Topic
Hybrid neural network
About: Hybrid neural network is a research topic. Over the lifetime, 1305 publications have been published within this topic receiving 18223 citations.
Papers published on a yearly basis
Papers
More filters
••
23 Jul 1991TL;DR: In this paper, a fully recurrent neural network and a rule-based expert system are combined in a hybrid architecture to provide power plant operators with an intelligent on-line advisory system.
Abstract: A fully recurrent neural network and a rule-based expert system are combined in a hybrid architecture to provide power plant operators with an intelligent on-line advisory system. Its purpose is to alert the operator to impending or occurring abnormal conditions related to the plant's boiler. The hybrid system is trained to provide a model of the boiler under normal operation, while the rules address a general set of diagnostic events. Deviation from normal conditions trigger rules to suggest corrective action. This system is intended to increase plant availability and efficiency by automatically deducing abnormal boiler conditions before they become critical. >
6 citations
••
01 Dec 2020TL;DR: In this article, a hybrid neural network text classification model which integrates time series convolutional network TGNet was proposed to capture the relationship between hidden features on different time scales, and uses Gated Tanh-ReLU Units (GTRU) as the activation layer to improve the expression ability of neural network to the model.
Abstract: Aiming at the problems of insufficient feature extraction and low classification accuracy in the classification of Chinese news texts by the neural network, this paper proposes a hybrid neural network text classification model which integrates time series convolutional network TGNet. The new model utilizes Temporal Convolutional Network to capture the relationship between hidden features on different time scale, and uses Gated Tanh-ReLU Units (GTRU) as the activation layer to improve the expression ability of neural network to the model. Meanwhile, the Gated Recurrent Unit networks (GRU) is used to learn the semantic features of context. Finally, the extracted features are fused and input into Softmax for classification. Experimental results show that the text classification model proposed in this paper has achieved better classification results in the published Chinese news data sets SougoCS and FuDan.
6 citations
•
01 Jan 1993
6 citations
••
TL;DR: By combining KL decomposition and neural networks, a reduced dynamical model of the Kuramoto-Sivashinsky (KS) equation is obtained.
Abstract: A hybrid approach consisting of two neural networks is used to model the oscillatory dynamical behavior of the Kuramoto-Sivashinsky (KS) equation at a bifurcation parameter α=84.25. This oscillatory behavior results from a fixed point that occurs at α=72 having a shape of two-humped curve that becomes unstable and undergoes a Hopf bifurcation at α=83.75. First, Karhunen-Loeve (KL) decomposition was used to extract five coherent structures of the oscillatory behavior capturing almost 100% of the energy. Based on the five coherent structures, a system offive ordinary differential equations (ODEs) whose dynamics is similar to the original dynamics of the KS equation was derived via KL Galerkin projection. Then, an autoassociative neural network was utilized on the amplitudes of the ODEs system with the task of reducing the dimension of the dynamical behavior to its intrinsic dimension, and a feedforward neural network was usedto model the dynamics at a future time. We show that by combining KL decomposition and neural networks, a reduced dynamical model of the KS equation is obtained.
6 citations
••
TL;DR: CBBI as discussed by the authors uses a hybrid neural network model to automatically learn the representative spatial and temporal features from the network traffic, such as the position relationship of the internal organization structure in network communication traffic, the time sequence of the data packets, and the duration of the network flow.
Abstract: With the rapid growth of the Internet of Things (IoT) devices, security risks have also arisen. The preidentification of IoT devices connected to the network can help administrators to set corresponding security policies according to the functionality and heterogeneity of the devices. However, the existing methods are based on manually extracted features and prior knowledge to identify the IoT devices, which increases the difficulty of the device identification task and reduces the timeliness. In this paper, we present CBBI, a novel IoT device identification approach. On the one hand, CBBI uses a hybrid neural network model Conv-BiLSTM to automatically learn the representative spatial and temporal features from the network traffic, such as the position relationship of the internal organization structure in network communication traffic, the time sequence of the data packets, and the duration of the network flow. On the other hand, CBBI contains the data augmentation module FGAN that solves the problem of data imbalance in deep learning and improves the accuracy of the model. Finally, we used the public dataset and laboratory dataset to evaluate CBBI from multiple dimensions. The evaluation results for different datasets show that our approach achieves the accurate identification of IoT devices.
6 citations