Limin Wang

Physician resources are fundamental assets of the healthcare system. The unequal distribution of these resources inevitably leads to patients seeking medical services elsewhere, increasing healthcare costs, and hindering the realization of health equ...

Evolutionary multitask optimization (EMTO) has been increasingly employed in addressing high-dimensional feature selection challenges, but current EMTO algorithms still have three deficiencies: First, in task generation, they just consider the linear correlation among features but ignore the nonlinear correlation. Second, they commonly encounter negative knowledge transfer. Third, they are hard to strike an optimal balance between global search capability and computational efficiency. This paper proposes an adaptive aggregative multitask competitive swarm optimization (AAMCSO) for high-dimensional feature selection, which contains three novel and effective strategies to address the above deficiencies. Firstly, AAMCSO proposes a linear-and-nonlinear-correlation-based task generation strategy to generate multiple tasks while considering both linear and nonlinear correlation between features and labels. Secondly, AAMCSO proposes an adaptive aggregative knowledge transfer strategy to adaptively transfer positive knowledge among related tasks. Thirdly, AAMCSO proposes a bi-directional asymmetric flip strategy to guide the population to search for a smaller feature subset with better classification performance. We have conducted extensive comparative experiments on AAMCSO and multiple state-of-the-art feature selection algorithms in high-dimensional feature selection problems with up to 10000 dimensions. The results show that AAMCSO achieves significantly superior performance to the state-of-the-art comparison algorithms in terms of both classification accuracy and feature number.

Unsupervised feature selection is an important tool in data mining, machine learning, and pattern recognition. Although data labels are often missing, the number of data classes can be known and exploited in many scenarios. Therefore, a structured graph, whose number of connected components is identical to the number of data classes, has been proposed and is frequently applied in unsupervised feature selection. However, methods based on the structured graph learning face two problems. First, their structured graphs are not always guaranteed to maintain the same number of connected components as the data classes with existing optimization algorithms. Second, they usually lack strategies for choosing moderate hyperparameters. To solve these problems, an efficient and stable unsupervised feature selection method based on a novel structured graph and data discrepancy learning (ESUFS) is proposed. Specifically, the novel structured graph, consisting of a pairwise data similarity matrix and an indicator matrix, can be efficiently learned by solving a discrete optimization problem. Data discrepancy learning focuses on features that maximize the difference among data and helps in selecting discriminative features. Extensive experiments conducted on various datasets show that ESUFS outperforms state-of-the-art methods not only in accuracy (ACC) but also in stability and speed.

In computer vision, a series of exemplary advances have been made in several areas involving image classification, semantic segmentation, object detection, and image super-resolution reconstruction with the rapid development of deep convolutional neural network (CNN). The CNN has superior features for autonomous learning and expression, and feature extraction from original input data can be realized by means of training CNN models that match practical applications. Due to the rapid progress in deep learning technology, the structure of CNN is becoming more and more complex and diverse. Consequently, it gradually replaces the traditional machine learning methods. This paper presents an elementary understanding of CNN components and their functions, including input layers, convolution layers, pooling layers, activation functions, batch normalization, dropout, fully connected layers, and output layers. On this basis, this paper gives a comprehensive overview of the past and current research status of the applications of CNN models in computer vision fields, e.g., image classification, object detection, and video prediction. In addition, we summarize the challenges and solutions of the deep CNN, and future research directions are also discussed.

To solve the problems of low convergence accuracy, slow speed, and common falls into local optima of the Chicken Swarm Optimization Algorithm (CSO), a performance enhancement strategy of the CSO algorithm (PECSO) is proposed with the aim of overcoming its deficiencies. Firstly, the hierarchy is established by the free grouping mechanism, which enhances the diversity of individuals in the hierarchy and expands the exploration range of the search space. Secondly, the number of niches is divided, with the hen as the center. By introducing synchronous updating and spiral learning strategies among the individuals in the niche, the balance between exploration and exploitation can be maintained more effectively. Finally, the performance of the PECSO algorithm is verified by the CEC2017 benchmark function. Experiments show that, compared with other algorithms, the proposed algorithm has the advantages of fast convergence, high precision and strong stability. Meanwhile, in order to investigate the potential of the PECSO algorithm in dealing with practical problems, three engineering optimization cases and the inverse kinematic solution of the robot are considered. The simulation results indicate that the PECSO algorithm can obtain a good solution to engineering optimization problems and has a better competitive effect on solving the inverse kinematics of robots.

The density-based spatial clustering of application with noise (DBSCAN) algorithm is able to cluster arbitrarily structured datasets. However, the clustering result of this algorithm is exceptionally sensitive to the neighborhood radius (Eps) and noise points, and it is hard to obtain the best result quickly and accurately with it. To solve the above problems, we propose an adaptive DBSCAN method based on the chameleon swarm algorithm (CSA-DBSCAN). First, we take the clustering evaluation index of the DBSCNA algorithm as the objective function and use the chameleon swarm algorithm (CSA) to iteratively optimize the evaluation index value of the DBSCAN algorithm to obtain the best Eps value and clustering result. Then, we introduce the theory of deviation in the data point spatial distance of the nearest neighbor search mechanism to assign the identified noise points, which solves the problem of over-identification of the algorithm noise points. Finally, we construct color image superpixel information to improve the CSA-DBSCAN algorithm’s performance regarding image segmentation. The simulation results of synthetic datasets, real-world datasets, and color images show that the CSA-DBSCAN algorithm can quickly find accurate clustering results and segment color images effectively. The CSA-DBSCAN algorithm has certain clustering effectiveness and practicality.

The Chinese Named Entity Recognition (CNER) is difficult to classify words accurately because of the lack of natural separators. Character embedding can effectively avoid the cumulative errors caused by word division, but carries relatively insufficient inter-character sequence information. In order to solve the problem of insufficient feature representation in Chinese NER model, this paper proposes a new CNER model (DBM-CNER), which has a dual-branch encoder structure. One branch is a new Transformer structure, in which the multi-head self-attention has dual-position encodings, and the other is a feedforward multi-scale dilated convolution fusion mechanism. This structure can better extract the global and local dependencies between characters in text from multiple perspectives, and enhance the ability to extract multi-level context features. Compared with other CNER models, the experimental results on the benchmark datasets in three different fields show that the proposed model is superior to state-of-the-art works.

The density peaks clustering (DPC) algorithm can identify clusters with various shapes and densities in the underlying dataset. However, the DPC algorithm cannot exactly find the true quantity of clustering centers when computing the local density, and it is difficult to handle non‐convex datasets. Moreover, the DPC algorithm is difficult to identify boundary points and outliers without a reasonable allocation strategy when dealing with low‐density points. To solve these limitations, a novel deviation density peaks clustering (DeDPC) algorithm is proposed. First, the local deviation of the spatial distance of datasets with different structures is utilised to replace the local density to generate a more reasonable clustering center decision graph. Second, a threshold is defined to further divide and process low‐density points. Finally, outliers in low‐density points can be accurately found to accurately cluster the dataset. To evaluate the performance of the DeDPC algorithm, experiments are conducted on synthetic and real‐world datasets and the DeDPC is compared with other clustering methods. The DeDPC is also applied to medical image segmentation to further demonstrate its capability for medical image processing. The simulation results show that the DeDPC method has good validity and utility for both non‐convex datasets and medical image segmentation.