Explanation-guided learning (EGL) has shown promise in aligning model predictions with interpretable reasoning, particularly in computer vision tasks. However, most approaches rely on external annotations or heuristic-based segmentation to supervise model explanations, which can be noisy, imprecise and difficult to scale. In this work, we provide both empirical and theoretical evidence that low-quality supervision signals can degrade model performance rather than improve it. In response, we propose ALIGN, a novel framework that jointly trains a classifier and a masker in an iterative manner. The masker learns to produce soft, task-relevant masks that highlight informative regions, while the classifier is optimized for both prediction accuracy and alignment between its saliency maps and the learned masks. By leveraging high-quality masks as guidance, ALIGN improves both interpretability and generalizability, showing its superiority across various settings. Experiments on the two domain generalization benchmarks, VLCS and Terra Incognita, show that ALIGN consistently outperforms six strong baselines in both in-distribution and out-of-distribution settings. Besides, ALIGN also yields superior explanation quality concerning sufficiency and comprehensiveness, highlighting its effectiveness in producing accurate and interpretable models.

Deep neural networks (DNNs) have achieved remarkable performance in many tasks, yet they often behave as opaque black boxes. Explanation-guided learning (EGL) methods steer DNNs using human-provided explanations or supervision on model attributions. These approaches improve interpretability but typically assume benign inputs and incur heavy annotation costs. In contrast, both predictions and saliency maps of DNNs could dramatically alter facing imperceptible perturbations or unseen patterns. Adversarial training (AT) can substantially improve robustness, but it does not guarantee that model decisions rely on semantically meaningful features. In response, we propose Explanation-Guided Adversarial Training (EGAT), a unified framework that integrates the strength of AT and EGL to simultaneously improve prediction performance, robustness, and explanation quality. EGAT generates adversarial examples on the fly while imposing explanation-based constraints on the model. By jointly optimizing classification performance, adversarial robustness, and attributional stability, EGAT is not only more resistant to unexpected cases, including adversarial attacks and out-of-distribution (OOD) scenarios, but also offer human-interpretable justifications for the decisions. We further formalize EGAT within the Probably Approximately Correct learning framework, demonstrating theoretically that it yields more stable predictions under unexpected situations compared to standard AT. Empirical evaluations on OOD benchmark datasets show that EGAT consistently outperforms competitive baselines in both clean accuracy and adversarial accuracy (+37%) while producing more semantically meaningful explanations, and requiring only a limited increase (+16%) in training time.

Website Fingerprinting (WF) attacks, a methodology that allows observers to infer visited websites through encrypted traffic analysis, are usually employed for evaluating the security of anonymous networks like Tor. Although deep learning-based methods have demonstrated significant success in recent studies, existing approaches exhibit limitations. They are insufficient in modeling the intricate characteristics of traffic patterns, which leads to compromised accuracy in identifying challenging samples. To address these limitations, we propose CTF, a novel WF attack comprising three key components: (1) extraction of Hybrid Multi-scale Traffic Features (HMTF) integrating spatio-temporal information, (2) a novel deep learning framework that synergizes CNN for spatial pattern recognition and Transformer for temporal dependency modeling within HMTF, and (3) a supervised contrastive learning mechanism to enhance discriminative capability across website fingerprints. Experimental evaluations across multiple benchmark datasets reveal that CTF achieves 99.09% accuracy in Closed-World scenarios, showing enhanced robustness against four SOTA WF defense mechanisms.

With the rapid advancement of diagnostic technology, the ability to detect pathological areas such as tumors and polyps has significantly improved. This progress provides medical imaging specialists with more precise visual information to support anomaly identification, diagnosis, treatment planning, and patient monitoring. However, existing unsupervised and semi-supervised anomaly detection methods struggle with data privacy constraints, limited annotated medical datasets, and challenges in generalization. Zero-Shot Anomaly Detection (ZSAD), which enables the detection of unseen categories without requiring class-specific training, has emerged as a promising solution by leveraging the vision-language alignment capabilities of Vision-Language Models (VLMs), such as Contrastive Language-Image Pretraining (CLIP). Despite recent progress, ZSAD remains hindered by high noise levels, sparse targets, and poor adaptability in complex medical imaging scenarios. To address these issues, we propose a novel framework: DiffusionCLIP, a diffusion-based VLM for zero-shot anomaly detection in two-dimensional medical images. Specifically, DiffusionCLIP integrates diffusion models into the VLM to progressively denoise multi-level features extracted from the CLIP visual encoder, enhancing feature robustness and discriminability. A multi-level feature fusion strategy is designed to aggregate multi-scale representations from different depths of the visual encoder, ensuring complementary semantic alignment across layers. In addition, a dynamically modulated weight loss function is introduced to adaptively balance the learning of hard and easy samples, further improving model generalization. Extensive experiments on multiple benchmark medical imaging datasets, demonstrate that the proposed method significantly outperforms existing zero-shot anomaly detection approaches in terms of accuracy, robustness, and generalization.

