查看更多>>摘要：The integration of visual and semantic information has been found to play a role in increasing the accuracy of social image clustering methods. However, existing approaches are limited by the heterogeneity gap between the visual and semantic modalities, and their performances significantly degrade due to the commonly sparse and incomplete tags in semantic modality. To address these problems, we propose a novel clustering framework to discover reasonable categories in unlabeled social images under the guidance of human explanations. First of all, a novel Explanation Generation Model (EGM) is proposed to automatically boost textual information for the sparse and incomplete tags based on an extra lexical database with human knowledge. Then, a novel clustering algorithm called Group Constrained Information Maximization (GCIM) is proposed to learn image categories. In this algorithm, a new type of constraint named group level side information is unprecedentedly defined to bridge the well-known heterogeneity gap between visual and textual modalities. Finally, an interactive draw-and-merge optimization method is proposed to ensure an optimal solution. Extensive experiments on several social image datasets including NUS-Wide, IAPRTC, MIRFlickr, ESP-Game and COCO demonstrate the superiority of the proposed approach to state-of-the-art baselines. (c) 2022 Elsevier Inc. All rights reserved.
查看更多>>摘要：Though many excellent methods have been designed for time series forecasting, they often entail sufficient training data from the same domain, which might be difficult to fulfill in some real-world applications. To alleviate this problem, a Relationship-Aligned Transfer Learning (aRATL) algorithm is proposed in this paper, in which the transfer learning process is implemented across different datasets. Whereas in some real scenarios, relevance between alternative source datasets and the target dataset may be ambiguous. For addressing this challenge, instead of calculating similarities at the data level, RATL tends to select the source model whose parameters facilitate the completion of the target task. The knowledge transfer in RATL contains representation relationship-aligned and regression relationship-aligned stages. The former one aims to enhance the representation ability of the target model by aligning relationships, in the form of triplets, between source and target models. The latter one aims to borrow regression experiences from the source model. Since predictions obtained by the source regression model are not always precise, RATL stresses the good results, but ignores the bad ones. Effectiveness of this proposed algorithm is underpinned by extensive experiments on five benchmark time series datasets, compared with several other state-of-the-art methods.(c) 2022 Elsevier Inc. All rights reserved.
查看更多>>摘要：This paper investigates the design of event-triggered (ET) fault detection (FD) observers for Takagi-Sugeno (T-S) fuzzy systems. In order to provide design flexibility, the membership functions (MFs) of the residual generator are different from those of the system and the observer. Unlike the existing asynchronous premise strategies that are designed based on inequality constraints between the MFs in the systems and those in the residual generators, an improved matching MF method is proposed, and the equality constraints are established under the proposed technical framework, such that the MFs of the residual generator can be obtained directly without calculation. By applying the constraints of the fuzzy weighting parameters and the differences of each MFs, new criteria in terms of linear matrix inequalities (LMIs) are derived to ensure the desired performance of the FD system. It is shown that the proposed method can overcome the shortcomings of the existing results. Finally, the effectiveness of the proposed FD scheme is verified by an example. (c) 2022 Published by Elsevier Inc.
查看更多>>摘要：We study opinion dynamics in multi-agent networks when a bias toward one of two pos-sible opinions exists, for example reflecting a status quo versus a superior alternative. Our aim is to investigate the combined effect of bias, network structure, and opinion dynamics on the convergence of the system of agents as a whole. Models of such evolving processes can easily become analytically intractable. In this paper, we consider a simple yet mathe-matically rich setting, in which all agents initially share an initial opinion representing the status quo. The system evolves in steps. In each step, one agent selected uniformly at ran -dom follows an underlying update rule to revise its opinion on the basis of those held by its neighbors, but with a probabilistic bias towards the superior alternative. We analyze con-vergence of the resulting process under well-known update rules. The framework we pro -pose is simple and modular, but at the same time complex enough to highlight a nonobvious interplay between topology and underlying update rule.(c) 2022 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
查看更多>>摘要：Deep learning models developed through multi-lead electrocardiogram (ECG) signals are considered the leading methods for the automated detection of arrhythmia on computer systems. However, due to the amplitudes of input signals, these models generate too many parameters for practical use. Therefore, they are rarely used on devices with limited computational resources in the newly-emerged technology of the Internet of medical things (IoMT). Knowledge distillation was utilized in this paper to propose a method for bridging the gap between the arrhythmia classification model with multi-lead ECG signals and the arrhythmia classification model with single-lead ECG signals by minimizing the performance decline. The proposed method consists of a teacher model with advanced architecture and a student model with simple architecture. The teacher model was already developed through multi-lead ECG signals, whereas the student model was developed through single-lead signals under the supervision of the teacher. Despite its simplicity, the student model receives the dark knowledge of multi-lead ECG signals from the teacher by imitating the teacher's behavior in the development process. According to the results, the student model was nearly 262.18 times more compressed than its teacher. Moreover, the student experienced approximately 0.81% of accuracy decline in Chapman ECG with 10646 patients. (c) 2022 Published by Elsevier Inc.
