Human cognitive systems excel at making approximate decisions in complex and uncertain environments, a capability particularly evident in visual perception. This inherent fuzzy decision-making ability has profound implications for multimodal image fusion, where the scarcity of ground-truth data for infrared and visible light integration presents a fundamental challenge for traditional deep learning approaches. Here we introduce a zero-shot transformer architecture that mirrors human cognitive flexibility by implementing fuzzy decision-making mechanisms for pixel-level fusion weight determination. Our approach circumvents the limitations of conventional pretraining requirements through a sparse attention mechanism that selectively preserves only 15-20% of the most salient cross-modal interactions, effectively filtering out redundancy and noise. To address the computational challenges of high-resolution data, we incorporate low-rank approximation techniques that reduce the complexity from quadratic to linear, capturing over 95% of cross-modal information using a projection dimension of merely 128. The method demonstrates remarkable stability across multiple benchmark datasets, achieving a peak signal-to-noise ratio of 65.904 dB under diverse environmental conditions. Our findings challenge the prevailing assumption that dense attention patterns are essential for effective feature integration, revealing instead that multimodal fusion inherently operates in a low-dimensional space. This biomimetic approach to zero-shot learning not only advances our understanding of cross-modal feature interactions but also provides a more generalizable framework for real-world applications where ground-truth data is scarce.

In the industrial Internet of Things (IIoT), edge computing is a pivotal power in enhancing system efficiency and responsiveness. However, traditional edge computing faces some challenges like poor flexibility in communication and susceptibility to blockages. Autonomous aerial vehicles (AAVs)-assisted edge computing can address these challenges due to their flexible deployment and strong Line of Sight (LoS) link capabilities. But it also confronts challenges like signal attenuation and resource constraints. To solve these problems, reconfigurable intelligent surface (RIS) emerges as a promising integration strategy to enhance network communication and computing capabilities. Integrating AAVs and RISs in complex dynamic edge computing system poses a notable challenge in achieving real-time and efficient decision-making. Digital twin (DT) technology is an advanced technology that establishes real-time mapping and interaction between the physical world and virtual models, thereby providing real-time status monitoring and precise offloading decisions for the system. Therefore, this article considers a novel DT-driven edge computing system supported by AAVs equipped with RIS in IIoT. In this system, we focus on the intelligent computation offloading problem, whose objective is to minimize the maximum execution time across all user devices (UDs). To tackle this nonconvex mixed-integer nonlinear optimization problem, we decompose it into the scheduling and offloading optimization problem and the allocation optimization problem. Then, we first propose a multitask reinforcement learning algorithm to solve the scheduling and offloading optimization problem by optimizing the AAV trajectories, UD offloading choices, and RIS phase shifts. Afterward, based on the solution of the scheduling and offloading optimization problem, we propose an alternating iterative algorithm to address the allocation optimization problem through optimizing the offloading ratio and resource allocation. Finally, through extensive simulation experiments, we validate the effectiveness and feasibility of our proposed solution.

Recently, there has been a growing interest in using chargers to extend the operational longevity of UAVs (drones). In this paper, we explore a charger-assisted drone application where a drone observes points of interest while a mobile charger moves to recharge its battery. We focus on the route and charging schedule of the drone and mobile charger to maximize observation utility in the shortest possible time, while ensuring continuous drone operation. In our problem, the drone and mobile charger cooperate to complete a task. Their discrete-continuous hybrid actions pose a major computational challenge. To address this issue, we present a hybrid-action deep reinforcement learning framework, called HaDMC, which uses a typical policy learning algorithm to generate latent continuous actions. We specifically design and train an action decoder. It involves two pipelines to convert the latent continuous actions into the original hybrid actions for the drone and mobile charger to directly interact with environment. We incorporate a mutual learning scheme into model training, emphasizing collaboration over individual actions. By extensive numerical experiments, we evaluate HaDMC and compare it with state-of-the-art approaches. The experimental results demonstrate the effectiveness and efficiency of our solution.

