Channel State Information (CSI)-based Human Activity Recognition (HAR) has emerged as a promising Wi-Fi sensing technology. Particularly in the power grids, ubiquitous sensing is crucial for monitoring personnel safety and terminal states. However, given the heterogeneous nature of communication protocol in real‑world infrastructure (e.g., power grids), the real‑world applicability of CSI‑based HAR is severely hampered by a fundamental limitation: mainstream deep learning methods are rigidly coupled to specific channel bandwidth and its corresponding subcarrier layouts. This lack of downward compatibility prevents models trained on 160 MHz Wi-Fi protocols from being deployed in environments using legacy 20 MHz protocols, despite the latter’s subcarriers being fully contained within the former. In this paper, we address this by learning bandwidth-invariant representations from CSI. Our key insight is that CSI measurements from disparate bandwidths are not disparate signals but rather complementary views of the same underlying activity-induced channel frequency response. We introduce Uni-Fi, a contrastive learning framework that pulls these views together in an embedding space, driving the model to learn shared frequency-domain activity features while being robust to specific bandwidth configuration. Extensive evaluations demonstrate that our framework empowers existing CSI-based HAR models with exceptional cross-bandwidth generalization: a model trained solely on 160 MHz CSI maintains high accuracy in a 20 MHz setting, where standard methods fail completely.

Personalized Federated Learning (PFL) enables client-specific models to address data heterogeneity but suffers from high communication overhead and unstable participation in resource-constrained and self-interested environments. Existing mainstream approaches predominantly prioritize training process efficiency but lack explicit consideration of incentive mechanisms and rational client behaviors, which may lead to clients behaving conservatively, reducing participation and limiting the effectiveness and scalability of PFL in real-world deployments. In this paper, we proposeIncenPNS, the first incentive-driven adaptive framework that jointly optimizes model pruning, neighbor selection, and incentive mechanisms for communication-efficient PFL.IncenPNSformulates the joint design as a unified optimization objective that balances personalization performance, communication efficiency, and incentive utility under dynamic and heterogeneous environments. To solve the resulting coupled and high-dimensional decision problem, we develop a Multi-Agent Soft Actor-Critic (MASAC)-based learning algorithm that enables clients to adapt pruning rates and collaboration decisions through online interaction. Moreover, a budget-balanced incentive mechanism is incorporated to operate under partial observability, aligning individual rationality with system-level objectives and ensuring reliable participation of self-interested clients. Extensive experiments on representative benchmarks show thatIncenPNSachieves up to 34.8% reduction in communication cost, 68.8% faster convergence, and 83.3% improvement in personalization accuracy compared with state-of-the-art baselines.

3D point cloud registration, which seeks the optimal rigid transformation to align two point clouds, is a fundamental task in autonomous systems. However, the 3D correspondences between point clouds are prone to substantial outliers (mismatches), leading to significant decreases in registration accuracy. Existing outlier-robust registration methods commonly have high computational complexity and, hence, are limited in time-sensitive applications. Inertial measurement unit (IMU) sensors are widespread in modern autonomous systems and can offer precise gravity directions. Accordingly, we propose a highly efficient voting-based outlier removal method by leveraging the gravity prior in this paper. This pre-processing step can significantly reduce the candidate correspondence set for subsequent estimation, thus accelerating robust point cloud registration. We then leverage pairwise invariant features to decompose the optimization of rotation and translation. Further, we propose a two-stage consensus maximization solver to optimize the rotation and translation sequentially, leading to deterministic and robust registration. Extensive experiments on both synthetic and real-world datasets indicate that our method effectively boosts registration efficiency while exhibiting comparable robustness to state-of-the-art methods.

The diffusion policy has introduced the excellent multimodal data modeling capability of the diffusion model to robotic imitation learning. However, extensive expert demonstration data is still needed during training, requiring considerable human effort to collect. Natural human action data is cheaper and easier to obtain than expert demonstration data, but its use in direct robot policy training is challenging due to the distribution gap. In this study, we propose the double diffusion policy learning paradigm, which incorporates low-cost human action data to diminish the reliance of diffusion policy on extensive expert demonstration data. Specifically, we extract human intentions from the diffusion policy modeling human actions, then integrate these intentions into training the diffusion policy for generating robot actions. This processing guides the policy model to train better on small expert demonstration datasets. Experiments show that our double diffusion policy outperforms the vanilla diffusion policy and other state-of-the-art imitation learning algorithms with limited expert demonstration data.

Edge computing plays a critical role in supporting low-latency IoT applications, yet the susceptibility of edge nodes to faults can disrupt services and degrade Quality of Service (QoS). Fault prediction offers a proactive solution by identifying potential failures through spatio-temporal feature learning from historical observations. However, existing spatio-temporal prediction models are typically designed for fixed network topologies with predefined input-output structures, which limits their effectiveness in dynamic edge networks where nodes are frequently added or removed. Adapting these models to topology variations often requires full retraining or architectural redesign, resulting in substantial computational overhead and limited real-time applicability. To overcome these limitations, this paper proposes ASTFNet, an adaptive spatio-temporal fault prediction framework for dynamic edge networks. The framework integrates a spatio-temporal fault prediction model that incorporates node identity embeddings to enable flexible representation learning under evolving topologies and an adaptive fine-tuning mechanism that detects topology changes and performs targeted model updates without full retraining. Experiments on real-world datasets demonstrate that ASTFNet significantly reduces retraining time while maintaining high prediction accuracy and achieves robust performance under dynamic node additions and removals.

With the rapid development of cloud computing, network architectures are moving towards distributed computing, which performs data processing at edge nodes to reduce latency, enabling more efficient and scalable network services. Nevertheless, this shift introduces significant security challenges due to the heterogeneity of communications protocols and the vulnerabilities of edge devices. To effectively secure these distributed networks, it is essential to perform multiple traffic analysis tasks, e.g. Network Intrusion Detection, Encrypted Traffic Classification, and Application Traffic Classification. However, existing methods have limited generic feature extraction and require the deployment of multiple models to solve multiple tasks, which exceeds the resource capacity of edge nodes. To address these challenges, we introduce a Lightweight Multitask Traffic Analysis Framework LiMTa, which novelly proposes a traffic pre-training method, FreqRec, and a lightweight multi-task model fine-tune method, MT-Adapter. FreqRec enables high-level semantic feature extraction by reconstructing the frequency features of traffic samples, and MT-Adapter efficiently performs multiple tasks by computing the pre-trained model only once. Experimental results demonstrate that our approach achieves state-of-the-art (SOTA) performance on six traffic analysis tasks. Moreover, the MT-Adapter module only fine-tunes a small number of parameters, accounting for only 6.37% of the pre-trained model’s parameters, and achieves the same result as the full fine-tuning. Compared to full fine-tuning, LiMTa reduces the time cost by 50.9% and the space cost by 57.4% in six edge traffic analysis tasks.

The combination of self-supervised learning and adversarial training (AT) can significantly improve the adversarial robustness of self-supervised models. However, the robustness of self-supervised adversarial training (self-AT) still lags behind that of state-of-the-art (SOTA) supervised AT (sup-AT), even though the performance of current self-supervised learning models has already matched or even surpassed that of SOTA supervised learning models. This issue raises concerns about the secure application of self-supervised learning models. The inclusion of adversarial training turns self-AT into a challenging joint optimization problem, and recent studies have shown that the data augmentation methods necessary for constructing positive pairs in self-supervised learning negatively impact the robustness improvement in self-AT. Inspired by this, we propose 3SAT, a simple self-supervised adversarial training framework. 3SAT conducts adversarial training on original, unaugmented samples, reducing the difficulty of optimizing the adversarial training subproblem and fundamentally eliminating the negative impact of data augmentation on robustness improvement. Additionally, 3SAT introduces a dynamic training objective scheduling strategy to address the issue of model training collapse during the joint optimization process when using original samples directly. 3SAT is not only structurally simple and computationally efficient, reducing self-AT training time by half, but it also improves the SOTA self-AT robustness accuracy by 16.19\% and standard accuracy by 11.41\% under Auto-Attack on the CIFAR-10 dataset. Even more impressively, 3SAT surpasses the SOTA sup-AT method in robust accuracy by a significant margin of 11.25\%. This marks the first time that self-AT has outperformed SOTA sup-AT in robustness, indicating that self-AT is a superior method for improving model robustness.