查看更多>>摘要：To make service robots plan a whole task execution sequence and deal with task failure intelligently in the face of object occlusion and incomplete home environment information, a hybrid offline and online task planning strategy is proposed. We establish object-level semantic map and object location probabilistic relations between dynamic and static objects. Semantic mapping helps to obtain semantic locations of static objects, and the probabilistic relationship between dynamic and static objects can obtain semantic loca-tions of dynamic objects through probabilistic reasoning. Probabilistic planning domain definition language (PPDDL) can generate offline action sequences, while partially observ-able Markov decision process (POMDP) can generate online action sequences. Therefore, a hybrid task planner that can receive semantic location information is constructed to gen-erate offline and online action sequences, and realizes the dynamic switching of the two kinds of sequences through the designed planner switching mechanism. In order to improve the robustness and intelligence of robot task planning, a task replanning mecha-nism considering the position relationship between robot and candidate static objects is designed. Experimental results in real environment and simulation environment show that this approach can effectively increase the intelligence of task planning and improve the robustness and efficiency of robot task execution.(c) 2022 Elsevier Inc. All rights reserved.
查看更多>>摘要：General type-2 fuzzy systems (GT2 FLSs) provide a more flexible way of overcoming an uncertain lack of uniformity in different applications. Centroid type reduction is one of the major component of GT2 FLSs, it is currently one of the key factors restricting GT2 FLSs efficiency. This paper proposes a novel and efficient method for centroid type reduction. The method is based on alpha-plane representation, where a general type-2 fuzzy set is decomposing into a series of alpha planes. The centroid for each plane is the calculated, layer by layer, from the top down, until the alpha = 0 plane is reached. In each alpha plane, a centroid type reduction calculation is performed using an improved enhanced opposite direction searching algorithm (IEODS). Finally, the centroids obtained for each plane are aggregated to obtain a type-1 fuzzy set, which form the centroid of general type-2 fuzzy set. Experiments show that this method results in less calculation time and fewer iterations than other alpha-plane representation-based reduction methods. By providing more efficient type reduction, the method improves the applicability of general type-2 fuzzy systems in more complex environments and embedded platforms, thereby enhancing their scope for future development. (c) 2022 Elsevier Inc. All rights reserved.
查看更多>>摘要：In this paper, a hybrid elementary cellular automata (HECA) composed of two chaotic global rules is designed to enhance the chaotic properties of the pseudo-random coupled map lattices (PRCML) model based on the Chirikov standard map (CSM). Theory analyses and simulations indicate that the PRCML-HECA model exhibits superior chaos. Subsequently, a novel image encryption scheme based on this model is implemented. The scheme consists of key seed generation, encryption, and decryption. A series-wound PRCML-HECA model is designed to generate the key seed related to the plain-image, initial key, and timestamp. This process is nonlinear and irreversible, ensuring the proposed scheme can resist the chosen-plaintext/ciphertext attacks. Furthermore, a dynamic substitution box changed with the key seed is also included in the proposed scheme. Moreover, plenty of statistical and security analyses, such as diffusion, histogram, correlation analyses, etc., are introduced in this paper. The results prove that the proposed image encryption scheme based on the PRCML-HECA model can resist differential and statistical attacks. And the scheme also possesses strong robustness and high efficiency.(c) 2022 Elsevier Inc. All rights reserved.
查看更多>>摘要：Feature importance and interaction are among the main issues in explainable artificial intelligence or interpretable machine learning. To measure feature importance and interaction, several methods, such as H-statistic and partial dependency, have been proposed. However, it is difficult to understand the practical implications of importance and interaction. In this paper, a new method for measuring feature importance and interaction is proposed. For the classification model, we observed correctly predicted cases in a predictive model and grouped them according to the characteristics of the cases. We derived a method for feature importance and interaction from group information. For the regression model, we grouped cases according to the change in the size of the prediction error. The proposed method supports the same rationale for feature importance and interaction. It also supports the decomposition of feature importance to feature power and feature interactions. To implement the proposed method, three visualization tools, including a feature interaction graph, are implemented. Through the proposed work, we can better understand the working mechanism of a predictive model.(c) 2022 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
查看更多>>摘要：Generative adversarial networks (GAN) training is subject to problems including mode col-lapse, gradient vanishing, and instability. Although many different losses have been pro-posed to alleviate these shortcomings, they heavily rely on a fixed-value function with limited expressive power in terms of robustness, whereby failing to perform consistently over multiple data sets. To solve this problem, we propose a parametric and robust ab-loss function that can improve the performances of GAN on different data sets. Specifically, unlike standard GAN loss function it exploits the alpha beta-divergence (AB-divergence) to weigh the likelihood ratio associated with each data point. This weighing mechanism makes the model robust to noises and yields better models in terms of FID score. To reduce the cost of searching for the optimal alpha and beta, we further propose an adaptive version to systematically update these parameters according to statistics of the discriminator's output. Moreover, alpha beta-loss can be reduced to Least Square GAN (LS-GAN) and standard GAN (SGAN) loss function as special cases. We conduct extensive experiments on both synthetic and real-world data sets. Experimental results over the synthetic data sets (2D Gaussian ring and grid) demonstrate that our approach can significantly alleviate the issue of mode collapse. Additionally, by constraining the gradient of the discriminator that is fed back to the generator via finely adjusting the hyper-parameters alpha and beta, our approach can improve the quality of synthetic images, as can be seen from the decrease of FID from 40 to 23.71 on the data set CIFAR10 using the SN-DCGAN architecture. (c) 2022 Elsevier Inc. All rights reserved.
NSTL
Elsevier