查看更多>>摘要：Supervised representation learning based on the teacher-student framework can extract quality-related features for soft sensors, in which the teacher network extracts representation information for the student network as supervision information. In traditional applications, the teacher network is heavy and is difficult to train, so the teacher network is conventionally pre-trained. However, the pre-training of the teacher network is unnecessary if the training process is not complicated so that it is meaningful to joint optimize the teacher-student network. In our application, the teacher-student framework is used to extract quality-related representation information for soft sensors. The objective is to maximize the mutual information of representation information and supervision information, in which the inconsistency of distributions between observed information and supervisory information is modeled as isotropic Gaussian noise. The objective is decoupled through analysis under some approximate assumptions so that the alternative iteration method can be used to update the parameters of the model. The proposed quality-related feature extraction method is applied to soft sensors combined with a traditional just-in-time learning method. Our experiments show that the prediction performance of our representation extraction method is better than other existing representation extraction algorithms. (C) 2022 Published by Elsevier Inc.
查看更多>>摘要：This paper presents a state estimation-based robust optimal control strategy for influenza epidemics in an interactive human society in the presence of modeling uncertainties. Interactive society is influenced by random entrance of individuals from other human societies whose effects can be modeled as a non-Gaussian noise. Since only the number of exposed and infected humans can be measured, the states of the influenza epidemics are first estimated by an extended maximum correntropy Kalman filter (EMCKF) to provide a robust state estimation in the presence of the non-Gaussian noise. An online quadratic program (QP) optimization is then synthesized subject to a robust control Lyapunov function (RCLF) to minimize susceptible and infected humans, while minimizing and bounding the rates of vaccination and antiviral treatment. The main contribution of this work is twofold. First, the joint QP-RCLF-EMCKF strategy meets multiple design specifications such as state estimation, tracking, pointwise control optimality, and robustness to parameter uncertainty and state estimation errors that have not been achieved simultaneously in previous studies. Second, the uniform ultimate boundedness (UUB)/convergence of all error trajectories is guaranteed by using a Lyapunov stability argument. Simulation results show that the proposed approach achieves appropriate tracking and state estimation performance with good robustness. (C) 2022 Elsevier Inc. All rights reserved.
查看更多>>摘要：Spatial co-location pattern (SCP) mining discovers subsets of spatial feature types whose objects frequently co-locate in a geographic space. Many existing methods treat the space as homogeneous, use absolute Euclidean distance to measure the neighbor relationship between objects and use a participation index to measure the prevalence of SCPs. Several issues arise: (1) it may be that the distance between objects cannot be accurately defined since it is a relative and fuzzy concept; (2) the degree of neighborliness and sharing relationships between objects are neglected; (3) current methods for collecting participating objects by generating candidate table instances utilizing combined search techniques are computationally expensive. In this paper, we propose a method based on fuzzy grid cliques to find all prevalent SCPs. Specifically, fuzzy theory is introduced to define the proximity between objects. The fuzzy participating contribution index (FPCI) is defined to measure the prevalence of SCPs, and it considers both the neighbor degree and sharing relationship between objects. Based on the defined proximity, a basic mining framework based on fuzzy grid cliques is proposed. We first design a naive algorithm based on the participating objects' filtering and verification called POFV, which uses a fuzzy grid clique search technology instead of combination search to collect participating objects and avoids enumerating all table instances. To solve a dilemma within POFV, we develop a maximal fuzzy grid cliques search based algorithm called MFGC, which can effectively reuse information. Experiments on both real and synthetic data sets verify the superiority of our proposed approaches, by showing that MFGC greatly outperforms the baseline algorithm and more efficiently captures SCPs. (c) 2022 Elsevier Inc. All rights reserved.
Biswas, MilonGaur, LoveleenAlenezi, FayadhSantosh, K. C....
13页
查看更多>>摘要：Chest X-ray (CXR) imaging is a low-cost, easy-to-use imaging alternative that can be used to diagnose/screen pulmonary abnormalities due to infectious diseaseX: Covid-19, Pneumonia and Tuberculosis (TB). Not limited to binary decisions (with respect to healthy cases) that are reported in the state-of-the-art literature, we also consider non-healthy CXR screening using a lightweight deep neural network (DNN) with a reduced number of epochs and parameters. On three diverse publicly accessible and fully categorized data -sets, for non-healthy versus healthy CXR screening, the proposed DNN produced the fol-lowing accuracies: 99.87% on Covid-19 versus healthy, 99.55% on Pneumonia versus healthy, and 99.76% on TB versus healthy datasets. On the other hand, when considering non-healthy CXR screening, we received the following accuracies: 98.89% on Covid-19 ver-sus Pneumonia, 98.99% on Covid-19 versus TB, and 100% on Pneumonia versus TB. To fur-ther precisely analyze how well the proposed DNN worked, we considered well-known DNNs such as ResNet50, ResNet152V2, MobileNetV2, and InceptionV3. Our results are comparable with the current state-of-the-art, and as the proposed CNN is light, it could potentially be used for mass screening in resource-constraint regions.(c) 2022 Elsevier Inc. All rights reserved.
查看更多>>摘要：Rumor detection is increasingly important in we-media. Most existing rumor detection models focus on mining features of text content, user profiles and propagation patterns. However, these models lack an efficient way to integrate information from multiple resources and have poor performance or low generalization on unseen data because they tend to capture event-specific features contained in visible data. In this study, we propose an end-to-end Multi-channel Graph ATtention network with Event-Sharing Module named MGAT-ESM. First, we parallelly build three subgraphs to model the propagation structures of source tweets and their responses, the relationships of source tweets and their words, and those of source tweets and their related users, respectively. We then design a path embedding method to learn the semantic information of propagation structures, and use graph attention neural network as a backbone to learn the representations of the other two subgraphs, and then aggregate the embedding representations of these three channel subgraphs with attention mechanism. Moreover, for learning event-invariant features in different rumors, we add an event-sharing module to the backbone network. Finally, we combine the learnt event-invariant features with the aggregated representations to get the final predictions. Experiments on two real-world benchmarks demonstrate that MGAT-ESM achieves the state-of-the-art performance. (c) 2022 Elsevier Inc. All rights reserved.
NSTL
Elsevier