The growing flexibility of 5G network architectures increases the risk of network faults. Such network fault types are diverse and variability, as 5G networks have been integrated in various vertical industries, e.g., power grids, resulting in these faults being widespread and challenging to diagnose. To reduce the costs associated with fault identification and remediation, we introduce Rsm-RCA as a novel framework for automated root cause analysis (RCA) in 5G networks. By training on massive amounts of data, Rsm-RCA is capable of identifying fault types from the fault signalling messages collected. Specifically, this framework efficiently collects and processes radio resource control (RRC) fault signalling messages from commercial networks to extract multidimensional attributes and KPI parameters. The processed signalling messages are then fed into a decision tree model, which enables accurate fault classification with minimal time expenditure after training. For power utilities private 5G networks, fast and accurate RCA is highly beneficial. Experimental results demonstrate that Rsm-RCacan classify faults in an extremely short time, achieving an average accuracy of 99 %.

The growing complexity and diversity of 5G network architecture (e.g., power grid 5G network) have made security risk assessment and root cause analysis increasingly challenging. Recent advances in large language models (LLMs) have the potential to transform this landscape. However, existing LLMs-based solutions primarily focus on understanding the language of 5G telecommunications, while overlooking potential security vulnerabilities in the data flows. To facilitate LLMs' in-depth application, this paper presents RCA-LLM, a novel fault root cause analysis framework for 5G networks developed from tailored LLMs-based solutions. In explicit terms, RCA-LLM is trained by inputting processed and organized fault information for fine-tuning, and combined with retrieval-augmented generation (RAG) technology to significantly improve the accuracy of 5G fault analysis. Our experimental results indicate that RCA-LLM performs well in fault analysis, effectively supporting users in diagnosing and resolving fault issues. Model evaluation results further demonstrate that the model significantly improves fault analysis accuracy and has high practical value. In addition, RCA-LLM provides important reference value for efficient operation and maintenance management of 5G and future power grid networks, while also offering new ideas for advancing intelligent fault analysis.

Self-supervised learning (SSL) is a technique used to learn feature representations from unlabeled data. However, existing SSL frameworks either rely too heavily on domain knowledge due to their design based on feature invariance, leading to a lack of domain transferability, or they are based on autoencoder designs, which generate features with redundant low-level semantics, resulting in suboptimal model representations. In this work, we introduce a novel Domain Agnostic Self-Supervised Learning framework called DASSL, which learns superior high-level feature representations of samples by reconstructing the samples’ missing information in the representation space. DASSL does not require additional domain priors, and compared to successful SSL methods, DASSL achieves competitive representation quality. Moreover, when DASSL incorporates domain-related data augmentation techniques, it outperforms successful methods across multiple datasets and evaluation protocols.

The emergence of Satellite-Terrestrial Integrated Networks (STIN) has significantly expanded terrestrial network coverage but introduced new security threats. Current Intrusion Detection Systems (IDSs) for STIN mostly consider the distributed nature of satellites, overlooking the computational limits of single satellites and the effect of satellite mobility on IDS generalization, where accuracy and adaptability may drop in dynamic environments. To address this, we propose an unsupervised IDS for STIN based on Federated Learning (FL) named STIN-IDS. We deploy IDS in a cross-layer distributed manner, distributing data processing tasks across multiple Low Earth Orbit (LEO) satellites, while Geostationary Earth Orbit (GEO) satellites act as FL clients responsible for anomaly detection, thereby alleviating the computational load on single satellites. Furthermore, to address changes in user regions and traffic patterns due to satellite mobility, FL clients use dynamic data from different LEO regions for collaborative training, improving adaptability to dynamic environments. Experiments across four datasets with varying network conditions show that STIN-IDS achieves strong generalization and outperforms similar methods.

Satellite-based cloud computing cybersecurity threats have long posed significant challenges, particularly for cloud infrastructure operators.While prior research has partially addressed these issues by mitigating threats and enhancing human response efficiency, this paper proposes a novel AI-Driven Threat Analysis and Response (TAR) framework.The study progresses in three main phases: (1) redefining urgent threats through a novel formula; (2) implementing a triage and analysis framework using augmented Large Language Models (LLMs); and (3) automating incident response via a Security Orchestration, Automation, and Response (SOAR) platform.Our prototype, tested in a simulated public cloud environments using real production threats, demonstrated a 17% improvement in handling low-and medium-urgency threats.Experimental results show our approach achieves 97.8% coverage in automatic threat classification, significantly outperforming traditional manual methods, which achieve 77.8% coverage.With high recall and precision in managing low-and medium-urgency threats, our method enhances manual efficiency through SOAR-enabled automation.Furthermore, * Corresponding Author.our augmented method surpasses the state-of-the-art GPT-4 Turbo model in addressing security threats containing Chinese characters.

Recently, Deep Learning (DL) models have been increasingly deployed on end-user devices as On-Device AI, offering improved efficiency and privacy. However, this deployment trend poses more serious Intellectual Property (IP) risks, as models are distributed on numerous local devices, making them vulnerable to theft and redistribution. Most existing ownership protection solutions (e.g., backdoor-based watermarking) are designed for cloud-based AI-as-a-Service (AIaaS) and are not directly applicable to large-scale distribution scenarios, where each user-specific model instance must carry a unique watermark. These methods typically embed a fixed watermark, and modifying the embedded watermark requires retraining the model. To address these challenges, we propose Hot-Swap MarkBoard, an efficient watermarking method. It encodes user-specific n-bit binary signatures by independently embedding multiple watermarks into a multi-branch Low-Rank Adaptation (LoRA) module, enabling efficient watermark customization without retraining through branch swapping. A parameter obfuscation mechanism further entangles the watermark weights with those of the base model, preventing removal without degrading model performance. The method supports black-box verification and is compatible with various model architectures and DL tasks, including classification, image generation, and text generation. Extensive experiments across three types of tasks and six backbone models demonstrate our method's superior efficiency and adaptability compared to existing approaches, achieving 100% verification accuracy.

Supervised quantum learning methods face notable limitations in dynamic, real-world environments due to their reliance on static labels and limited adaptability. To address these challenges, we propose a novel online learning framework — Quantum Contextual Bandit (QCB) — that integrates quantum neural networks (QNNs) with contextual bandit (CB) algorithms. The QCB framework enables adaptive decision-making by incorporating bandit-based exploration into QNN training, making it particularly suitable for applications such as recommender systems. To mitigate the adverse effects of quantum noise—including depolarizing, Pauli, and shot noise, the framework leverages a gradient-free optimization approach, enhancing robustness and convergence stability. Experimental results on various datasets demonstrate that QCB consistently outperforms traditional QNN training methods with identical circuit architectures. Notably, the model achieves over 99% accuracy under ideal conditions and sustains high performance under noisy quantum environments. These results underscore the potential of QCB as a scalable, noise-resilient solution for adaptive learning in quantum machine learning systems.

The signaling collected in mobile communication networks can intuitively display the operational status of the system, which can use to locate faults. This paper proposes a novel signaling-based end-to-end fine-grained troubleshooting (simFGT) method for 6G NTN networks. First, the signaling collected from the core network is analyzed to extract multidimensional KPIs and attribute information. Second, a root cause localization algorithm is employed for fine-grained fault localization. Third, a lightweight data-driven ensemble learning method is adopted, with the abnormal KPI of the localized root cause node as inputs, and precise fault classification is achieved through data-driven weight optimization. Experiments results show that the proposed lightweight SimFGT method achieves best balance between precision and recall, resulting in highest F1 score, outperforming current state-of-the-art solutions.

Medical artificial intelligence (AI) holds significant potential in identifying signs of health conditions in nasal endoscopic images, thereby accelerating the diagnosis of diseases and systemic disorders. However, the performance of AI models heavily relies on expert annotations, and these models are usually task-specific with limited generalization performance across various clinical applications. In this paper, we introduce NasVLM, a Nasal Vision-Language foundation Model designed to extract universal representations from unlabeled nasal endoscopic data. Additionally, we construct a large-scale nasal endoscopic pre-training dataset and three downstream validation datasets from routine diagnostic records. The core strength of NasVLM lies in its ability to learn cross-modal semantic representations and perform multi-granular report-image alignment without depending on expert annotations. Furthermore, to the best of our knowledge, it is the first medical foundation model that effectively aligns medical report with multiple images of different anatomic regions, facilitated by a well-designed hierarchical report-supervised learning framework. The experimental results demonstrate that NasVLM has superior generalization performance across diverse diagnostic tasks and surpasses state-of-the-art self- and report-supervised methods in disease classification and lesion localization, especially in scenarios requiring label-efficient fine-tuning. For instance, NasVLM can distinguish normal nasopharynx (NOR) from abnormalities (benign hyperplasia, BH, and nasopharyngeal carcinoma, NPC) with an accuracy of 91.38% (95% CI, 90.59 to 92.17) and differentiate NPC from BH and NOR with an accuracy of 81.45% (95% CI, 80.21 to 82.67) on the multi-center NPC-Screen dataset using only 1% labeled data, on par with the performance of traditional supervised methods using 100% labeled data.