For large-scale LoRa-based Internet-of-Things (IoT) systems, using UAVs to collect data is a promising method, which not only leverages the long-range communication advantages of LoRa but also avoids the high costs associated with deploying fixed-location LoRa gateways. In this paper, we take into account the concurrent data reception capability inherent to LoRa, and propose an innovative algorithm that provides a bi-criteria approximation solution for UAV-assisted LoRa data collection, maximizing the sensor coverage under the UAV's energy budget. Our algorithm is twofold: first, we formulate the problem within a constrained submodular maximization framework incorporating sensor allocation and a Traveling Salesman Problem. Second, proving the NP-hardness of this sensor allocation problem, we develop a constant-factor approximation algorithm for energy-minimum sensor allocation. We present a bi-criteria approximation algorithm that employs a greedy strategy to maximize sensor coverage while adhering to the UAV's energy budget. We evaluate our designs through extensive numerical experiments, demonstrating their efficiency and effectiveness.

Mobile crowdsourcing (MCS) is a promising way to monitor urban-scale data by leveraging the crowds’ power and has attracted much attention recently. How to recruit suitable workers for requesters to perform the published sensing tasks is always a crucial problem and also a research hotspot. Many attempts have been made in past literature to maximize social welfare or to motivate workers to participate in the mobile crowdsourcing (MCS). However, most existing works do not consider the individual sensing quality requirements of tasks, which may not be suitable for some special scenarios, such as monitoring tasks of locations with different importance levels. In this work, we investigate the optimal worker selection problem for collaborative MCS, in which we study the recruitment cost minimization problem to meet individual sensing quality requirements of tasks for the requester-centric MCS, as well as the profit maximization problem for the platform-centric MCS. Both of the studied problems are proved to be NP-hard, and thus we design corresponding approximation algorithms for them. Specifically, to solve the recruitment cost minimization problem for requester-centric MCS, we design two different polynomial time algorithms, both of which have performance guarantees. For the profit maximization problem for platform-centric MCS, we introduce a double-greedy-based algorithm and then use the iterative pruning technique to ensure the performance guarantee of our algorithm with a much weaker condition. Finally, we evaluate our algorithms through numerical simulation experiments and validate the effectiveness of our designs by comparing them with baselines under different parameter settings.

Smart contract is the core of blockchain operation, and contract vulnerability will cause huge economic losses. Therefore, effective smart contract vulnerability detection is of vital importance and attracts more and more attention. In this paper, we propose a vulnerability detection model (VDM-AEI) based on automatic feature extraction and feature interaction. For the first time, this model converts smart contracts into gray images and uses VGG16 and GRU models to automatically extract vulnerability features and filter effective features, respectively. Then, a contract graph and an expert knowledge feature vector are constructed by using commonly used methods as part of feature construction. Next, AutoInt and DCN networks are used to build a dual feature interaction network to obtain more abundant vulnerability feature information, which extracts high-dimensional nonlinear features from the low and sparse features of the contract graph feature vector and the expert knowledge-defined feature vector. Finally, all ouput features of GRU, AutoInt and DCN networks are integrated to obtain vulnerability classification results through fully connected neural networks. We conducted extensive experiments on the ESC and VSC datasets for reentrancy vulnerabilities, timestamp dependency vulnerabilities, and infinite loop vulnerabilities. The experimental results prove the effectiveness and accuracy of the VDM-AEI model. Compared with the latest vulnerability detection model CGE, the accuracy rates of the 3 types of vulnerability detection are improved by 10.85%, 6.18%, and 12.34%, respectively. In addition, the predicted F1 scores of VDM-AEI are all greater than 95%, and the recall rate is no less than 94%.

