For the safety and effectiveness of autonomous driving, it is crucial to accurately detect lane lines and road structures. Modern methods based on the Transformer architecture, such as Maptr, have demonstrated powerful capability for online construction of high-definition maps on GPUs. However, for practical vehicular platforms, the deployment and optimization of related models have not caught sufficient attention in practice. To address such issue, this paper proposes an integrated deployment framework, which is tailored specifically for embedded automotive systems and can overcome the limitations of isolated optimization through co-optimizing the model algorithm and the target hardware platform. Furthermore, to validate the effectiveness of the proposed framework and demonstrate its practical performance, we implemented a complete inference visualization on the MDC610 platform. This visualization pipeline can directly project the structured map elements generated by the model, such as detected lane lines, onto the original sensor data.

The swift propagation of fake news via online social networks severely undermines public trust and social harmony, compelling the development of resilient and efficient detection strategies. While prevailing methodologies primarily focus on either content-based analysis, leveraging linguistic and visual features, or context-based methods, utilizing social network dynamics, they often fail to fully integrate the semantic depth of news content with the relational complexity of its propagation. To tackle these challenges, this research introduces an innovative framework that synergizes Hypergraph Neural Networks (HGNN) with Large Language Models (LLMs) to enhance fake news detection. The framework leverages LLMs to extract intricate linguistic and semantic features from news content, while using HGNN to capture the complex, multi-faceted interactions within the news propagation network. The integration of these components fortifies the framework’ ability to discern fake news, synergizing the advantages of both content-driven and context-aware methodologies. Experiments demonstrate that HGNN significantly outperforms traditional Graph Neural Networks (GNNs) in capturing complex relational information, leading to superior robust detection performance. The integration of LLMs further improves text encoding accuracy. This research underscores the efficacy of leveraging advanced language models in conjunction with hypergraph-based architectures, providing a more resilient and holistic approach to fake news detection within social media context. Future work will focus on addressing dynamic hypergraph construction challenges and scalability improvements for real-world applications.

Distributed computing is an important topic in the field of wireless communications and networking, and its high efficiency in handling large amounts of data is particularly noteworthy. Although distributed computing benefits from its ability of processing data in parallel, the communication burden between different servers is incurred, thereby the computation process is detained. Recent researches have applied coding in distributed computing to reduce the communication burden, where repetitive computation is utilized to enable multicast opportunities so that the same coded information can be reused across different servers. To handle the computation tasks in practical heterogeneous systems, we propose a novel coding scheme to effectively mitigate the “straggling effect” in distributed computing. We assume that there are two types of servers in the system and the only difference between them is their computational capabilities, the servers with lower computational capabilities are called stragglers. Given any ratio of fast servers to slow servers and any gap of computational capabilities between them, we achieve approximately the same computation time for both fast and slow servers by assigning different amounts of computation tasks to them, thus reducing the overall computation time. Furthermore, we investigate the information-theoretic lower bound of the inter-communication load and show that the lower bound is within a constant multiplicative gap to the upper bound achieved by our scheme. Various simulations also validate the effectiveness of the proposed scheme.

Graph pooling refers to the operation that maps a set of node representations into a compact form for graph-level representation learning. However, existing graph pooling methods are limited by the power of the Weisfeiler-Lehman (WL) test in the performance of graph discrimination. In addition, these methods often suffer from hard adaptability to hyper-parameters and training instability. To address these issues, we propose Hi-PART, a simple yet effective graph neural network (GNN) framework with Hierarchical Partition Tree (HPT). In HPT, each layer is a partition of the graph with different levels of granularities that are going toward a finer grain from top to bottom. Such an exquisite structure allows us to quantify the graph structure information contained in HPT with the aid of structural information theory. Algorithmically, by employing GNNs to summarize node features into the graph feature based on HPT's hierarchical structure, Hi-PART is able to adequately leverage the graph structure information and provably goes beyond the power of the WL test. Due to the separation of HPT optimization from graph representation learning, Hi-PART involves the height of HPT as the only extra hyper-parameter and enjoys higher training stability. Empirical results on graph classification benchmarks validate the superior expressive power and generalization ability of Hi-PART compared with state-of-the-art graph pooling approaches.