Training robots by model-free deep reinforcement learning (DRL) to carry out robotic manipulation tasks without sufficient successful experiences is challenging. Hindsight experience replay (HER) is introduced to enable DRL agents to learn from failure experiences. However, the HER-enabled model-free DRL still suffers from limited training performance due to its uniform sampling strategy and scarcity of reward information in the task environment. Inspired by the progress incentive mechanism in human psychology, we propose Progress Intrinsic Motivation-based HER (P-HER) in this work to overcome these difficulties. First, the Trajectory Progress-based Prioritized Experience Replay (TPPER) module is developed to prioritize sampling valuable trajectory data thereby achieving more efficient training. Second, the Progress Intrinsic Reward (PIR) module is introduced in agent training to add extra intrinsic rewards for encouraging the agents throughout the exploration of task space. Experiments in challenging robotic manipulation tasks demonstrate that our P-HER method outperforms original HER and state-of-the-art HER-based methods in training performance. Our code of P-HER and its experimental videos in both virtual and real environments are available athttps://github.com/weixiang-smart/P-HER. Note to Practitioners—This work is motivated to develop a fast and effective learning method for intelligent robotic manipulation of typical industrial tasks, including pushing, picking, and placing workpieces, which are essential and fundamental processing plan activities for accomplishing robotic machining and assembly applications towards smart manufacturing. The introduction of reinforcement learning enables robots to learn manipulation tasks autonomously, which can save the effort for engineers to teach or hard program the robot and also reduce labor costs. However, the existing HER-based reinforcement learning algorithms are with low training efficiency and performance due to the uniform sampling and scant task reward. Inspired by human learning, this work introduces a progress incentive mechanism to identify valuable trajectory data for effective training. In addition, a novel rewarding method, that applies additional intrinsic rewards for agents learning valuable trajectory space, results in fast and robust learning. The setting of important weight parameters in the rewarding method is given in the paper, which provides a practical reference for applying the proposed algorithm. The average success rate of two actual manipulation tasks in simulation and real robotic manipulation environments are 96% and 92.5%, respectively, which demonstrates that the method is effective for both environments and there is 3.5% average gap of successful rate dropping from simulation scenarios to real ones due to the inherent mismatches between simulation and reality. The high success rate demonstrated in the real Workpieces-sorting task exemplifies the potential of the trained policies for application in industrial scenarios.

Object six degrees of freedom (6-DoF) pose estimation is the powerful vision algorithm for the robot-environment interaction. However, current robust pose estimation algorithms rely heavily on labeled real data with high-cost collection, making it difficult to apply the algorithm. Many studies discuss the use of synthetic data as a complement to real datasets. However, reducing the gap between synthetic and real data is still a challenging problem. Based on the consistency of object geometric characteristics between real data and synthetic data, we argue that multi-input, rather than image-only input, is more suitable for transfer from synthetic to real, because it strengthens the extraction of object geometric feature. Therefore, we propose a semantic-fusion 6-DoF object pose estimation method that effectively capture common features across various resolutions by employing the designed pyramid transformer feature-fusion module. Extensive experiments show that the proposed method performs better than the state-of-the-art (SOTA), indicating that the proposed method can effectively extract and fuse different representations. Furthermore, in response to the lack of industrial scene datasets, we also develop a synthetic pose dataset and conduct the human-robot collaboration experiment to verify the robustness of the proposed method. Note to Practitioners—The purpose of this paper is to bridge the gap between synthetic and real data for pose estimation of industrial tools. Our method can be trained only on synthetic data and accurately estimate pose parameters in real scenes. Combining physically-based renderer and industrial tools, such as hammers and screwdrivers, a synthetic dataset of industrial scenes can be produced using the data production pipeline proposed in this paper. In this case, the trained model can assist the robot vision system to understand object pose information in a real production workshop. Extensive dataset experiments and human-robot collaboration experiments demonstrate the effectiveness of the proposed method. In addition, based on the actual robot working environment, practitioners can produce industrial datasets from multiple angles, objects, and scenes. Sufficient datasets can enhance the model’s generalization and robustness.

Despite being widely used in network intrusion detection systems (NIDSs), machine learning (ML) has proven to be vulnerable to adversarial attacks. White-box and black-box adversarial ML attacks of NIDS have been explored in several studies. However, white-box attacks unrealistically assume that the attackers have full knowledge of the target NIDSs. Meanwhile, existing black-box attacks can not achieve high attack success rate due to the weak adversarial transferability between models (e.g., neural networks and tree models). Additionally, neither of them explains why adversarial examples exist and why they can transfer across models. To address these challenges, this paper introduces ETA, anExplainableTransfer-based Black-Box AdversarialAttack framework. ETA aims to achieve two primary objectives: 1) create transferable adversarial examples applicable to various ML detectors and 2) provide insights into the existence of adversarial examples and their transferability within NIDSs. Specifically, we first provide a general transfer-based adversarial attack method applicable across the entire ML space. Following that, we exploit a unique insight based on cooperative game theory and perturbation interpretations to explain adversarial examples and adversarial transferability. On this basis, we propose an Important-Sensitive Feature Selection (ISFS) method to guide the search for adversarial examples, achieving stronger transferability and ensuring traffic-space constraints. Finally, the experimental results on three NIDSs datasets show that our method performs significantly effectively against several classical and state-of-the-art ML classifiers, outperforming the latest baselines. We conduct three interpretation experiments and two cases to verify our interpretation method's correctness. Meanwhile, we uncover two major misconceptions about applying machine learning to NIDSs systems.

To bridge the gap of conventional single-hop task offloading schemes in infrastructure-free scenarios, multi-hop task offloading schemes for IoT devices in Mobile Edge Computing (MEC) are desired to jointly optimize task offloading decisions and routing paths. In this paper, we investigate a hierarchical multi-hop edge computing framework and propose a joint Task Offloading and Relay Selection (TORS) scheme. It considers real-time computation at each relay node and employs directional searches to facilitate the task execution and results reporting at the fastest speed. However, finding the optimal TORS solution is a formidable challenge due to the time-varying network environments, the strong interdependence of decision sets across different time slots, and the high computational complexity. To address these challenges, we first leverage Lyapunov optimization to transform the stochastic TORS problem into a deterministic per-slot block problem, avoiding the need for extensive system prior knowledge. Subsequently, we propose a Soft Actor-Critic (SAC)-based algorithm, SAC-TORS, to find a satisfactory TORS solution with minimal computational complexity in a distributed manner. Accordingly, each IoT device can independently make self-determined and directional decisions with observable network information. Through extensive experiments, we demonstrate that the SAC-TORS outperforms state-of-the-art solutions, achieving performance improvements of up to 66%.

The ever-increasing complexity and scale of 5G communication networks pose huge challenges to network operations. Root cause analysis is considered as a promising method for fault detection. However, it still suffers challenges of severely uneven distribution of fault data, low accuracy in root cause detection, and long time consumption due to a large search space in 5G cellular networks. To address the above challenges, we introduce SimRCA to effectively analyze the faults’ root causes in 5G networks using signalling messages. By designing a novel confidence threshold value and pruning technique, SimRCA can significantly reduce the search space of signalling messages while maintaining the accuracy of root cause analysis. Moreover, SimRCA is proven to be able to handle unbalanced data distribution in 5G networks. We collected over 10GB of signalling data from Huawei 5G commercial network and conducted extensive experiments on this dataset. Experimental results demonstrate that SimRCA can complete root cause localization and fault classification within 11 seconds with an average F1-score over 0.93 which outperforms the current state-of-the-art solutions.

The increasing complexity and openness of 5G networks naturally enlarge the attack surface and introduce new vulnerabilities, thereby posing challenges to the performance of existing intrusion detection systems (IDSs). Current IDSs solely rely on statistical features, which may suffer from low accuracy due to the complex traffic patterns in 5G networks. Additionally, recent IDSs apply contrastive learning to improve detection capabilities, but the reliance on costly manual labeling hinders the adaptability to complex attacks in 5G networks.In this paper, we present Manticore, an unsupervised intrusion detection system based on contrastive learning for 5G networks. Specifically, Manticore leverages both statistical features and original features of packets to capture the holistic information of traffic in 5G networks. Moreover, it automatically establishes positive and negative pairs without manual labeling. We further explore the combination patterns between reconstruction loss and contrastive loss to attain a more precise model. Our experimental evaluation of two datasets demonstrates the proposed Manticore outperforms the relevant state-of-the-art methods.

