Cross-site scripting (XSS) attacks pose a significant threat to web applications and user privacy, with the number of such attacks rapidly increasing. Although existing machine learning and deep learning-based XSS attack detection models are effective against common XSS attacks, these models all overlook their own security and often fail to defend against adversarial samples that exploit model vulnerabilities, allowing attackers to successfully bypass these models by using XSS adversarial samples. To address this challenge, in this paper, we propose a novel XSS attack detection model based on two-stage Abstract Syntax Tree (AST) analysis and Long Short-Term Memory (LSTM) neural networks, effectively mitigating the impact of adversarial samples. Our model leverages the ability of AST parsing and analysis of HTML and JavaScript code to effectively eliminate redundant information and adversarial perturbations introduced by adversarial samples. The two-stage process first extracts JavaScript code from the HTML AST, then identifies malicious code fragments from the JavaScript AST. Finally, the LSTM neural network is trained to classify samples as malicious or benign. By analyzing the HTML and JavaScript components of web pages, our model identifies and eliminates adversarial perturbations that interfere with detection, significantly enhancing the security and reliability of the detection process. Extensive experiments on real datasets demonstrate our model's superior performance, achieving an accuracy rate of 0.991 and an F1 score of 0.998 against standard XSS samples, outperforming existing models. More importantly, when facing adversarial XSS samples, most existing detection models exhibit severe robustness degradation with the detection rate (DR) below 0.880, whereas our model maintains a detection rate of over 0.982, significantly higher than state-of-the-art models and demonstrating its significant effectiveness in defending against XSS adversarial attacks.

DDoS (Distributed Denial of Service) attacks are increasingly becoming a major threat in the field of cybersecurity. They overwhelm target servers by sending large-scale requests from multiple locations, causing the servers to become unresponsive. The distributed nature of DDoS attacks also makes detection and defense even more challenging. As the damage caused by such attacks grows, the development of efficient detection and mitigation mechanisms has become an urgent priority. While traditional machine learning methods are useful, they still require manual feature extraction, which not only involves significant human intervention, but also is time-consuming. In contrast, deep learning offers an automated approach to feature extraction and can learn more abstract patterns, which leads to improved detection performance. Therefore, this paper reviews and analyzes existing deep learning methods for DDoS attack detection. We provide a comprehensive analysis of the various types of DDoS attacks and explore different deep learning models employed for attack detection. Additionally, we explore techniques such as federated learning that can be integrated with deep learning, and analyze their related literature in DDoS attack detection. Finally, we specify future research directions on DDoS attack detection using deep learning.

The growing adoption of 3D point cloud in applications like autonomous driving has heightened concerns about their vulnerability to adversarial attacks. Existing defense methods face two fundamental challenges: ineffective detection of imperceptible adversarial examples and poor restoration of severely distorted point cloud. In this paper, we present ADDR, an end-to-end defense framework that integratesBinary Geometric Feature Anomaly Detection (BGFAD)andDistorted point cloud Restoration (DPCR). BGFAD employs a dual threshold mechanism combining global distance statistics and local curvature analysis to detect both substantial and imperceptible adversarial perturbations. DPCR leverages attention enhanced feature encoding to reconstruct missing geometric structures while preserving semantic integrity through bidirectional Chamfer loss optimization. Our framework uniquely bridges traditional geometric priors with deep learning mechanisms, achieving attack-agnostic defense without classifier retraining. Extensive experiments on ModelNet40, ShapeNet and ScanObjectNN datasets demonstrate state-of-the-art performance, with about 12% higher robustness against structural attacks and 6× better restoration fidelity than existing methods. ADDR maintains real-time processing capabilities while reducing adversarial success rates to <5% across diverse attacks. The code is available at https://github.com/whwh456/ADDR.