Many complex networks in the real world have community structures. Typical examples include online social networks and ecology networks. While the identification of communities bears numerous practical applications, with the increasing awareness of data security and privacy concerns, the need to protect the community affiliations of individuals from disclosing by attackers emerges. This raises the community deception (CD) problem, that is, the opposite of community detection, which asks for ways to minimally perturb the network structures by rewiring nodes so that the target communities maximally hide themself from community detection algorithms. To this end, we investigate the CD problem through a Laplacian spectrum lens and propose a method named $\mathtt {ComDeceptor}$ to hide a flexible target set of communities, which is more universal than most existing methods that either focus on hiding the entire communities or a single community. The key idea of $\mathtt {ComDeceptor}$ is to first allocate the resources of perturbations fairly and effectively. By proving that hiding communities through intercommunity edge addition and intracommunity edge deletion correspond to maximizing the second smallest eigenvalue $\lambda _{2}$ and minimizing the largest eigenvalue $\lambda _{n}$ of the graph Laplacian, respectively, $\mathtt {ComDeceptor}$ then incorporates efficient heuristics for approximately solving the problems, thus selecting the appropriate edge to perturb. Experimental results over nine real-world networks and six community detection algorithms not only demonstrate the efficiency of $\mathtt {ComDeceptor}$ , but also the superior performance on obfuscating community structures over the baselines.

This article studies connectivity maintenance in uncertain networks under adversarial attack, where a defender conceals crucial links to prevent the largest connected component from being decomposed by an attacker. In contrast with its static counterpart, connectivity maintenance in uncertain networks involves additional probing on links to determine their existence. Therefore, by modeling an uncertain network as a random graph with each link associated with an existence probability and a probing cost, our goal is to design a defensive strategy for link selection that maximizes the expected size of the largest remaining connected component with the minimum expected probing cost, and moreover, the strategy should be independent of the attacking patterns. To this end, we first unravel the computational complexity of the problem by proving its NP-hardness, and then propose optimal defensive strategies based on dynamic programming and multi-objective optimization. Due to the prohibitive computational cost of optimality, two approximate defensive strategies are further designed to pursue decent performance with quasilinear complexity, in which the first one is a heuristic approach that quantifies the link vulnerability through an analogy from the degree centrality of a vertex in static networks to the connectivity weight of a link in uncertain networks, and the second one is an adaptive greedy policy incorporating the minimax rule from game theory, which minimizes the possible loss suffered by the defender in a worst-case scenario and has a constant approximation ratio. Extensive experiments on both synthetic and real-world network datasets under diverse attacking patterns demonstrate the superiority of the proposed strategies over baselines.

This paper focuses on optimally determining the existence of connected paths between some given nodes in random ring-based graphs.Serving as a fundamental underlying structure in network modeling,ring topology appears as commonplace in many realistic scenarios.Regarding this,we consider graphs composed of rings,with some possible connected paths between them.Without prior knowledge of the exact node permutations on rings,the existence of each edge can be unraveled through edge testing at a unit cost in one step.The problem examined is that of determining whether the given nodes are connected by a path or separated by a cut,with the minimum expected costs involved.Dividing the problem into different cases based on different topologies of the ring-based networks,we propose the corresponding policies that aim to quickly seek the paths between nodes.A common feature shared by all those policies is that we stick to going in the same direction during edge searching,with edge testing in each step only involving the test between the source and the node that has been tested most.The simple searching rule,interestingly,can be interpreted as a delightful property stemming from the neat structure of ring-based networks,which makes the searching process not rely on any sophisticated behaviors.We prove the optimality of the proposed policies by calculating the expected cost incurred and making a comparison with the other class of strategies.The effectiveness of the proposed policies is also verified through extensive simulations,from which we even disclose three extra intriguing findings:i) in a onering network,the cost will grow drastically with the number of designated nodes when the number is small and will grow slightly when that number is large;ii) in ring-based network,Depth First is optimal in detecting the connectivity between designated nodes;iii) the problem of multi-ring networks shares large similarity with that of two-ring networks,and a larger number of ties between rings will not influence the expected cost.