Recently, Self-Supervised Learning (SSL) has achieved great success in various famous applications e.g., BERT and ChatGPT. However, when applying SSL to safety-critical downstream tasks, such as self-driving cars, potential adversarial attacks can completely change the final decisions and thus lead to serious security issues. To overcome this issue, existing methods combine adversarial training with pre-training to improve the adversarial robustness of SSL. However, combining these two computationally complex processes may largely amplify the computation cost. Moreover, whether performing adversarial training in pre-training or fine-tuning, current methods may degrade the accuracy due to the famous catastrophic forgetting problem. The computation cost of current adversarial training methods based on full parameter updating is still high even in the fine-tuning stage.To address the above challenges, we propose an effective robust fine-tuning framework for SSL based on Low-Rank Adaptation (LoRA), named LoFT. First, LoFT performs adversarial training in the fine-tuning stage rather than in the pre-training stage. Second, LoFT innovatively and elaborately integrates LoRA into adversarial training to avoid the catastrophic forgetting problem. Third, LoFT exploits a low-rank matrix in LoRA which enables efficient fine-tuning by updating only a small set of parameters, which contains only 1%-5% of the parameters of the pre-trained model. The whole pre-training and fine-tuning stages take only 9.44 hours, which reduces training time by 3× over the current SOTA method. Furthermore, compared with existing SOTA robust pre-training methods for SSL, LoFT improves accuracy by 5.97% (77.41%⇒83.38%) and robustness by 13% (45.04%⇒58.44%) on the CIFAR-10 dataset.

Fisheye cameras, offering a wide field of view (FOV) of 360◦, are extensively employed for surround-view perception in autonomous driving. Compared with the object detection on the standard images, it lacks studies for fisheye images. Moreover, efficient perception is crucial for autonomous vehicles with limited computational capability. In this work, we introduce a lightweight fisheye object detection network with transformer-based feature enhancement for autonomous driving. Specifically, we leverage ShuffleNet V2 as a feature extraction network to reduce computation complexity and develop a transformer-based feature enhancement module (TFEM) to integrate multi-level features. Notably, we observe that data augmentation methods like mix-up and mosaic, effective on standard images, do not yield positive results on fisheye images. The results on the WoodScape dataset demonstrate that our method can achieve better performance with fewer parameters and floating-point operations per second (FLOPs). Extending our evaluation to the Microsoft Common Objects in Context (MS COCO) dataset shows that the proposed method has excellent generalization capability.

Estimating the rigid transformation with 6 degrees of freedom based on a putative 3D correspondence set is a crucial procedure in point cloud registration. Existing correspondence identification methods usually lead to large outlier ratios (>95% is common), underscoring the significance of robust registration methods. Many researchers turn to parameter search-based strategies (e.g., Branch-and-Bround) for robust registration. Although related methods show high robustness, their efficiency is limited to the high-dimensional search space. This paper proposes a heuristics-guided parameter search strategy to accelerate the search while maintaining high robustness. We first sample some correspondences (i.e., heuristics) and then just need to sequentially search the feasible regions that make each sample an inlier. Our strategy largely reduces the search space and can guarantee accuracy with only a few inlier samples, therefore enjoying an excellent trade-off between efficiency and robustness. Since directly parameterizing the 6-dimensional nonlinear feasible region for efficient search is intractable, we construct a three-stage decomposition pipeline to reparameterize the feasible region, resulting in three lower-dimensional sub-problems that are easily solvable via our strategy. Besides reducing the searching dimension, our decomposition enables the leverage of 1-dimensional interval stabbing at all three stages for searching acceleration. Moreover, we propose a valid sampling strategy to guarantee our sampling effectiveness, and a compatibility verification setup to further accelerate our search. Extensive experiments on both simulated and real-world datasets demonstrate that our approach exhibits comparable robustness with state-of-the-art methods while achieving a significant efficiency boost.

Point cloud registration is challenging in the presence of heavy outlier correspondences. This paper focuses on addressing the robust correspondence-based registration problem with gravity prior that often arises in practice. The gravity directions are typically obtained by inertial measurement units (IMUs) and can reduce the degree of freedom (DOF) of rotation from 3 to 1. We propose a novel transformation decoupling strategy by leveraging the screw theory. This strategy decomposes the original 4-DOF problem into three sub-problems with 1-DOF, 2-DOF, and 1-DOF, respectively, enhancing computation efficiency. Specifically, the first 1-DOF represents the translation along the rotation axis, and we propose an interval stabbing-based method to solve it. The second 2-DOF represents the pole which is an auxiliary variable in screw theory, and we utilize a branch-and-bound method to solve it. The last 1-DOF represents the rotation angle, and we propose a global voting method for its estimation. The proposed method solves three consensus maximization sub-problems sequentially, leading to efficient and deterministic registration. In particular, it can even handle the correspondence-free registration problem due to its significant robustness. Extensive experiments on both synthetic and real-world datasets demonstrate that our method is more efficient and robust than state-of-the-art methods, even when dealing with outlier rates exceeding 99%.

This paper develops an extremely robust solution for absolute pose estimation with known prior gravity direction by motion decoupling. Absolute pose estimation is a fundamental problem in computer vision, and recently the prior known vertical direction is commonly applied to help solve the pose estimation problem. In this paper, we explore the geometrical constraints of the absolute pose estimation with a known direction. We find that the rigid pose can be decoupled with the help of the known direction. Thereby, absolute pose estimation algorithms, which decouple rigid motion, are proposed. Notably, in real applications, there may be imperfect inputs, i.e., outliers, due to incorrect 2D-3D matches. Unfortunately, these outliers may lead to unacceptable results. To suppress the outliers, the decoupled absolute pose estimation problem is solved by branch-and-bound algorithm and globally voting, which can provide the optimal solution with provable guarantees. Moreover, in extreme case, the proposed method can solve absolute pose estimation problem without knowing the 2D-3D correspondences, which is also known as simultaneous camera pose correspondence estimation. To demonstrate the feasibility and the superiority of the proposed methods, comprehensive comparison experiment are conduced. The source code is available at https://github.com/Liu-Yinlong/algorithm-for-PnP-with-known-vertical-direction.

Posting programs to code hosting platforms such as GitHub is common for developers, but it will lead to privacy leakage issues in hosted programs. Though there are some detection methods for privacy leakage, they are difficult to be applied in practice. First, existing works mainly focus on detection algorithms, while ignoring the complete detection system from a holistic perspective. Second, the system will be blocked when acquiring programs because code hosting platforms usually have protection mechanism. Third, high-performance privacy detection algorithms need hardware devices in practice, and their effectiveness in real scenarios has not been verified since there is no public real hosted program privacy dataset.To address the above problems, we implement and commercialize a user-friendly privacy information leakage detection system for actual hosted programs. Firstly, we provide a system frame-work that can automatically complete "program acquisition-privacy detection-alert", allowing subscribers receive alerts if there is a privacy information leakage. Secondly, we propose a novel multi-random crawler scheme that can flexibly cope with the limitations of GitHub when acquiring hosted programs. Thirdly, we skillfully apply a cost-effective detection approach based on fuzzy matching, which can detect the subscriber customized privacy information with high performance. Based on this, we further provide a high-quality dataset obtained in real scenarios. Finally, we conduct comprehensive experiments to evaluate our system. Experimental results demonstrate the effectiveness of our crawler scheme and detection approach in providing a user-friendly system.

This paper describes our submission to the ICASSP 2023 Signal Processing Grand Challenge (SPGC), which focuses on multilingual Alzheimer’s disease (AD) recognition through spontaneous speech. Our approaches include using a variety of acoustic features and silence-related information for AD detection and mini-mental state examination (MMSE) score prediction, and fine-tuning wav2vec2.0 models on speech in various frequency bands for AD detection. Our overall results on the test data outperform the baseline provided by the organizers, achieving 73.9% accuracy in AD detection by fine-tuning our bilingual wav2vec2.0 pre-trained model on the 0-1000Hz frequency band speech, and 4.610 RMSE (r = 0.565) in MMSE prediction through the fusion of eGeMAPS and silence features.

