Complex & Intelligent Systems最新文献

The support vector machine (SVM) method is an important basis of the current popular multi-class classification (MCC) methods and requires a sufficient number of samples. In the case of a limited number of samples, the problem of over-learning easily occurs, resulting in unsatisfactory classification. Therefore, this work investigates an MCC method that requires only a small number of samples. During model construction, raw data are converted into two-dimensional form via preprocessing. Via feature extraction, the learning network is measured and the loss function minimization principle is considered to better solve the problem of learning based on a small sample. Finally, three examples are provided to illustrate the feasibility and effectiveness of the proposed method.

Numerous graph neural network (GNN) models have been used for sentiment analysis in recent years. Nevertheless, addressing the issue of over-smoothing in GNNs for node representation and finding more effective ways to learn both global and local information within the graph structure, while improving model efficiency for scalability to large text sentiment corpora, remains a challenge. To tackle these issues, we propose a novel Bert-based unlinked graph embedding (BUGE) model for sentiment analysis. Initially, the model constructs a comprehensive text sentiment heterogeneous graph that more effectively captures global co-occurrence information between words. Next, by using specific sampling strategies, it efficiently preserves both global and local information within the graph structure, enabling nodes to receive more feature information. During the representation learning process, BUGE relies solely on attention mechanisms, without using graph convolutions or aggregation operators, thus avoiding the over-smoothing problem associated with node aggregation. This enhances model training efficiency and reduces memory storage requirements. Extensive experimental results and evaluations demonstrate that the adopted Bert-based unlinked graph embedding method is highly effective for sentiment analysis, especially when applied to large text sentiment corpora.

Flocking cohesion is critical for maintaining a group’s aggregation and integrity. Designing a potential function to maintain flocking cohesion unaffected by social distance is challenging due to the uncertainty of real-world conditions and environments that cause changes in agents’ social distance. Previous flocking research based on potential functions has primarily focused on agents’ same social distance and the attraction–repulsion of the potential function, ignoring another property affecting flocking cohesion: well depth, as well as the effect of changes in agents’ social distance on well depth. This paper investigates the effect of potential function well depths and agent’s social distances on the multi-agent flocking cohesion. Through the analysis, proofs, and classification of these potential functions, we have found that the potential function well depth is proportional to the flocking cohesion. Moreover, we observe that the potential function well depth varies with the agents’ social distance changes. Therefore, we design a segmentation potential function and combine it with the flocking control algorithm in this paper. It enhances flocking cohesion significantly and has good robustness to ensure the flocking cohesion is unaffected by variations in the agents’ social distance. Meanwhile, it reduces the time required for flocking formation. Subsequently, the Lyapunov theorem and the LaSalle invariance principle prove the stability and convergence of the proposed control algorithm. Finally, this paper adopts two subgroups with different potential function well depths and social distances to encounter for simulation verification. The corresponding simulation results demonstrate and verify the effectiveness of the flocking control algorithm.

Unsupervised domain adaptation transfers empirical knowledge from a label-rich source domain to a fully unlabeled target domain with a different distribution. A core idea of many existing approaches is to reduce the distribution divergence between domains. However, they focused only on part of the discrimination, which can be categorized into optimizing the following four objectives: reducing the intraclass distance between domains, enlarging the interclass distances between domains, reducing the intraclass distances within domains, and enlarging the interclass distances within domains. Moreover, because few methods consider multiple types of objectives, the consistency of data representations produced by different types of objectives has not yet been studied. In this paper, to address the above issues, we propose a zeroth- and first-order difference discrimination (ZFOD) approach for unsupervised domain adaptation. It first optimizes the above four objectives simultaneously. To improve the discrimination consistency of the data across the two domains, we propose a first-order difference constraint to align the interclass differences across domains. Because the proposed method needs pseudolabels for the target domain, we adopt a recent pseudolabel generation method to alleviate the negative impact of imprecise pseudolabels. We conducted an extensive comparison with nine representative conventional methods and seven remarkable deep learning-based methods on four benchmark datasets. Experimental results demonstrate that the proposed method, as a conventional approach, not only significantly outperforms the nine conventional comparison methods but is also competitive with the seven deep learning-based comparison methods. In particular, our method achieves an accuracy of 93.4% on the Office+Caltech10 dataset, which outperforms the other comparison methods. An ablation study further demonstrates the effectiveness of the proposed constraint in aligning the objectives.

Multi-scale representation provides an effective answer to the scale variation of objects and entities in semantic segmentation. The ability to capture long-range pixel dependency facilitates semantic segmentation. In addition, semantic segmentation necessitates the effective use of pixel-to-pixel similarity in the channel direction to enhance pixel areas. By reviewing the characteristics of earlier successful segmentation models, we discover a number of crucial elements that enhance segmentation model performance, including a robust encoder structure, multi-scale interactions, attention mechanisms, and a robust decoder structure. The attention mechanism of the asymmetric non-local neural network (ANNet) is merged with multi-scale pyramidal modules to accelerate model segmentation while maintaining high accuracy. However, ANNet does not account for the similarity between pixels in the feature map channel direction, making the segmentation accuracy unsatisfactory. As a result, we propose EMSNet, a straightforward convolutional network architecture for semantic segmentation that consists of Integration of enhanced regional module (IERM) and Multi-scale convolution module (MSCM). The IERM module generates weights using four or five-stage feature maps, then fuses the input features with the weights and uses more computation. The similarity of the channel direction feature graphs is also calculated using ANNet’s auxiliary loss function. The MSCM module can more accurately describe the interactions between various channels, capture the interdependencies between feature pixels, and capture the multi-scale context. Experiments prove that we perform well in tests using the benchmark dataset. On Cityscapes test data, we get 82.2% segmentation accuracy. The mIoU in the ADE20k and Pascal VOC datasets are, respectively, 45.58% and 85.46%.