This article studies time-sensitive multicast in wireless sensor networks (WSNs) with link uncertainty, where information from the source needs to be delivered to multiple receivers within an imposed delay constraint. Prior art on static WSNs minimizes the multicast delay via the construction of a multicast tree that approximates the Steiner tree in length, which, however, may be invalidated by the time-varying network topology of WSNs with uncertain link states. Moreover, for multicast in WSNs with link uncertainty, the possible link failure necessitates a suitable measurement of the uncertain communication distance and calls for the performance guarantee in both delay and delivery ratio. In this work, by modeling a WSN as a random graph with each link associated with a transmission probability, we propose FlowerCast, an efficient multicast scheme, to jointly minimize the expected multicast delay and to maximize the expected delivery ratio of multicast under delay constraint. The core of FlowerCast is to quantify the uncertain communication distance by the expected transmission delay of a time-varying path, based on which a delay-optimal multicast tree is constructed in accordance with the directionality of delay. Candidate paths with a high expected delivery ratio and low expected delay are then selected in a distributed manner to conditionally connect adjacent multicast members and thus transform the multicast tree into a multicast flower. Despite the NP-hardness of optimal candidate paths’ addition, the transformation with the highest expected delivery ratio of multicast under delay constraint can be guaranteed through a pseudo-polynomial time derandomization-based greedy approach. We further demonstrate the time and energy efficiency of FlowerCast through asymptotic analysis. To make full use of the possible overlapping links in a multicast flower, a hybrid routing strategy is presented to wisely switch between sequential routing and synchronous routing for extra enhancement of the multicast performance. Extensive experiments on various datasets verify the superiority of FlowerCast and hybrid routing over baselines and indicate their wide applicability to practical scenarios.

The popularity of intelligent devices provides straightforward access to the Internet and online social networks. However, the quick and easy data updates from networks also benefit the risk spreading, such as rumor, malware, or computer viruses. To this end, this article studies the problem of source detection, which is to infer the source node out of an aftermath of a cascade, that is, the observed infected graph GN of the network at some time. Prior arts have adopted various statistical quantities such as degree, distance, or infection size to reflect the structural centrality of the source. In this article, we propose a new metric that we call the infected tree entropy (ITE), to utilize richer underlying structural features for source detection. Our idea of ITE is inspired by the conception of structural entropy [21], which demonstrated that the minimization of average bits to encode the network structures with different partitions is the principle for detecting the natural or true structures in real-world networks. Accordingly, our proposed ITE based estimator for the source tries to minimize the coding of network partitions brought by the infected tree rooted at all the potential sources, thus minimizing the structural deviation between the cascades from the potential sources and the actual infection process included in GN. On polynomially growing geometric trees, with increasing tree heterogeneity, the ITE estimator remarkably yields more reliable detection under only moderate infection sizes, and returns an asymptotically complete detection. In contrast, for regular expanding trees, we still observe guaranteed detection probability of ITE estimator even with an infinite infection size, thanks to the degree regularity property. We also algorithmically realize the ITE based detection that enjoys linear time complexity via a message-passing scheme, and further extend it to general graphs. Extensive experiments on synthetic and real datasets confirm the superiority of ITE to the baselines. For example, ITE returns an accuracy of 85%, ranking the source among the top 10%, far exceeding 55% of the classic algorithm on scale-free networks.

This paper proposes a novel one-stage detection method with cascade bounding box refinement, called CasDet. Generally one-stage detection methods generate less accurate bounding boxes than two-stage methods. Motivated by the location refinement of bounding boxes of two-stage methods, we present a cascade bounding box refinement for one-stage detection networks. By refining the detection results for multiple times, CasDet can generate high quality bounding boxes. To further improve the feature representation, we present a multi-scale feature aggregation structure (MBFA). By fusing features with adjacent three scales and adding multiple top-down and bottom-up aggregation paths, MBFA introduces rich contextual information and local details into each detection feature. The experimental results on COCO benchmark confirm that our CasDet outperforms existing state-of-the-art detection methods and achieves a better balance between speed and accuracy.