With the rapid development of the cutting edge cloud computing technology, millions of vulnerabilities have been identified, there is a growing concern that organizations should devote plenty of time and lots of resources to secure. The overarching objective of remediation is to prioritize the vulnerabilities. Hence, define the severity and the urgency of the vulnerabilities and remediate them automatically is very important. Although the recognized Common Vulnerability Scoring System (CVSS) 4.0 method addresses this issues partly, they are difficult to be implemented in practices on the cloud because of the complication and lack of risk based factors.To this end, we propose a Cost-effective Massive Automation Method of Vulnerability Analysis and Remediation Based on Cloud Native Framework. Specifically, considering that the current CVSS is more like a severity of vulnerabilities, we design a novel formula to define the urgency of vulnerabilities. The formula takes the advantaged of the capabilities of modern cloud-based infrastructure and simplifies the CVSS. Besides, we propose an algorithm of risk reduction by leveraging the cloud native security capabilities, which cut down unnecessary patching time and workload. Particularly, in order to remediation the risk on the cloud, we implement an automatic scheme to harden the vulnerabilities by invoking the cloud native APIs based on the Security Orchestration, Automation and Response (SOAR) platform. Finally, we conduct comprehensive experiments to evaluate our system. Experimental results demonstrate the effectiveness of ours approach has a high ratio of urgency risk recognition of 99.24%. Meanwhile, ours approach shows a maximum risk reduction by downgrade the fixable vulnerability with a average of 79% risk reduction rate in application level and 99% of risk reduction rate in operating system level respectively. As a result, our approach lightens the workload of patching greatly in the real cloud computing environment.

Point cloud registration is a fundamental problem in 3D computer vision. Previous learning-based methods for LiDAR point cloud registration can be categorized into two schemes: dense-to-dense matching methods and sparse-to-sparse matching methods. However, for large-scale outdoor LiDAR point clouds, solving dense point correspondences is time-consuming, whereas sparse keypoint matching easily suffers from keypoint detection error. In this paper, we propose SDMNet, a novel Sparse-to-Dense Matching Network for large-scale outdoor LiDAR point cloud registration. Specifically, SDMNet performs registration in two sequential stages: sparse matching stage and local-dense matching stage. In the sparse matching stage, we sample a set of sparse points from the source point cloud and then match them to the dense target point cloud using a spatial consistency enhanced soft matching network and a robust outlier rejection module. Furthermore, a novel neighborhood matching module is developed to incorporate local neighborhood consensus, significantly improving performance. The local-dense matching stage is followed for fine-grained performance, where dense correspondences are efficiently obtained by performing point matching in local spatial neighborhoods of high-confidence sparse correspondences. Extensive experiments on three large-scale outdoor LiDAR point cloud datasets demonstrate that the proposed SDMNet achieves state-of-the-art performance with high efficiency.

Point cloud registration is a fundamental problem in 3D computer vision. Outdoor LiDAR point clouds are typically large-scale and complexly distributed, which makes the registration challenging. In this paper, we propose an efficient hierarchical network named HRegNet for large-scale outdoor LiDAR point cloud registration. Instead of using all points in the point clouds, HRegNet performs registration on hierarchically extracted keypoints and descriptors. The overall framework combines the reliable features in deeper layer and the precise position information in shallower layers to achieve robust and precise registration. We present a correspondence network to generate correct and accurate keypoints correspondences. Moreover, bilateral consensus and neighborhood consensus are introduced for keypoints matching, and novel similarity features are designed to incorporate them into the correspondence network, which significantly improves the registration performance. In addition, we design a consistency propagation strategy to effectively incorporate spatial consistency into the registration pipeline. The whole network is also highly efficient since only a small number of keypoints are used for registration. Extensive experiments are conducted on three large-scale outdoor LiDAR point cloud datasets to demonstrate the high accuracy and efficiency of the proposed HRegNet. The source code of the proposed HRegNet is available at https://github.com/ispc-lab/HRegNet2.

Radar is among the most popular sensors in modern Intelligent Transportation Systems (ITSs), enabling weather-robust perception. The orientation and position of the traffic radar relative to the ITS coordinate system are necessary for the perception fusion in ITSs. However, due to the unknown target association, sparseness and noisiness of traffic radar measurements, the robust and accurate extrinsic calibration of traffic radar is challenging. In this paper, we propose a targetless traffic radar calibration method based on GPS to overcome the inconvenience during ITS operation, because the installation of a dedicated calibration target on the highway is impractical and dangerous. On the other hand, the high-precision GPS device installed on the moving vehicle can provide traffic radar with accurate positioning information of the detection target. Furthermore, during the optimization process of extrinsic calibration, we propose a globally optimal registration method, which is robust to noise and outliers in radar measurements, and is called Gaussian Mixture Robust Branch and Bound (GMRBnB). Specifically, we first construct the robust objective function by utilizing the Gaussian Mixture Model (GMM). Then, we derive novel relaxation bounds and present the GMRBnB algorithm that overcomes the susceptibility to local minima and the dependence on initialization of traditional optimization methods. Compared with existing methods, extensive experiments in synthetic and real-world data demonstrate that our method is not only globally optimal, but also more accurate and robust.

With the advancements in Machine Learning (ML) and edge computing, increasing efforts have been devoted toedge video analytics. However, most of the existing works fail to consider the cooperation of edge nodes for ML model caching and video frame scheduling, thus less efficient in practical scenarios with diverse requirements. In this article, we propose a novel approach named ESMO (joint framEScheduling andMOdel caching) to jointly optimize Frame Scheduling and Model Caching (FSMC), aiming at enhancing the performance of edge video analytics. In detail, we decompose the FSMC as three sub-problems, where the first two sub-problems (i.e., user's transmit power and edge computing resources allocation problems) are proven to be quasi-convex and strictly convex, respectively; while the third main sub-problem (i.e., trade-off among the video analytics (VA) accuracy, service delay and energy consumption) is NP-hard. Therefore, an efficient Two-layers Genetic Algorithm based algorithm (i.e., TGA-FSMC) is designed to find the close-to-optimal frame scheduling and the model caching decisions in an iterative manner. Finally, we deploy a target recognition prototype to comprehensively evaluate the practical performance in diverse edge nodes and CNN models. Extensive experiments demonstrate the empirical superiority of the ESMO over alternatives on real-world edge video analytics platforms, and it achieves 37.5%$\sim$87.2% performance improvement.

Effective adversarial attacks simulation is essential for the deployment of ensemble Intrusion Detection Systems (en- semble IDSs) in Satellite-Terrestrial Integrated Network (STIN). This is because it can automatically generate a large amount of adversarial samples to evaluate the robustness of different classifiers. Based on the result, it can further guide the STIN engineers to select proper classifiers in ensemble IDSs. Moreover, it can help the IDSs improve detect performance by their self- learning property in the adversarial attack process. However, the existing adversarial attack approaches suffer from the problems of low success rate and high overhead of communication and calculation due to the limited computing resources and long communication links of STIN. This results in their inefficiency in STIN. To address the above problems, we provide StinAttack as a robustness evaluation scheme for STIN. First, StinAttack provides a comprehensive and automatic robustness evaluation framework for IDSs in STIN with only few times interactions between terrestrial and satellite nodes. Second, StinAttack proposes an effective adversarial attack simulation based on lightweight gradient evaluation for ensemble IDSs. Third, we conduct experiments on 11 typical IDSs, 4 baseline popular adversarial attacks and our StinAttack. Experimental results show that our approach can effectively attack ensemble IDSs and the evaluation results based on real STIN dataset are instructive for designing secure networks.

One key challenge to maintain a large-scale Content Delivery Network (CDN) is to minimize the service downtime when severe system problems happen (e.g., hardware failures). In this case, a critical step is to quickly and accurately identify the range of users with performance degradation, termed impact identification. Successful impact identification not only helps identify impacted users but also provides meaningful information for troubleshooting. However, current practice of impact identification usually takes network engineers several hours to manually identify impacted users, which may lead to a huge business loss. The main challenges for automatic impact identification in large CDNs include the inaccuracy of underlying anomaly detection, huge search space of impact identification and severe long-tail distribution of user traffic. In this paper we propose iSwift, a system that is specifically designed for impact identification in large-scale CDNs in order to address aforementioned challenges. We evaluate the performance of iSwift on semi-synthetic datasets and the results show that iSwift can achieve a F1-score greater than 0.85 within ten seconds, which significantly outperforms state-of-the-art solutions. Furthermore, iSwift has been deployed in a production CDN around one year as a pilot project and demonstrated its online performance confirmed by the network operators.