Federated learning (FL) can complete collaborative model training without transferring local data, which can greatly improve the training efficiency. However, FL is susceptible data and model backdoor attacks. To address data backdoor attack, in this article, we propose a defense method named TSF. TSF transforms data from time domain to frequency domain and subsequently designs a low-pass filter to mitigate the impact of high-frequency signals introduced by backdoor samples. Additionally, we undergo homomorphic encryption on local updates to prevent the server from inferring user’s data. We also introduce a defense method against model backdoor attack named ciphertext field similarity detect differential privacy (CFSD-DP). CFSD-DP screens malicious updates using cosine similarity detection in the ciphertext domain. It perturbs the global model using differential privacy mechanism to mitigate the impact of model backdoor attack. It can effectively detect malicious updates and safeguard the privacy of the global model. Experimental results show that the proposed TSF and CFSD-DP have 73.8% degradation in backdoor accuracy while only 3% impact on the main task accuracy compared with state-of-the-art schemes. Code is available athttps://github.com/whwh456/TSF.

Model inversion attacks (MIAs) aim to recover private data from inaccessible training sets of deep learning models, posing a privacy threat. MIAs primarily focus on the white-box scenario where attackers have full access to the model’s structure and parameters. However, practical applications are usually in black-box scenarios or label-only scenarios, i.e., the attackers can only obtain the output confidence vectors or labels by accessing the model. Therefore, the attack models in existing MIAs are difficult to effectively train with the knowledge of the target model, resulting in sub-optimal attacks. To the best of our knowledge, we pioneer the research of a powerful and practical attack model in the label-only scenario. In this paper, we develop a novel MIA method, leveraging a conditional diffusion model (CDM) to recover representative samples under the target label from the training set. Two techniques are introduced: selecting an auxiliary dataset relevant to the target model task and using predicted labels as conditions to guide training CDM; and inputting target label, pre-defined guidance strength, and random noise into the trained attack model to generate and correct multiple results for final selection. This method is evaluated using Learned Perceptual Image Patch Similarity as a new metric and as a judgment basis for deciding the values of hyper-parameters. Experimental results show that this method can generate similar and accurate samples to the target label, outperforming generators of previous approaches.

In recent years, Bitcoin has become one of the most popular cryptocurrencies. The most significant mechanism of Bitcoin is PoW (Proof-of-Work), but it also brings opportunities for mining attacks. In our last study, we proposed a Selfish Mining-based Denial-of-Service Attack (SDoS), which can cause serious threats to the Bitcoin system. On this basis, we further put forward three greedier SDoS attack strategies: a competitive greedy SDoS attack strategy ESDoS, a trail greedy SDoS attack strategy TSDoS, a hybrid greedy SDoS attack strategy ETSDoS, and a more public SDoS attack strategy PSDoS. Besides, we also study the adversary’s optimal strategies under different conditions. The experimental results show that if the adversary adopts the SDoS optimal strategy, his revenue increase rate will be further improved and significantly higher than the other existing mining attacks. If the adversary masters 14% of the total mining power, he has a chance to improve his revenue (25% in Selfish Mining, 19.6% in SDoS), and if the adversary masters 15% of the total mining power, he is capable of launching a 51% attack.

Privacy-preserving record linkage (PPRL) aims to link records of the same real-world entity from different databases without exposing any private information about the entity. Bloom filters are widely used in PPRL due to their effectiveness in encoding records while enabling fast approximate linkage in the case of attribute value errors and changes. However, the basic Bloom filters used for PPRL can be subject to cryptanalysis attacks that expose the plain-text values encoded in them. Recent studies have successfully attacked some improved Bloom filter encodings in PPRL but require specific conditions or knowledge of various encoding parameters to obtain high accuracy. This paper presents a novel attack based on differential analysis against Bloom filters used for PPRL. The attack exploits graphs to model the relationship between attribute value variation and the difference between Bloom filters. Then, features are generated for the node in graphs according to a clustering algorithm that we propose. Thus, we can match nodes with similar features to re-identify encoded records. Experiments on two real-world databases show that even with improved Bloom filter encoding and some hardening techniques, our attack can re-identify private information from encoded records with high accuracy and require less priori knowledge.

