Evolutionary algorithms exhibit flexibility and global search advantages in multi-UAV path planning, effectively addressing complex constraints. However, when there are numerous obstacles in the environment, especially narrow passageways, the algorithm often struggles to quickly find a viable path. Additionally, collaborative constraints among multiple UAVs complicate the search space, making algorithm convergence challenging. To address these issues, we propose a novel hybrid particle swarm optimization algorithm called PPSwarm. This approach initially employs the RRT* algorithm to generate an initial path, rapidly identifying a feasible solution in complex environments. Subsequently, we adopt a priority planning method to assign priorities to UAVs, simplifying collaboration among them. Furthermore, by introducing a path randomization strategy, we enhance the diversity of the particle swarm, thereby avoiding local optimum solutions. The experimental results show that, in comparison to algorithms such as DE, PSO, ABC, GWO, and SPSO, the PPSwarm algorithm demonstrates significant advantages in terms of path quality, convergence speed, and runtime when addressing path planning issues for 40 UAVs across four different scenarios. In larger-scale experiments involving 500 UAVs, the proposed algorithm also exhibits excellent processing capability and scalability.

Jiduan Liu, Jiahao Liu, Qifan Wang, Jingang Wang, Wei Wu, Yunsen Xian, Dongyan Zhao, Kai Chen, Rui Yan. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

When dealing with UAV path planning problems, evolutionary algorithms demonstrate strong flexibility and global search capabilities. However, as the number of UAVs increases, the scale of the path planning problem grows exponentially, leading to a significant rise in computational complexity. The Cooperative Co-Evolutionary Algorithm (CCEA) effectively addresses this issue through its divide-and-conquer strategy. Nonetheless, the CCEA needs to find a balance between computational efficiency and algorithmic performance while also resolving convergence difficulties arising from the increased number of decision variables. Moreover, the complex interrelationships between the decision variables of each UAV add to the challenge of selecting appropriate decision variables. To tackle this problem, we propose a novel collaborative algorithm called CCEA-ADVS. This algorithm reduces the difficulty of the problem by decomposing high-dimensional variables into sub-variables for collaborative optimization. To improve the efficiency of decision variable selection in the collaborative algorithm and to accelerate the convergence speed, an adaptive decision variable selection strategy is introduced. This strategy selects decision variables according to the order of solving single-UAV constraints and multi-UAV constraints, reducing the cost of the optimization objective. Furthermore, to improve computational efficiency, a two-stage evolutionary optimization process from coarse to fine is adopted. Specifically, the Adaptive Differential Evolution with Optional External Archive algorithm (JADE) is first used to optimize the waypoints of the UAVs to generate a basic path, and then, the Dubins algorithm is combined to optimize the trajectory, yielding the final flight path. The experimental results show that in four different scenarios involving 40 UAVs, the CCEA-ADVS algorithm significantly outperforms the Grey Wolf Optimizer (GWO), Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC), and JADE algorithms in terms of path performance, running time, computational efficiency, and convergence speed. In addition, in large-scale experiments involving 500 UAVs, the algorithm also demonstrates good adaptability, stability, and scalability.

Clouds have an enormous influence on the hydrological cycle, Earth’s radiation budget, and climate changes. Accurate automatic recognition of cloud shape based on ground-based cloud images is beneficial to analyze the atmospheric motion state and water vapor content, and then to predict weather trends and identify severe weather processes. Cloud type classification remains challenging due to the variable and diverse appearance of clouds. Deep learning-based methods have improved the feature extraction ability and the accuracy of cloud type classification, but face the problem of lack of labeled samples. In this paper, we proposed a novel classification approach of ground-based cloud images based on contrastive self-supervised learning (CSSL) to reduce the dependence on the number of labeled samples. First, data augmentation is applied to the input data to obtain augmented samples. Then contrastive self-supervised learning is used to pre-train the deep model with a contrastive loss and a momentum update-based optimization. After pre-training, a supervised fine-tuning procedure is adopted on labeled data to classify ground-based cloud images. Experimental results have confirmed the effectiveness of the proposed method. This study can provide inspiration and technical reference for the analysis and processing of other types of meteorological remote sensing data under the scenario of insufficient labeled samples.

Though it has been easier to build large face datasets by collecting images from the Internet in this Big Data era, the time-consuming manual annotation process prevents researchers from constructing larger ones, which makes the automatic cleaning of noisy labels highly desirable. However, identifying mislabeled faces by machine is quite challenging because the diversity of a person’s face images that are captured wildly at all ages is extraordinarily rich. In view of this, we propose a graph-based cleaning method that mainly employs the community detection algorithm and deep CNN models to delete mislabeled images. As the diversity of faces is preserved in multiple large communities, our cleaning results have both high cleanness and rich data diversity. With our method, we clean the extremely large MS-Celeb-1M face dataset (approximately 10 million images with noisy labels) and obtain a clean version of it called C-MS-Celeb (6,464,018 images of 94,682 celebrities). By training a single-net model using our C-MS-Celeb dataset, without fine-tuning, we achieve 99.67% at Equal Error Rate on the LFW face recognition benchmark, which is comparable to other state-of-the-art results. This demonstrates the data cleaning positive effects on the model training. To the best of our knowledge, our C-MS-Celeb is the largest clean face dataset that is publicly available so far, which will benefit face recognition researchers.

Learning to rank algorithm has become important in recent years due to its successful application in information retrieval, recommender system, and computational biology, and so forth. Ranking support vector machine (RankSVM) is one of the state-of-art ranking models and has been favorably used. Nonlinear RankSVM (RankSVM with nonlinear kernels) can give higher accuracy than linear RankSVM (RankSVM with a linear kernel) for complex nonlinear ranking problem. However, the learning methods for nonlinear RankSVM are still time-consuming because of the calculation of kernel matrix. In this paper, we propose a fast ranking algorithm based on kernel approximation to avoid computing the kernel matrix. We explore two types of kernel approximation methods, namely, the Nyström method and random Fourier features. Primal truncated Newton method is used to optimize the pairwise L2-loss (squared Hinge-loss) objective function of the ranking model after the nonlinear kernel approximation. Experimental results demonstrate that our proposed method gets a much faster training speed than kernel RankSVM and achieves comparable or better performance over state-of-the-art ranking algorithms.