System instance clustering is crucial for large-scale Web services because it can significantly reduce the training overhead of anomaly detection methods. However, the vast number of system instances with massive time points, redundant metrics, and noise bring significant challenges. We propose OmniCluster to accurately and efficiently cluster system instances for large-scale Web services. It combines a one-dimensional convolutional autoencoder (1D-CAE), which extracts the main features of system instances, with a simple, novel, yet effective three-step feature selection strategy. We evaluated OmniCluster using real-world data collected from a top-tier content service provider providing services for one billion+ monthly active users (MAU), proving that OmniCluster achieves high accuracy (NMI=0.9160) and reduces the training overhead of five anomaly detection models by 95.01% on average.

Timely anomaly detection of key performance indicators (KPIs),e.g., service response time, error rate, is of utmost importance to Web services. Over the years, many unsupervised deep learning-based anomaly detection approaches have been proposed. To achieve good performance, they require a long period of KPI data for model training, which is not easy to guarantee with frequent service changes. Additionally, the training overhead is too significant for the vast number of KPIs in large-scale Web services. To address the problems, we propose an unsupervised KPI anomaly detection approach, namedAnoTransfer, by combining a novel Variational Auto-Encoder (VAE)-based KPI clustering algorithm with an adaptive transfer learning strategy. Extensive evaluation experiments using real-world data collected from several large-scale Web service providers demonstrate thatAnoTransferreduces the average initialization time by 65.71% and improves the training efficiency by 50.62 times, without significantly degrading anomaly detection accuracy.

In man-made environments, most of the objects and structures are organized in the form of orthogonal and parallel planes. These planes can be approximated by an Atlanta world assumption, in which the normals of planes can be represented by Atlanta frames. The Atlanta world assumption has one vertical frame and multiple horizontal frames. Conventionally, given a set of inputs such as surface normals, the Atlanta frame estimation problem can be solved by a branch-and-bound (BnB) algorithm. However, the runtime of the BnB algorithm will increase greatly when the dimensionality (i.e., the number of horizontal frames) increases. In this paper, we estimate only the vertical direction, instead of all Atlanta frames at once. Accordingly, we propose a vertical direction estimation method by considering the relationship between the vertical frame and horizontal frames. Concretely, our approach employs a BnB algorithm to search the vertical direction, thereby guaranteeing global optimality without requiring prior knowledge of the number of Atlanta frames. In order to guarantee convergence, four novel bounds are investigated, by mapping a 3D hemisphere to a 2D region. We verify the feasibility of the proposed method using various challenging synthetic and real-world data.

Estimating the pose of a calibrated camera relative to a 3D point set from one image is an important task in computer vision. Perspective-n-Point algorithms are often used if perfect 2D-3D correspondences are known. However, it is difficult to determine 2D-3D correspondences perfectly, and then the simultaneous pose and correspondence determination problem is needed to be solved. Early methods aimed to solve this problem by local optimization. Recently, several new methods are proposed to globally solve this problem by using branch-and-bound (BnB) method, but they tend to be slow because the time complexity of the BnB-based methods is exponential to the dimensionality of the parameter space, and they directly search the 6D parameter space. In this paper, we propose to decompose the joint searching into two separate searching processes by introducing a rotation invariant feature (RIF). Specifically, we construct RIFs from the original 3D and 2D point sets and search for the globally optimal translation to match these two RIFs first. Then, the original 3D point set is translated and matched with the 2D point set to find a globally optimal rotation. Experiments on challenging data show that the proposed method outperforms state-of-the-art methods in terms of both speed and accuracy.

Predicting the future can significantly improve the safety of intelligent vehicles, which is a key component in autonomous driving. 3D point clouds can accurately model 3D information of surrounding environment and are crucial for intelligent vehicles to perceive the scene. Therefore, prediction of 3D point clouds has great significance for intelligent vehicles, which can be utilized for numerous further applications. However, due to point clouds are unordered and unstructured, point cloud prediction is challenging and has not been deeply explored in current literature. In this paper, we propose a novel motion-based neural network named MoNet. The key idea of the proposed MoNet is to integrate motion features between two consecutive point clouds into the prediction pipeline. The introduction of motion features enables the model to more accurately capture the variations of motion information across frames and thus make better predictions for future motion. In addition, content features are introduced to model the spatial content of individual point clouds. A recurrent neural network named MotionRNN is proposed to capture the temporal correlations of both features. Moreover, an attention-based motion align module is proposed to address the problem of missing motion features in the inference pipeline. Extensive experiments on two large-scale outdoor LiDAR point cloud datasets demonstrate the performance of the proposed MoNet. Moreover, we perform experiments on applications using the predicted point clouds and the results indicate the great application potential of the proposed method.

Although great progress has been made in generic object detection by advanced deep learning techniques, detecting small objects from images is still a difficult and challenging problem in the field of computer vision due to the limited size, less appearance, and geometry cues, and the lack of large-scale datasets of small targets. Improving the performance of small object detection has a wider significance in many real-world applications, such as self-driving cars, unmanned aerial vehicles, and robotics. In this article, the first-ever survey of recent studies in deep learning-based small object detection is presented. Our review begins with a brief introduction of the four pillars for small object detection, includingmultiscale representation, contextual information, super-resolution, and region-proposal. Then, the collection of state-of-the-art datasets for small object detection is listed. The performance of different methods on these datasets is reported later. Moreover, the state-of-the-art small object detection networks are investigated along with a special focus on the differences and modifications to improve the detection performance comparing to generic object detection architectures. Finally, several promising directions and tasks for future work in small object detection are provided. Researchers can track up-to-date studies on this webpage available at:https://github.com/tjtum-chenlab/SmallObjectDetectionList.

LiDAR point cloud streams are usually sparse in time dimension, which is limited by hardware performance. Generally, the frame rates of mechanical LiDAR sensors are 10 to 20 Hz, which is much lower than other commonly used sensors like cameras. To overcome the temporal limitations of LiDAR sensors, a novel task named Point Cloud Frame Interpolation is studied in this paper. Given two consecutive point cloud frames, Point Cloud Frame Interpolation aims to generate intermediate frame(s) between them. To achieve that, we propose a novel framework, namely Point Cloud Frame Interpolation Network (PointINet). Based on the proposed method, the low frame rate point cloud streams can be upsampled to higher frame rates. We start by estimating bi-directional 3D scene flow between the two point clouds and then warp them to the given time step based on the 3D scene flow. To fuse the two warped frames and generate intermediate point cloud(s), we propose a novel learning-based points fusion module, which simultaneously takes two warped point clouds into consideration. We design both quantitative and qualitative experiments to evaluate the performance of the point cloud frame interpolation method and extensive experiments on two large scale outdoor LiDAR datasets demonstrate the effectiveness of the proposed PointINet. Our code is available at https://github.com/ispc-lab/PointINet.git.

Point cloud registration is a fundamental problem in 3D computer vision. Outdoor LiDAR point clouds are typically large-scale and complexly distributed, which makes the registration challenging. In this paper, we propose an efficient hierarchical network named HRegNet for large-scale out-door LiDAR point cloud registration. Instead of using all points in the point clouds, HRegNet performs registration on hierarchically extracted keypoints and descriptors. The overall framework combines the reliable features in deeper layer and the precise position information in shallower layers to achieve robust and precise registration. We present a correspondence network to generate correct and accurate keypoints correspondences. Moreover, bilateral consensus and neighborhood consensus are introduced for keypoints matching and novel similarity features are designed to in-corporate them into the correspondence network, which significantly improves the registration performance. Besides, the whole network is also highly efficient since only a small number of keypoints are used for registration. Extensive experiments are conducted on two large-scale outdoor LiDAR point cloud datasets to demonstrate the high accuracy and efficiency of the proposed HRegNet. The project website is https://ispc-group.github.io/hregnet.

In remote regions (e.g., mountain and desert), cellular networks are usually sparsely deployed or unavailable. With the appearance of new applications (e.g., industrial automation and environment monitoring) in remote regions, resource-constrained terminals become unable to meet the latency requirements. Meanwhile, offloading tasks to urban terrestrial cloud (TC) via satellite link will lead to high delay. To tackle above issues, Satellite Edge Computing architecture is proposed, i.e., users can offload computing tasks to visible satellites for executing. However, existing works are usually limited to offload tasks in pure satellite networks, and make offloading decisions based on the predefined models of users. Besides, the runtime consumption of existing algorithms is rather high. In this paper, we study the task offloading problem in satellite-terrestrial edge computing networks, where tasks can be executed by satellite or urban TC. The proposed Deep Reinforcement learning-based Task Offloading (DRTO) algorithm can accelerate learning process by adjusting the number of candidate locations. In addition, offloading location and bandwidth allocation only depend on the current channel states. Simulation results show that DRTO achieves near-optimal offloading cost performance with much less runtime consumption, which is more suitable for satellite-terrestrial network with fast fading channel.

