Abstract Aiming at the problems of incomplete dehazing of a single image and unnaturalness of the restored image, a multi‐scale single‐image defogging network with local features fused with global features is proposed, using fog and non‐fogging image pairs train the network in a direct end‐to‐end manner. The network is divided into global feature extraction module, multi‐scale feature extraction module and deep fusion module. The global feature extraction module extracts global features that characterize the contour; multi‐scale feature extraction module extracts features at different scales to improve learning accuracy; in the deep fusion module, the convolutional layer extracts the local features that describe the image content, and then the local features and the global features are merged through skip connections. Comparative experiments were carried out on artificially synthesized fog images and real fog images. The experimental results show that the algorithm proposed here can achieve the ideal dehazing effect, and is superior to other comparison algorithms in subjective and objective aspects.
A safe and reliable task planning method is a prerequisite for the collaborative execution of ocean observation data collection tasks by multiple unmanned surface vessels (multi-USVs). Deep Reinforcement Learning (DRL) combines the powerful nonlinear function-fitting capabilities of deep neural networks with the decision-making and control abilities of reinforcement learning, providing a novel approach to solving the multi-USV task planning problem. However, when applied to the field of multi-USV task planning, it faces challenges such as a vast exploration space, extended training times, and unstable training process. To this end, this paper proposes a multi-USV task planning method based on improved deep reinforcement learning. The proposed method draws on the idea of a value decomposition network, breaking down the multi-USV task planning problem into two subproblems: task allocation and autonomous collision avoidance. Different state spaces, action spaces, and reward functions are designed for the various subproblems. Based on this, a deep neural network is used to map the state space of each subproblem to the action space of each USV, and the generated strategy of the deep neural network is assessed based on the corresponding reward function. This successfully integrates task allocation and path planning into a comprehensive task planning framework. Deep neural networks consist of the Actor networks and the Critic networks. During the training phase of the Critic network, different methods are used to train different Critic networks to improve the convergence speed of the algorithm. An improved temporal difference error method is specifically applied to train the Critic network for evaluating autonomous collision avoidance strategies, resulting in improving the autonomous collision avoidance ability of USVs. At the same time, to improve the efficiency of task allocation, hierarchical mechanisms, and regional division mechanisms are introduced to construct sub-system task planning models, which further decompose the task planning problem. A combination of successor features and an improved temporal difference error method is specifically applied to train another Critic network for evaluating the sub-systems task allocation schemes and collaborative motion trajectories, aiming to enhance the allocation efficiency of the sub-systems. Furthermore, transfer learning is employed to merge the sub-system task planning, using it as a constraint to direct the exploration and assessment of both the cluster task allocation schemes and the cluster collaborative motion trajectories. This enables rapid and accurate learning for task allocation within the multi-USV cluster. During the training phase of the Actor network, the introduction of the experience replay method and target network technique is employed to enhance the proximal policy optimization algorithm. This facilitates distributed joint training of the Actor network, thereby improving the accuracy of the algorithm. Simulation results validate the effectiveness and superiority of this method.
This study aims to present a model of the formation generation for multiple agents using a modified binary particle swarm optimisation (MBPSO). The major objective of this study is to maximise the formation combat capability and reduce the formation generation cost. We treat the ratio of the aforementioned two values as a measure of formation combat effectiveness. Additionally, chaos theory is adopted in the initialisation of MBPSO to acquire diversified particle population. Moreover, particle diversity is utilised to dynamically adjust the particle position updating process to guarantee the global convergence. A case study for multi-agent formation generation model in a naval battlefield is conducted. It is shown that the proposed algorithm can accomplish multi-agent formation generation under multiple constraints. Compared with the existing related algorithms, the proposed algorithm has improved search performance and better convergence characteristics.