In the domain of intelligent manufacturing, automatic anomaly detection plays a pivotal role and holds great significance for improving production efficiency and product quality. However, the scarcity and uncertainty of anomalous data pose significant challenges in this field. Data augmentation methods, such as Cutout, which are widely adopted in existing methodologies, tend to generate patterned data, leading to biased data and compromised detection performance. To deal with this issue, we propose an approach termed self-supervised anomaly detection with pixel-point random walk (P2 Random Walk), which combines data augmentation and Siamese neural networks. We develop a pixel-level data augmentation technique to enhance the randomness of generated data and establish a two-stage anomaly classification framework. The effectiveness of the P2 Random Walk method has been demonstrated on the MVTec dataset, achieving an AUROC of 96.2% and 96.3% for classification and segmentation, respectively, by using only data augmentation-based techniques. Specifically, our method outperforms other state-of-the-art methods in several categories, improving the AUROC for classification and segmentation by 0.5% and 0.3%, respectively, which demonstrates the high performance and strong academic value of our method in anomaly detection tasks.

The fault-tolerant tracking control problem is studied for the discrete-time systems with actuator faults. To lessen adverse impacts of actuator fault, a PPD information-driven fault estimation algorithm is established to adaptively estimate actuator fault information online, which avoids the additional construction and training process of neural network. With the aid of the adaptive fault compensation, a model-free adaptive fault-tolerant tracking control algorithm is constructed to ensure that the expected output reference trajectory can be tracked by system output. Moreover, only input and output data are employed throughout the design process, system dynamics are not demanded. Ultimately, the availability of developed strategy is proved through a simulation.

Mutual selection (the process of two types of objects choosing each other) often occurs in practical applications, such as those concerning financial credit. Considering the increasing demands for credibility, traditional artificial methods often cannot satisfy the corresponding requirements for security and transparency. Blockchain technology has the characteristics of decentralization, traceability, transparency, and being tamper-resistant, making it a potential method for solving the abovementioned problems. However, the existing consensus algorithms have some shortcomings in terms of efficiency, fault tolerance, security, and other relevant aspects, rendering them unsuitable for direct implementation in a mutual selection scenario. In this study, a system for mutual selection operations, denoted as “MuSelect Chain," is established. First, the institution information initialization method on blockchain is developed via a smart contract, ensuring the authenticity of information stored on the chain. Second, a mutual selection relationship confirmation algorithm is designed to ensure a reliable automated mutual selection process. Next, considering the characteristics of nodes participating in the network, a consensus algorithm called “Proof-of-Leadership” is proposed to ensure consistency of information stored by different nodes on the chain. Subsequently, an incentive mechanism is established with the focus on improving MuSelect Chain efficiency. Finally, a MuSelect Chain prototype is built. Simulation results prove that the proposed MuSelect Chain is secure with strong fault tolerance.

Wireless capsule endoscopy (WCE) enables imaging and diagnostics of the gastrointestinal (GI) tract to be performed without any discomfort. Despite this, several characteristics, including efficacy, tolerance, safety, and performance, make it difficult to apply and modify widely. The use of automated WCE to collect data and perform the analysis is essential for finding anomalies. Medical specialists need a significant amount of time and expertise to examine the data generated by WCE imaging of the patient’s digestive tract. To address these challenges, several computer vision-based solutions have been designed; nevertheless, they do not achieve an acceptable level of accuracy, and more advancements are required. Thus, in this study, we proposed four multi-classification deep learning (DL) models i.e., Vgg-19 + CNN, ResNet152V2, Gated Recurrent Unit (GRU) + ResNet152V2, and ResNet152V2 + Bidirectional GRU (Bi-GRU) and applied it on different publicly available databases for diagnosing ulcerative colitis, polyps, and dyed-lifted polyps using WCE images. To our knowledge, this is the only study that uses a single DL model for the classification of three different GI diseases. We compared the classification performance of the proposed DL classifiers in terms of many parameters such as accuracy, loss, Matthew's correlation coefficient (MCC), recall, precision, negative predictive value (NPV), positive predictive value (PPV), and F1-score. The results revealed that the Vgg-19 + CNN outperforms the three other proposed DL models in classifying GI diseases using WCE images. The Vgg-19 + CNN model achieved an accuracy of 99.45%. The results of four proposed DL classifiers are also compared with recent state-of-the-art classifiers and the proposed Vgg-19 + CNN model has performed better in terms of improved accuracy.

Federated learning (FL) represents a promising distributed machine learning paradigm for resolving data isolation due to data privacy concerns. Nevertheless, most vanilla FL algorithms, which depend on a server, encounter the problem of reliability and a high communication burden in real cases. Decentralized federated learning (DFL) that does not follow the star topology faces the challenges of weight divergence and inferior communication efficiency. In this paper, a novel DFL framework called federated incremental subgradient-proximal (FedISP) is proposed that utilizes the incremental method to perform model updates to alleviate weight divergence. In our setup, multiple clients are distributed in a ring topology and communicate in a cyclic manner, which significantly mitigates the communication load. A convergence guarantee is given under the convex condition to demonstrate the impact of the learning rate on our algorithms, which further improves the performance of FedISP. Extensive experiments on benchmark datasets validate the effectiveness of the proposed approach in both independent and identically distributed (IID) and non-IID settings while illustrating the advantages of the FedISP algorithm in achieving model consensus and saving communication costs.