Mobile edge computing (MEC) has been envisaged as one of the most promising technologies in the fifth generation (5G) mobile networks. It allows mobile devices to offload their computation‐demanding and latency‐critical tasks to the resource‐rich MEC servers. Accordingly, MEC can significantly improve the latency performance and reduce energy consumption for mobile devices. Nonetheless, privacy leakage may occur during the task offloading process. Most existing works ignored these issues or just investigated the system‐level solution for MEC. Privacy‐aware and user‐level task offloading optimization problems receive much less attention. In order to tackle these challenges, a privacy‐preserving and device‐managed task offloading scheme is proposed in this paper for MEC. This scheme can achieve near‐optimal latency and energy performance while protecting the location privacy and usage pattern privacy of users. Firstly, we formulate the joint optimization problem of task offloading and privacy preservation as a semiparametric contextual multi‐armed bandit (MAB) problem, which has a relaxed reward model. Then, we propose a privacy‐aware online task offloading (PAOTO) algorithm based on the transformed Thompson sampling (TS) architecture, through which we can (1) receive the best possible delay and energy consumption performance, (2) achieve the goal of preserving privacy, and (3) obtain an online device‐managed task offloading policy without requiring any system‐level information. Simulation results demonstrate that the proposed scheme outperforms the existing methods in terms of minimizing the system cost and preserving the privacy of users.

Advanced Persistent Threat (APT) is a stealthy, continuous and sophisticated method of network attacks, which can cause serious privacy leakage and millions of dollars losses. In this paper, we introduce a new game-theoretic framework of the interaction between a defender who uses limited Security Resources(SRs) to harden network and an attacker who adopts a multi-stage plan to attack the network. The game model is derived from Stackelberg games called a Multi-stage Maze Network Game (M2NG) in which the characteristics of APT are fully considered. The possible plans of the attacker are compactly represented using attack graphs(AGs), but the compact representation of the attacker’s strategies presents a computational challenge and reaching the Nash Equilibrium(NE) is NP-hard. We present a method that first translates AGs into Markov Decision Process(MDP) and then achieves the optimal SRs allocation using the policy hill-climbing(PHC) algorithm. Finally, we present an empirical evaluation of the model and analyze the scalability and sensitivity of the algorithm. Simulation results exhibit that our proposed reinforcement learning-based SRs allocation is feasible and efficient.

With the rapid growth of multimedia content in the social content-centric network (SocialCCN), in-network caching and caching strategy are becoming more and more important for efficient content delivery, but it also brings huge challenges to the cache space and computing capabilities in the network. In order to increase cache space and improve the computing capability in SocialCCN, in this paper, we integrate Mobile edge computing with SocialCCN (MeSoCCN) and design a novel caching strategy in MeSoCCN. Firstly, we proposed MeSoCCN, a novel architecture that integrates Mobile Edge Computing (MEC) in SocialCCN. Then, in MeSoCCN, a caching strategy based on popularity prediction is designed, which can increase the cache hit rate and reduce hop redundancy. We predict content popularity in the future and make cache placement and replacement decisions based on the prediction results. Finally, we conducted experiments and verified the effectiveness of the proposed caching strategy in MeSoCCN.

Anomaly detection on attributed networks aims to differentiate rare nodes that are significantly different from the majority. It plays an important role in various practical scenarios, such as intrusion detection and fraud detection. However, existing graph-based methods mainly adopt shallow models that cannot capture the highly non-linear interactions between nodes in an attribute network consisting of different information modalities. To tackle the above issues, in this paper, we propose a novel deep model named DeepAE for anomaly detection which (a) can capture the high non-linearity in both topological structure and nodal attributes through graph convolutional autoencoder, (b) fully exploits the intrinsic information of the network with the description of various proximities, (c) and preserve the differences between anomalies and the majority by applying Laplacian sharpening. We perform anomaly detection by measuring the reconstruction errors of nodes. Experimental results on realworld datasets demonstrate that DeepAE outperforms the stateof-art baselines.

Keypoint detector and descriptor are two main components of point cloud registration. Previous learning-based keypoint detectors rely on saliency estimation for each point or farthest point sample (FPS) for candidate points selection, which are inefficient and not applicable in large scale scenes. This paper proposes Random Sample-based Keypoint Detector and Descriptor Network (RSKDD-Net) for large scale point cloud registration. The key idea is using random sampling to efficiently select candidate points and using a learning-based method to jointly generate keypoints and corresponding descriptors. To tackle the information loss of random sampling, we exploit a novel random dilation cluster strategy to enlarge the receptive field of each sampled point and an attention mechanism to aggregate the positions and features of neighbor points. Furthermore, we propose a matching loss to train the descriptor in a weakly supervised manner. Extensive experiments on two large scale outdoor LiDAR datasets show that the proposed RSKDD-Net achieves state-of-the-art performance with more than 15 times faster than existing methods. Our code is available at https://github.com/ispc-lab/RSKDD-Net.

Anomaly Detection in Dynamic Networks plays a critical role in various real-world applications such as cyberse-curity, e-commerce, and social media. Recent approaches based on graph neural networks have achieved fruitful results in static networks, and most of the researches focus on learning node embeddings to represent the large-scale complex networks, thereby facilitating downstream anomaly detection task. However, these methods require the whole network to extract intrinsic properties, so they are sensitive to the frequent changes of nodes and incapable of capturing the dynamism as realworld networks evolve over time. In this paper, we propose a novel framework DynAD for anomalous edge detection on time-evolving networks, which performs adaptive parameter learning in an end-to-end manner. In particular, DynAD first extracts fixed-size node embedding from each snapshot with temporal graph convolution and pooling operations. Then, it captures the temporal information of the graph sequence using Gated recurrent units (GRU) for anomaly detection, where an attention mechanism is employed to highlight the differences between the dynamic patterns. Experimental results on three real-world datasets illustrate that DynAD significantly outperforms the state-of-the-art baseline methods in anomaly detection.

Content-Centric Networking (CCN) is one of the most important next-generation Internet architectures. The in-network caching mechanism in CCN can bring higher efficiency and lower traffic to the network in terms of content distribution, but it also poses a great privacy risks. In this paper, we propose a privacy-preserving scheme based on fragments storage and fragments recombination (FS&FR) to solve the user's privacy leakage problem caused by timing attack in CCN. Firstly, the content in the network can be divided into different privacy levels according to the content provider, content consumer and router. Secondly, the optimal number of content fragments can be obtained by binary linear regression model based on content popularity, node betweenness and content privacy levels. Finally, the FS&FR algorithm is proposed and applied to the private content for content distribution and achieving fine-grained privacy protection. The simulation results show that the proposed scheme is secure yet highly efficient again timing attack compared to random-K delay algorithm. More specifically, the FS&FR algorithm can make the round-trip delays obtained by the attacker requesting the same content change, and then protect users' behavior privacy without sacrificing distribution performance.

One of the most important security threats in Content Centric Networking (CCN) is cache privacy disclosure, where an adversary can obtain the privacy information of a legitimate user through timing attack. In view of this security problem, we propose a cache privacy protection strategy based on content privacy and user security classification (CPUSC). While protecting user privacy, CPUSC also takes into account the efficiency of content distribution. Firstly, CPUSC starting from user's subjective point of view to determine content privacy level. Secondly, the router in the system assigns a credit value to each user according to the user's behavior and sets a security level for each of them. Then, in the light of the credit value, the user's security level can be defined. Finally, a latency is set for each content response based on the content privacy level and the user's security level. Performance analysis show that the proposed scheme can effectively protect the user's privacy while ensuring the content distribution efficiency of the network.

Simultaneously determining the relative pose and correspondence between a set of 3D points and its 2D projection is a fundamental problem in computer vision, and the problem becomes more difficult when the point sets are contaminated by noise and outliers. Traditionally, this problem is solved by local optimization methods, which usually start from an initial guess of the pose and alternately optimize the pose and the correspondence. In this paper, we formulate the problem as optimizing the pose of the 3D points in the SE(3) space to make its 2D projection best align with the 2D point set, which is measured by the cardinality of the inlier set on the 2D projection plane. We propose four geometric bounds for the position of the projection of a 3D point on the 2D projection plane and solve the 2D-3D point set registration problem by combining a global optimal rotation search and a grid search of translation. Compared with existing global optimization approaches, the proposed method utilizes a different problem formulation and more efficiently searches the translation space, which improves the registration speed. Experiments with synthetic and real data showed that the proposed approach significantly outperformed state-of-the-art local and global methods.