Self-supervised graph-level representation learning has recently received considerable attention. Given varied input distributions, jointly learning graphs’ unique and common features is vital to downstream tasks. Inspired by graph contrastive learning (GCL), which targets maximizing the agreement between graph representations from different views, we propose an Adaptive self-supervised framework, Ada-MIP, considering both Mutual Information between views (unique features) and inter-graph Proximity (common features). Specifically, Ada-MIP learns graphs’ unique information through a learnable and probably injective augmenter, which can acquire more adaptive views compared to the augmentation strategies applied by existing GCL methods; to learn graphs’ common information, we employ graph kernels to calculate graphs’ proximity and learn graph representations among which the precomputed proximity is preserved. By sharing a global encoder, graphs’ unique and common information can be well integrated into the graph representations learned by Ada-MIP. Ada-MIP is also extendable to semi-supervised scenarios, with our experiments confirming its superior performance in both unsupervised and semi-supervised tasks.

Given an aftermath of an information spreading, i.e., an infected network G N after the propagation of malicious rumors, malware or viruses, how can we identify the sources of the cascade? Answering this problem, which is known as the multiple source detection (MSD) problem, is critical whether for forensic use or insights to prevent future epidemics.Despite the recent considerable effort, most of them are built on a preset propagation model, which limits their application range. Some attempts aim to break this limitation via a label propagation scheme where the nodes surrounded by a large proportion of infected nodes are highlighted. Nonetheless, the detection accuracy may suffer since the node labels are simply integers with all infected or uninfected nodes sharing the same initialization setting respectively, which fall short of sufficiently distinguishing their structural properties. To this end, we propose a community based label propagation (CLP) framework that locates multiple sources through exploiting the community structures formed by infected subgraphs of different sources. Besides, CLP tries to enhance the detection accuracy by incorporating node prominence and exoneration effects, namely the nodes surrounded by larger proportions of infected nodes are more likely to be sources, and the uninfected nodes or infected ones in neighboring community hold the key in exonerating an infected node from being the source. As such, CLP is applicable in more propagation models, and is provably convergent. Experiments on both synthetic and real-world networks further validate the superiority of CLP to the state-of-the-art, boosting the F-score from few percents to approximately 68× in large-scale networks.

Single image super-resolution (SISR) is of great importance and has found wide applications. In recent years, convolutional neural network (CNN) has been implemented to SISR and achieved great successes. However, many existing CNN-based SR networks improve the performance by stacking a large number of blocks, resulting in more parameters and higher computation cost. To solve this problem, we propose a lightweight dynamic hybrid attention network (DHAN) which is made up of chained dynamic hybrid attention blocks (DHABs). A DHAB is mainly composed of a non-attention branch and a hybrid attention branch based on 3D convolution kernels, and can combine channel attention and spatial attention to extract powerful representations of inter-channel and intra-channel feature response. Furthermore, we design a dynamic weight module of DHAB, which generates two sum-to-one weights according to the input features of each block and dynamically adjusts the contribution of the hybrid attention branch and the non-attention branch. Based on the hybrid attention branch and the dynamic weight module, DHAN can not only efficiently reduce the number of multiply-add operations (Multi-Adds) and the number of parameters, but also accurately recover the textural details of the high-resolution image from the low-resolution feature map. Experiments on benchmark datasets are done to demonstrate that the proposed DHAN achieves excellent performance compared to some state-of-the-art networks of similar sizes.

He rapid development of the Internet has brought about exponential growth in the amount of data in the network. With the growth of data volume in the network and the increasing dependence of humans on the Internet, network intrusion has become increasingly frequent, posing a serious threat to network security. In view of this, network intrusion detection has gradually become an important research topic in the field of network security. Due to the increasing maturity of deep learning technology, a large amount of intrusion detection research is currently based on deep learning technology. Deep learning technology can certainly improve the accuracy of intrusion detection, but it requires a large amount of data and computing resources and can easily cause overfitting when the data volume is insufficient, thereby reducing the recognition rate of unknown intrusions. This paper proposes an intrusion detection algorithm based on Markov chain Monte Carlo method, which can complete network intrusion detection tasks with low resource consumption. Compared with the intrusion detection algorithm based on basic Extreme learning machine, this method has better detection performance.