As one of the basic protocols of the wireless sensor network (WSN), neighbor discovery aims at initializing the network topology and maintaining the runtime connectivity. It is a quite big challenge to reduce the neighbor discovery latency in low-duty-cycle WSN, because the long dormancy of nodes handicaps the quick neighborhood establishment in their vicinity. Recently preliminary efforts have been directed toward the referring-based neighbor discovery, which proactively refers or recommends potential neighbors to nearby nodes in order to reduce the total discovery latency. Without delicate considerations, however, such proactive references will incur a great waste of node energy. In response to this limitation, we design and implement xBOND, a novel referring-based neighbor discovery protocol for the WSN with low-duty cycles. The xBOND protocol involves four perspectives of featuring the physical proximity of nodes. These features can be easily evaluated, and thus the node willing to do neighbor recommendation does not need to mull over them. With the features evaluated, xBOND leverages the XGBoost (a powerful classifier in machine learning) to achieve precise neighbor recommendations during neighbor discovery and then a better tradeoff between energy overhead and discovery latency. We also concretize the xBOND so that it can complete the neighbor discovery in a distributed way. Finally, we validate xBOND through extensive simulation experiments that show significant efficiency gains in discovery latency and energy consumption over the state-of-the-art referring-based approach.

Mobile crowd sensing (MCS) is a new paradigm for urban-scale monitoring. This article concentrates on the Cost-Fair Task Allocation (CFA) problem for the MCS scenario where the collaboration of multiple probabilistic mobilephone users is needed to yield more reliable observation. CFA aims to allocate sensing tasks to users so that the sensing costs undertaken by all users are as balancing as possible, while the requirement of the requester for data reliability can be satisfied. CFA is greatly important to MCS campaigns in terms of reliability and sustainability. We design two algorithms to solve the CFA problem in the offline and online cases, respectively. Specifically, we propose a novel penalty-based model to reformulate the offline CFA problem, and based on this model, we design an offline algorithm, which can yield a computation-efficient ε-solution with any small ε > 0. For the online case, we design a polynomial-time approximation algorithm, which struggles to allocate each of the sequentially arriving tasks to users as fairly as possible, and can achieve an upper-bounded competitiveness relative to the optimal CFA solution. Finally, we conduct extensive numeric analyses to validate the performance of our algorithms under diverse experimental setups.

As a promising technology to achieve perpetual operation of battery-powered wireless sensor devices, wireless power transfer has attracted much attention recently. In wireless power transfer, the charger enables the energy to be wirelessly transmitted to the rechargeable sensor devices that are hungry for energy. Previous works mainly focus on maximizing the charging utility or minimizing the charging delay. This paper concerns a more practical issue of placing wireless chargers, which aims at minimizing the deployment cost of chargers while satisfying the overall requirement for charging utility. We investigate the above cost-minimum charger placement problem under two typical scenarios in which omni chargers and directional chargers are used, respectively. To resolve this problem under the two charging models, we first prove its NP-hardness and then propose two approximation algorithms with proven performance guarantees. Finally, we conduct extensive simulation experiments to validate our designs, and the experimental results demonstrate that the proposed algorithms significantly outperform the baselines.

Mobile crowdsensing (MCS) is a new and promising tool in urban sensing. It exploits a crowd of smartphone-carried mobile users and transfers their sensory data to requesters who usually publish spatio-temporal tasks of sensing city area. In reality, mobile users can probabilistically move in the sensing region in their daily mobility and stay there for a period of time; and then these probabilistic users can be recruited to collaboratively perform MCS sensing tasks. Such an MCS depending on the probabilistic collaboration of mobile users is usually called nondeterministic MCS. In this paper, we focus on the budget-feasible user recruitment (BFUR) problem in non-deterministic MCS, which is the first work to maximize the requester's utility under a given budget constraint. Because of the NP-hardness of BFUR, we reformulate it as a monotone submodular maximization problem and propose a greedy algorithm (called uMax) with provable constant-factor competitiveness. Unlike previous works for nondeterministic MCS, however, this paper specially puts effort on predicting the mobility patterns of users, especially their stay time in requester's sensing region, and then designs an effective predictor based on bi-directional long short-term memory neural network. Such a prediction of user's stay time not only connects the BFUR problem modeling defined in this paper and the actual mobility uncertainty of users, but also can apply to any nondeterministic MCS campaign that depends on the knowledge of user's stay patterns. We finally validate the performance of the proposed predictor under a real-world dataset of wireless mobile networks, and evaluate algorithm uMax by comparing it with two other baseline algorithms.