Malware has been researched hot off the press in cyber security for a long time. With the rise of machine learning algorithms, many research works attempt to apply machine learning-based methods in malware classification. However, existing machine learning-based malware classification methods rely on many various features extracted from malware samples, which may make their system lose processing speed and generality between different operational environments. These methods can not cope with massive malware samples. To improve generality and speed of classification system, we proposed a new model to classify malware with function call graphs (FCGs) extracted from their assembly code. According to previous studies, FCG is a generic feature and it is stable against metamorphic malware. Moreover, FCG extraction is not a time-consuming process. We select DGCNN (Deep Graph Convolutional Neural Network) to embed structural information inherent in FCGs for malware classification. It can make the best of the structure information stored in FCGs and make the results more convincing and accurate compared with other methods using traditional features. We use two large datasets from different operational environments containing nearly 20K malware samples to evaluate our proposed model. The experimental results show that it can classify malware represented as FCG with satisfactory accuracy and faster processing speed.

Inspired by the success of graph neural network in graph data classification, graph neural networks have been widely used in malware classification and they have been proven to be the state-of-the-art malware classification models. However, most of existing adversarial samples generation techniques against machine learning-based malware classification models modify malware samples by inserting dead codes or modifying binaries directly, which is less effective against graph neural network-based malware classification models. In this paper, we propose an adversarial attack framework powered by reinforcement learning to spoof the deep graph convolutional neural network (DGCNN)-based malware classifiers called Intelligent Malware Evader (IMaler). We construct functionality-preserved manipulations based on traditional obfuscation techniques that can modify both node features and structural features of malware. The reinforcement learning agent can make optimal decisions on how to obfuscate malware with functionality-preserved manipulations. We use a large dataset with more than 10,000 samples to evaluate the performance of IMaler and use a random agent attack as a baseline attack. The experiment results show that IMaler can achieve a significantly higher evasion rate (88.26%) than the random agent attack with fewer query times.

This paper uses game theory, analyses the relevant enterprises and the government's different motives and the behavior of game in the process of the ecological industrial chain forming, then conceives the incomplete information dynamic grey game model and identifies the optimal strategic combination between the government and enterprises. Based on the analysis of the game, get the government and the market is two important forces to promote the ecological industrial chain formed. Game balance results show that the government supports eco-oriented enterprises, and enterprises adopt eco-development model can achieve a win-win business with the results between government and enterprises. The results also show the importance that government departments strengthen supervision and management.

The effect of mobility on the critical transmitting range in wireless ad hoc networks is discussed in this paper. A new parameter, pause probability P/sub pause/, which is the function of the minimum velocity V/sub min/, the maximum velocity V/sub max/ and the pause time t/sub pause/ of random waypoint (RWP) mobility model, is defined. Simulation results show that the effect of RWP parameter on the critical transmitting range is determined by P/sub pause/ that is, if P/sub pause/ smaller than 0.5, the critical transmitting range of mobile network is about 10 percent larger than that of stationary one, else it equals that of stationary network approximately. If the values of t/sub pause/, V/sub min/ and V/sub max/ are given, their effect on the critical transmitting range is determined since P/sub pause/ is the function of them.

A multiple access control (MAC) protocol - user adaptive scheduling multiple access (UASMA) protocol is proposed in this paper, which can perfectly schedule the packet transmission according to the exact number of active mobile terminals (MTs) determined by self-organizing algorithm, and adjust the number of packets sending by one node in one frame properly. UASMA protocol employs an especial frame structure, which makes it possible to allocate channel resource to uplink and downlink traffic according o their respective service requirements. And the concept of referenced-frame-length is proposed, which can ensure the frame length various in a certain range, and consequently increase the utilization of the channel greatly. Meanwhile, UASMA protocol uses an efficient collision resolution algorithm to guarantee that active MTs can access the channel rapidly. Finally, the performance of UASMA protocol is evaluated by simulation and compared with UPMA protocol, the simulation results show that the former has better performance than the latter.

The IEEE 802.11 multiple access control protocol was modified for use in multi-channel, multi-hop ad hoc network, through the use of a new channel-status indicator. In particular, we have evaluated the improvement due to the multi-channel use. We report in this paper on the results of the throughput per node and the end-to-end delay for the modified IEEE 802.11 protocol for different network sizes. Using these results, we were able to propose a number of throughput scaling laws. Our simulation results show that the throughputs per node with multiple channels for the line and the grid ad hoc network topologies will increase by 47.89%, and by 1.39-163%, respectively, for networks with 16 to 64 nodes, as compared with that of single channel.