. Comprehensive experiments are carried out to testify POBA-GA's performances. Both simulation and application results prove that our method is better than current state-of-art black-box attack methods in aspects of attack capability and perturbation control.

Jinyin Chen, Yangyang Wu, Lu Fan et al.

Recommender systems are becoming more and more important in our daily lives. However, traditional recommendation methods are challenged by data sparsity and efficiency, as the numbers of users, items, and interactions between the two in many real-world applications increase fast. In this work, we propose a novel clustering recommender system based on node2vec technology and rich information network, namely N2VSCDNNR, to solve these challenges. In particular, we use a bipartite network to construct the user-item network, and represent the interactions among users (or items) by the corresponding one-mode projection network. In order to alleviate the data sparsity problem, we enrich the network structure according to user and item categories, and construct the one-mode projection category network. Then, considering the data sparsity problem in the network, we employ node2vec to capture the complex latent relationships among users (or items) from the corresponding one-mode projection category network. Moreover, considering the dependency on parameter settings and information loss problem in clustering methods, we use a novel spectral clustering method, which is based on dynamic nearest-neighbors (DNN) and a novel automatically determining cluster number (ADCN) method that determines the cluster centers based on the normal distribution method, to cluster the users and items separately. After clustering, we propose the two-phase personalized recommendation to realize the personalized recommendation of items for each user. A series of experiments validate the outstanding performance of our N2VSCDNNR over several advanced embedding and side information based recommendation algorithms. Meanwhile, N2VSCDNNR seems to have lower time complexity than the baseline methods in online recommendations, indicating its potential to be widely applied in large-scale systems.

Jinyin Chen, Jian Zhang, Xuanheng Xu et al.

Predicting the potential relations between nodes in networks, known as link prediction, has long been a challenge in network science. However, most studies just focused on link prediction of static network, while real-world networks always evolve over time with the occurrence and vanishing of nodes and links. Dynamic network link prediction thus has been attracting more and more attention since it can better capture the evolution nature of networks, but still most algorithms fail to achieve satisfied prediction accuracy. Motivated by the excellent performance of Long Short-Term Memory (LSTM) in processing time series, in this paper, we propose a novel Encoder-LSTM-Decoder (E-LSTM-D) deep learning model to predict dynamic links end to end. It could handle long term prediction problems, and suits the networks of different scales with fine-tuned structure. To the best of our knowledge, it is the first time that LSTM, together with an encoder-decoder architecture, is applied to link prediction in dynamic networks. This new model is able to automatically learn structural and temporal features in a unified framework, which can predict the links that never appear in the network before. The extensive experiments show that our E-LSTM-D model significantly outperforms newly proposed dynamic network link prediction methods and obtain the state-of-the-art results.

Jinyin Chen, Haibin Zheng, Hui Xiong et al.

Machine learning models have been shown vulnerable to adversarial attacks launched by adversarial examples which are carefully crafted by attacker to defeat classifiers. Deep learning models cannot escape the attack either. Most of adversarial attack methods are focused on success rate or perturbations size, while we are more interested in the relationship between adversarial perturbation and the image itself. In this paper, we put forward a novel adversarial attack based on contour, named FineFool. Finefool not only has better attack performance compared with other state-of-art white-box attacks in aspect of higher attack success rate and smaller perturbation, but also capable of visualization the optimal adversarial perturbation via attention on object contour. To the best of our knowledge, Finefool is for the first time combines the critical feature of the original clean image with the optimal perturbations in a visible manner. Inspired by the correlations between adversarial perturbations and object contour, slighter perturbations is produced via focusing on object contour features, which is more imperceptible and difficult to be defended, especially network add-on defense methods with the trade-off between perturbations filtering and contour feature loss. Compared with existing state-of-art attacks, extensive experiments are conducted to show that Finefool is capable of efficient attack against defensive deep models.