Abstract Metabolic reprogramming of host cells plays critical roles during viral infection. Itaconate, a metabolite produced from cis-aconitate in the tricarboxylic acid cycle (TCA) by immune responsive gene 1 (IRG1), is involved in regulating innate immune response and pathogen infection. However, its involvement in viral infection and underlying mechanisms remain incompletely understood. Here, we demonstrate that the IRG1-itaconate axis facilitates the infections of VSV and IAV in macrophages and epithelial cells via Rab GTPases redistribution. Mechanistically, itaconate promotes the retention of Rab GTPases on the membrane via directly alkylating Rab GDP dissociation inhibitor beta (GDI2), the latter of which extracts Rab GTPases from the membrane to the cytoplasm. Multiple alkylated residues by itaconate, including cysteines 203, 335, and 414 on GDI2, were found to be important during viral infection. Additionally, this effect of itaconate needs an adequate distribution of Rab GTPases on the membrane, which relies on Rab geranylgeranyl transferase (GGTase-II)-mediated geranylgeranylation of Rab GTPases. The single-cell RNA sequencing data revealed high expression of IRG1 primarily in neutrophils during viral infection. Co-cultured and in vivo animal experiments demonstrated that itaconate produced by neutrophils plays a dominant role in promoting viral infection. Overall, our study reveals that neutrophils-derived itaconate facilitates viral infection via redistribution of Rab GTPases, suggesting potential targets for antiviral therapy.
The location of distress object in the maritime search area is difficult to determine, which has brought great difficulties to the search path planning. Aiming at this problem, a search path planning algorithm based on the probability of containment (POC) model for a distress object is proposed. This algorithm divides the area to be searched into several subareas by grid method and dynamically evaluates the POC of the distress object in each subarea using the Monte Carlo random particle method to build the POC model. On this basis, the POC is dynamically updated by employing the Bayes criterion within the constraint of the time window. Then, the sum of the POC of the object in the subareas is regarded as the weight of the search path. And the proposed algorithm dynamically executes the search path planning according to the maximum path weight. In comparison with the parallel line search path planning algorithm given in the “International Aeronautical and Maritime Search and Rescue Manual,” the simulation results show that the search path planning algorithm based on the POC model of the distress object can effectively improve the search efficiency and the probability of search success of the distress object.
The redox-degradable nano-micelle-reversed drug resistance by combination chemotherapy strategy of salinomycin (SL) that could specifically inhibit A/MCF-7 cells and a traditional broad-spectrum antitumor drug, doxorubicin (DOX).
A series of pullulan–doxorubicin conjugates (Pu-DOXs) were investigated for effectively delivering DOX to nuclei of hepatic carcinoma cells in subcutaneous tumor model. These Pu-DOXs were prepared by conjugating DOX onto pullulan molecule via pH-responsive hydrazone bond using spacers with different alkane chain length. The highest drug loading content of Pu-DOXs went up to nearly 50%, and the diameter of Pu-DOX nanoparticles ranged from 50 to 170 nm, as measured by DLS and TEM. These Pu-DOX nanoparticles could rapidly release DOX in the acidic environment at pH = 5.0 while being kept relatively stable in neural conditions. The in vitro cell coculture experiments revealed that these Pu-DOX nanoparticles were selectively internalized by hepatic carcinoma cells through receptor-mediated endocytosis via asialoglycoprotein receptor on the hepatic carcinoma cell surface. DOX was rapidly released from Pu-DOX nanoparticles in acidic endosome/lysosome, diffused into cell nuclei due to its strong affinity to nucleic acid, inhibited the cell proliferation, and accelerated the cell apoptosis. In the nude mice subcutaneous hepatic carcinoma model, Pu-DOX nanoparticles efficiently accumulated in the tumor site through the enhanced permeation and retention effect. Then DOX was specifically internalized by hepatic carcinoma cells and rapidly diffused into the nuclei of cells. Compared with the control group in in vivo experiments, these Pu-DOX nanoparticles effectively inhibited solid tumor growth, prolonging the lifetime of the experimental animal. These pH sensitive nanoparticles might provide an important clinical implication for targeted hepatic carcinoma therapy with high efficiency and low systematic toxicity.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)
'/%3e%3c/svg%3e)