Abstract To address the problem of high network energy consumption and imbalanced load distribution in data transmission, this paper proposes a novel deep reinforcement learning algorithm to jointly optimize energy consumption and link load distribution. We design a three-layer back-propagation neural network. It iteratively adjusts weights based on past states and actions, and minimizes the loss to improve action prediction accuracy. It guides the agent to make reasonable routing decisions within complex network environments. Based on this, we leverage Q-learning to seek paths for transmission demands. By dynamically aggregating and balancing traffic, energy efficiency and load balance can be achieved. We design reward functions from the viewpoint of link and node, respectively, for different optimization objectives. In order to obtain high efficiency and good robustness, we improve roulette-based Chebyshev scalarization function to solve the weight selection problem among multi-objectives. We update the Pareto set via multiple state transitions to approximate optimal solution. We use the Euclidean distance to measure optimizing effect of both objectives. Simulation results illustrate that our algorithm can effectively reduce network energy consumption and balance load distribution.

Crowd counting plays a crucial role in applications like public safety and smart cities, where estimating the number of people in images or videos is essential. However, since most of the current crowd counting datasets are visible light images, the counting performance is limited by the light intensity. To tackle this problem, we collect a drone based thermal infrared crowd counting dataset named LYU-DroneInfrared, which includes 64,210 images and 2,997,352 head annotation points, covering different scenes such as schools, streets, squares, sports grounds, etc. In addition, we propose the IncepTNet, an inception transformer network based on the transformer architecture. It consists of two parts: low frequency feature extraction and high frequency feature extraction. In low-frequency feature extraction, average pooling is first applied, followed by multi-head self-attention to capture contextual information in the image. On the other hand, to be able to pay more attention to the fine-grained details of the images, the parallel approach of convolution and maximum pooling are used to extract the high frequency features. The proposed method is validated on two benchmark datasets JHU-Crowd++ and NWPU-Crowd, and a newly collected LYU-DroneInfrared dataset. Extensive experimental results have shown that the IncepTNet method exhibits excellent crowd counting performance on different types of datasets.

Very-High-Resolution (VHR) remote sensing images are characterized by extremely high spatial resolution, incorporating higher pixel density and larger image sizes, which pose challenges for existing methods to extract complex texture features. Furthermore, due to the wide-area and high-resolution imaging strategy, VHR change detection images suffer from a severe imbalance between change pixels and non-change pixels, increasing the difficulty of handling change detection tasks. To address these challenges, we introduced the Omnidirectional Selective Scan Module (OSSM), which has the capability to process long sequences. By integrating it with the lightweight Siamese Feature Pyramid Network (SFPN), we designed a hybrid CNN-Mamba backbone, referred to as SPMamba. This backbone captures both global and local information within bitemporal feature maps at each stage, enhancing the precision of texture feature extraction. Additionally, to integrate the semantic features from each branch in SPMamba and reduce noise interference from non-target change areas, we developed a Hybrid Fusion Module (HFM). The HFM consists of two fusion modules: the High-Low Channel Fusion Module (HLM) and the Bilateral Channel Fusion Module (BCM), which facilitates both feature-level and channel-level integration, enhancing the sensitivity of the model to subtle changes. Extensive experimental results demonstrate that SPMNet achieves the highest F1-score of 91.80%, 90.99%, and 96.04% on the WHU-CD, LEVIR-CD, and CDD-CD datasets, respectively, outperforming eleven state-of- the-art methods. Moreover, the effects of varying image sizes on model training are thoroughly analyzed.

Drones have been widely used in a variety of applications, e.g., aerial photography and military security, because of their high maneuverability and broad views compared with fixed cameras. Multi-drone tracking systems can provide rich information about targets by collecting complementary video clips from different views, especially when targets are occluded or disappear in some views. However, it is challenging to handle cross-drone information interaction and multi-drone information fusion in multi-drone visual tracking. Recently, Transformer has shown significant advantages in automatically modeling the correlation between templates and search regions for visual tracking. To leverage its potential in multi-drone tracking, we propose a novel cross-drone Transformer network (TransMDOT) for visual object tracking tasks. The self-attention mechanism is used to automatically capture the correlation between multiple templates and the corresponding search region to achieve multi-drone feature fusion. During the tracking process, a cross-drone mapping mechanism is proposed by using the surrounding information of the drone with promising tracking status as reference, assisting drones that lost targets to re-calibrate, which implements real-time cross-drone information interaction. As the existing multi-drone evaluation metrics only consider spatial information while ignore temporal information, we further present a system perception index (SPFI) that combines both temporal and spatial information to evaluate the tracking status of multiple drones. Experiments on the MDOT dataset prove that TransMDOT greatly surpasses the state-of-the-art methods in both single-drone performance and multi-drone system fusion performance. Our code will be available onhttps://github.com/cgjacklin/transmdot.

Extracting domain knowledge from databases can facilitate the development of Web ontologies. In this paper, a formal approach and an automated tool for constructing ontologies from Object-oriented database models (OODMs) are developed. The approach and tool can automatically translate an OODM and its corresponding database instances into the ontology structure and ontology instances, respectively. Case studies show that the approach is feasible and the automated construction tool is efficient.

Rules have been playing an increasingly important role in the Semantic Web. However, general rule languages are not capable of representing much imprecise and uncertain knowledge in the Semantic Web, nor are if-then rules. Combining if-then rules with OWL DL in the framework of fuzzy sets and possibility distribution, we propose fuzzy Semantic Web if-then Rule Language (f-SW-if-then-RL), and investigate its syntax and semantics. Considering nonmonotonicity, we employ unless rules to extend f-SW-if-then-RL, and f-SW-if-then-unless-RL appears. Two kinds of negation are introduced to express semantics of unless rules. And we extend rule interchange format R2ML to encode nonmonotonic fuzzy rules.

To deal with the problem of too many results returned from a Web database in response to a user query, this paper proposes a novel approach, which takes advantage of the contextual preferences to precompute a few representative orders of tuples and uses them to expeditiously provide ranked answers factoring in the information contained in the query. Contextual preferences take the form that item i1 is preferred to item i2 with an interest degree in the context of X. This paper formally defines contextual preferences, provides algorithms for creating tuple orders, clustering orders and processing queries, and presents experimental results to show their efficiency.

How to quickly and cheaply construct Web ontologies has become a key technology to enable the Semantic Web. However, information imprecision and uncertainty exist in many real-world applications. Thus constructing fuzzy ontology by extracting domain knowledge from fuzzy database model such as fuzzy ER model can profitably support fuzzy ontology development. In this paper, firstly, we give the formal definition and semantics of fuzzy ER model. Then, we introduce a kind of fuzzy extension of OWL DL, named fuzzy OWL DL. Furthermore, based on the fuzzy OWL DL, the formal definition and Model-Theoretic semantics of fuzzy OWL DL ontology are given. Whatpsilas more, we realize the formal translation from fuzzy ER model to fuzzy OWL DL ontology by a semantics-preserving translation algorithm. Finally, since a fuzzy OWL DL ontology is being equivalent to a description logic f-SHOIN(D) knowledge base, the reasoning problem of satisfiability, subsumption, and redundancy of fuzzy ER model may reason automatically through reasoning mechanism of f-SHOIN(D) is also investigated, which can contribute to constructing fuzzy OWL DL ontologys exactly that meet applicationpsilas needs.