Panoramic image stitching is widely implemented to many application scenarios, such as satellite remote sensing detection, street view mapping and medical image generation. In this paper, we study two processes that have great impact on the visual effects of the stitched images and are the most concerned by researchers, namely image registration and stitching seam removal. In contrast to the traditional panoramic image stitching algorithms that use only one global projection transformation matrix for image alignment, this paper proposes a spatially-varying deformation method that uses multiple local transformation matrices, including the APAP (As-Projective-As-Possible) method that can effectively reduce the image alignment error and geometric transformation distortion, the SPHP (Shape-Preserving Half-Projective) method that can combine good alignment accuracy and reduced projection distortion, and the AANAP (Adaptive As-Natural-As-Possible) method that can improve the perspective of non-overlapping regions of the stitched image and good accuracy. The image stitching performance can be well improved.

Driver distraction and fatigue detection systems can effectively reduce car accidents and ensure the safety of traffic participants. Most of the existing vision-based approaches use facial landmarks as driver’s states. However, facial landmark detection is inaccurate under the large angle of head posture, which impacts the accuracy of further processing. This paper presents an effective method using convolution neural networks (CNNs). The method firstly deploys a modified MTCNN to detect the face region. Then, a lightweight multi-task CNN is proposed to detect eye regions, mouth landmarks and 3D head pose, and a simple CNN is used to detect eye closure independently. An angle-adapted loss function is applied to improve the landmark detection accuracy under the large posture. Finally, multiple abnormal behaviors are recognized to determine distraction and fatigue driving. Experiments show that our proposed method is superior to existing methods in both accuracy and running speed.

In modern information retrieval systems, the thesaurus is playing an increasingly important role. In order to better describe and analyze the domain news, this paper proposes a method of domain keyword extraction, and further constructs an effective domain thesaurus. Compared with the previous research, this paper grasps the core information in the field by extracting and combining domain keywords, and improves the domain effectiveness of the thesaurus. In addition, this paper conducts both manual analysis and automated processing to construct high-quality thesaurus, which has practical application value. The final results provide support for the process of indexing, organizing, retrieving and recommending news.

This paper studies the event-triggered quantized consensus problem for discrete-time scalar linear multi-agent systems. Due to the limited communication resources, the information exchanged among neighboring agents should be quantized into finite bits and transmitted over an unreliable digital network suffers from dropouts. The existence of nondeterministic packet losses exacerbates the effect of quantization error and damages the consensus performance. To handle these issues, an event-triggering strategy including a local event-triggering function for each agent and a dynamic quantizing-updating method is designed. It is theoretically proved that the estimation error is always bounded by an enlarged event-triggering function and the asymptotic consensus of concerned systems can be guaranteed at a finite bit rate. A simulation example is provided to confirm the effectiveness of our event-triggered control protocol. It shows that the communication frequency among neighbors can be reduced a lot while guaranteeing the system performance.

Fault diagnosis of solar panels is essential for production capacity and safety of solar energy, and has caught considerable attention. This paper presents an efficient method based on a convolution neural network (CNN) to detect solar panels' edges in infrared images. An accurate classifier is proposed to identify defective solar panels with the obtained obtained edges of solar panels. The location and classification algorithms of solar panels with faults are realized through a lightweight CNN, and implemented on embedded devices. With many images collected from different photovoltaic plants, we verified the effectiveness of the proposed fault diagnosis algorithms.

Solar energy is the most techno-economically viable renewable source of energy and can be effectively converted into electrical power by Photovoltaic (PV) systems. As partial shading (PS) may reduce the harvested energy of PV systems, this paper proposes a novel tunicate swarm algorithm (TSA) based MPPT (maximum power point tracking) strategy to tackle the PS issue. More specifically, the simple and effective modeling of TSA with a search and skipping (SAS) scheme is utilized to minimize the tracking time and search area. The SAS scheme can discard the voltage range lacking global maximum power point (GMPP) and significantly reduce computation time. Consequently, power tracking, tracking time, and robustness can be greatly enhanced. The performance of the proposed TSA strategy is comprehensively analyzed against state-of-the-art techniques, including incremental conductance (InC), improved particle swarm optimization (IPSO), grey wolf optimization (GWO), and cuckoo search algorithm (CSA), through detailed case studies, which include standard array configurations, PS conditions, varying irradiance patterns, fast-changing temperature, and the field atmospheric data. TSA is further validated on a low-cost hardware setup, confirming its superior performance. The results provide insightful validation of the practicality of the proposed TSA strategy in the real-world applications.