With the rapid development of the Internet of Things, location information becomes increasingly important for various applications. However, the localization information of network devices is vulnerable to various attacks and not always trustworthy. In this paper, we propose a location verification scheme that allows the access point (AP) to verify credibility of a reported location from a network node with assistance of an obstacle that moves actively in the network at a random speed. When the obstacle blocks the transmissions between the AP and a network node, the received signal strength (RSS) at the node is reduced. For location verification, a network node is asked to report its location and the RSS for a period of time. Based on the changes in the reported RSS and the mobility information of the obstacle, the AP can determine whether the network node has reported the correct location. An analytical model is developed to find the performance of the proposed scheme. Simulation results show that the proposed scheme achieves high probability of detecting malicious nodes and low probability of treating legitimate nodes as malicious. Simulation results have also verified the accuracy of the analysis.

Content Centric Mobile Ad-hoc NETwork (CCMANET) applies the advantages of Content Centric networking (CCN) into Mobile Ad Hoc Network(MANET) to overcome the drawbacks of low efficiency and unstable in transmission. Caching scheme is one of the key components of CCMANET. However, the caching scheme in CCMANET has not been well explored. In this paper, a novel cooperative caching scheme based on generalized dominating set and local content popularity for CCMANET is proposed. First, a virtual backbone in CCMANET is constructed by generalized dominating set to make the arbitrary topology become two-level hierarchy and a collaborative cache placement scheme between the two hierarchies is designed. Second, we propose a method of computing the local content popularity and a cache replacement scheme based on the local content popularity. Simulation results show that the proposed caching scheme can effectively reduces both path stretch value and server load, and improves the hit ratio compared with the existing caching schemes.

Securing cyberspaceCyberspace, the ubiquitous space that exists in relation to the Internet, is usually referred to as a dynamic broad domain ranging from Internet and its infrastructures to social networks.More research work in security has been extended from securing computers to securing Cyberspace, which includes the physical-level security, the network-level security, and the application-level security and addresses improvements in Cyberspace management.As a result, recent years have witnessed increasing research attention on securing Cyberspace, and many interesting methods have been proposed to locate suspicious IP, detect gossip content, prevent illegal information publication and distribution, manage social software and applications, and profile user behavior and opinion.This trend has provided the motivation to launch this special issue.Based on an open call in this area and invited best papers from The Fifth International Conference on Applications and Techniques for Information Security (ATIS 2014) and The first International Workshop on Curbing Cyber-Crimes (C 3 2014), five submissions have been accepted to best illustrate the main development and perspectives.The papers in this issue report a variety of methods used to tackle the security issues in cyberspace.They aim at improving security in applications ranging from RFID systems, location based service, discovery of software vulnerability to private medical records, and outsourcing in Multi-Cloud.The problems discussed in these papers are also related to disciplines including data mining, network security, digital forensics, and behavioral and psychological sciences.Here, we provide an integrative perspective of this special issue by summarizing each contribution contained therein.In [1], to address security and privacy issue in RFID systems, a new off-line reading orderindependent grouping-proof protocol is proposed to generate a proof that a group of tags have been scanned simultaneously in the range of a reader.The proposed protocol defines an ideal groupingproof functionality aiming at capturing the secure grouping-proof generation for a group of RFID tags in the UC framework.The new protocol maintains its security properties when composed concurrently with an unbounded number of instances of arbitrary protocol.In addition, the protocol conforms to the computational constraints of EPC Class-Gen-2 passive RFID tags.It is suitable for low-cost passive RFID tags, which are widely used in practical applications.In [2], the authors proposed an algorithm to address the problem of preserving privacy for individual users in location-aware applications.They define a novel distance measurement that combines the semantic and Euclidean distance to address the privacy-preserving issue.They conduct performance experiments on the proposed algorithm and distance metric, and results suggest that they can successfully retain the utility of the location services.In [3], to discover software vulnerability, an effective and efficient mechanism is proposed.The method also helps programmers to write secure code to avoid the existence of vulnerability at the early stage of software development.The proposed mechanism uses code clone verification to discover vulnerability in software programs and reduces the false positive of detection by combining the advantages of static and dynamic analysis.In addition, it also mitigates the path explosion problem in the testing process when verifying the existence of vulnerability.As a result, the proposed approach effectively improves the security of software systems, applications, and utilities in various areas of Cyberspace.In particular, it helps to create a reliable environment for the communications of all the social media participants.In [4], the authors analyze the security of Fair Remote Retrieval (FRR) model that is used to ensure the integrity of remote medical records.They show that FRR model fails to achieve its security goals, therefore present an improved protocol, called IFR 2, to fix the security minor faults

In LTE system, HTTP streaming services sometimes experience video quality deterioration because of the constraint of available wireless resources, which leads to reduce of Quality of Experience (QoE). In this paper, we propose a new scheduling scheme which is QoE-oriented from user's perceptive to improve the performance of HTTP streaming service in a LTE system. First of all, we adopt the jerkiness, frame freezing for example, perceived at an end user as the prominent QoE factor and implement a jerkiness measuring algorithm on end devices. Secondly, we implement a prioritized traffic flow scheduling algorithm at the base station, and put the end users who experience the maximum jerkiness to be scheduled with the highest priority. Finally, we compare the performance of the proposed QoE-oriented scheme with the traditional scheduling algorithm (i.e. Round Robin scheduling algorithm) in terms of jerkiness for HTTP streaming service. Simulation result shows that the proposed scheme can dramatically improve QoE from user's perspective by effectively decrease the jerkiness of HTTP streaming video.

Information-Centric Networking (ICN) is a new emerging concept, in which the principal paradigm shifts from the traditional end-to-end connection to the information-centric communication model. In ICN, information unit, such as content or service, is distributed in different sites or data centers to provide large-scale services. A common supporting approach for scalable service provisioning is deploying multiple replica servers throughout the network. Accordingly, an efficient and flexible scheme is needed to direct distributed requests to an appropriate replica server. In this paper, we first propose an incrementally deployable ICN architecture based on edge/core separation. And then, a practical anycast-based service provisioning scheme is presented with joint considerations of both servers contexts and the underlying network conditions. Extensive experiments have been performed to evaluate the proposed scheme. Experiment results show that efficient context-aware distributed service provisioning can be achieved.

Information-Centric Networking (ICN) enables caching of addressable content chunks in every cache-equipped router. So it is crucial to make the cached content to be visible locally, in order to satisfy the subsequent request. We achieve this goal in this paper with a novel caching scheme, which combines the content placement with dynamic request routing by selectively creating trails along the content delivery path. In addition, in-network caching requires search or replacement of cached chunks at line-speed, which resorts to the low complexity collaborative caching scheme and limits the cache size at each network level. Coordinate with the feasible cache size allocation, the proposed scheme automatically distributes the content to the proper cache location according to its popularity. Therefore, the proposed self assembly caching scheme could reduce caching redundancy and in turn, make more efficient utilization of available cache resources. The simulation results show that the proposed scheme outperforms existing algorithms.

Cooperation among mobile terminals (MTs) in heterogeneous wireless networks is currently widely investigated. Cooperation among MTs has a lot of advantages such as increasing network throughput, decreasing file download time, saving energy and so on. Since different cooperate MTs have different influence on quality of service and user's experience, MT selection becomes a key issue in MTs cooperation. To improve the quality of service and user's experience, we propose a novel mechanism for MT selection which we call NMC (Network MTs Cooperation) mechanism in this paper. In NMC mechanism, MTs and its access network are integrated as a virtual MT. Then we propose a Cooperative Terminal Selection (CTS) algorithm to select the optimal virtual MT. Simulation results show that NMC can help to save energy consumption of MTs as well as ensure the download rates.

This paper addresses how to improve the energy efficiency of OFDMA femtocell networks with sleep mode. Firstly, the energy consumption of the sleep mode is analyzed. Considering the requirements of improving the energy efficiency and ensuring the QoS of users, we extend the definition of effective capacity to formulate the tradeoff relationship between the energy efficiency and the waiting delay of users. Then two optimization schemes of sleep mode parameter are proposed: (1) Maximize the energy efficiency with effective capacity constraint; and (2) Maximize the effective capacity with energy efficiency constraint. And the performances are analyzed, including average energy consumption, energy efficiency, effective capacity and waiting time. Simulation results show that the effective capacity-based parameter optimization of sleep mode can achieve a good tradeoff between energy efficiency and QoS of users in OFDMA femtocell networks.