The rapid growth of GPS technology and mobile devices has led to a massive accumulation of location data, bringing considerable benefits to individuals and society. One of the major usages of such data is travel time prediction, a typical service provided by GPS navigation devices and apps. Meanwhile, the constant collection and analysis of the individual location data also pose unprecedented privacy threats. We leverage the notion of geo-indistinguishability, an extension of differential privacy to the location privacy setting, and propose a procedure for privacy-preserving travel time prediction without collecting actual individual GPS trace data. We propose new concepts to examine the impact of geo-indistinguishability-based sanitization on the usefulness of GPS traces and provide analytical and experimental utility analysis for privacy-preserving travel time prediction. We also propose new metrics to measure the adversary error in learning individual GPS traces from the collected sanitized data. Our experiment results suggest that the proposed procedure provides travel time prediction with satisfactory accuracy at reasonably small privacy costs.

In this paper, we focus on mining attacks targeting the Proof of Work (PoW) consensus mechanism in blockchain-based systems. Specifically, we model mining as a game and propose a mining attack – the Selfish mining-based denial of service (SDoS) attack. By studying the choices (mining or stopping) of honest miners under the attack and the adversary’s revenue, we demonstrate that selfish mining is incentive-compatible with game-level denial of service attack, and that SDoS can be more threatening than existing mining attacks. Even under the worst assumption, the adversary only needs to master more than 19.6% of the total mining power to increase the revenue, and can launch a 51% attack with much less than 50%. In addition, we show that honest miners may make decisions based on the overall or current utility, and choosing the current utility is more beneficial to the adversary.

Non-negative Matrix Factorization (NMF), a classical method for dimensionality reduction, has been applied in many fields. It is based on the idea that negative numbers are physically meaningless in various data-processing tasks. Apart from its contribution to conventional data analysis, the recent overwhelming interest in NMF is due to its newly discovered ability to solve challenging data mining and machine learning problems, especially in relation to gene expression data. This survey paper mainly focuses on research examining the application of NMF to identify differentially expressed genes and to cluster samples, and the main NMF models, properties, principles, and algorithms with its various generalizations, extensions, and modifications are summarized. The experimental results demonstrate the performance of the various NMF algorithms in identifying differentially expressed genes and clustering samples.

With the development of deep sequencing technologies, many RNA-Seq data have been generated. Researchers have proposed many methods based on the sparse theory to identify the differentially expressed genes from these data. In order to improve the performance of sparse principal component analysis, in this paper, we propose a novel class-information-based sparse component analysis (CISCA) method which introduces the class information via a total scatter matrix. First, CISCA normalizes the RNA-Seq data by using a Poisson model to obtain their differential sections. Second, the total scatter matrix is gotten by combining the between-class and within-class scatter matrices. Third, we decompose the total scatter matrix by using singular value decomposition and construct a new data matrix by using singular values and left singular vectors. Then, aiming at obtaining sparse components, CISCA decomposes the constructed data matrix by solving an optimization problem with sparse constraints on loading vectors. Finally, the differentially expressed genes are identified by using the sparse loading vectors. The results on simulation and real RNA-Seq data demonstrate that our method is effective and suitable for analyzing these data.

Many methods have been considered for gene selection and analysis of gene expression data. Nonetheless, there still exists the considerable space for improving the explicitness and reliability of gene selection. To this end, this paper proposes a novel method named robust graph regularized non-negative matrix factorization for characteristic gene selection using gene expression data, which mainly contains two aspects: Firstly, enforcing L21-norm minimization on error function which is robust to outliers and noises in data points. Secondly, it considers that the samples lie in low-dimensional manifold which embeds in a high-dimensional ambient space, and reveals the data geometric structure embedded in the original data. To demonstrate the validity of the proposed method, we apply it to gene expression data sets involving various human normal and tumor tissue samples and the results demonstrate that the method is effective and feasible.