This paper proposes an intrusion detection model based on a convolutional neural network. First, a one-dimensional convolutional neural network structure is used to speed up the convergence of the model and prevent overfitting. Then, using the method based on SMOTE-GMM, the sample data is equalize...
Improving graph convolutional networks based on relation-aware attention for end-to-end relation extraction. IEEE Access 2020, 8, 51315–51323. [Google Scholar] [CrossRef] Krizhevsky, A.; Sutskever, I.; Hinton, G.E. Imagenet classification with deep convolutional neural networks. In Advances ...
we improve upon prior understanding and performance in complex-valued convolutional neural networks. Using novel derivations of convolutional, down-sampling, non-linear, and affine layers implemented in a complex-valued counterpart to Caffe, we proved results when evaluated against real-valued models in...
Recently, the hybrid deep neural network (DNN)-hidden Markov model (HMM) has been shown to significantly improve speech recognition performance over the conventional Gaussian mixture model (GMM)-HMM. The performance improvement is partially attributed to the ability of the DNN to model complex c...
We propose a novel, end-to-end model based on convolutional and recurrent neural networks for speech enhancement. Our model is purely data-driven: it does not make any assumptions about the type or the stationarity of the noise. In contrast to existing methods that use multilayer perceptrons (...
Spatial pyramid pooling in deep convolutional networks for visual recognition. IEEE Trans. Pattern Anal. Mach. Intell. 37(9), 1904–1916 (2015). Article PubMed Google Scholar Girshick, R. Fast r-cnn. In Proceedings of the IEEE International Conference on Computer Vision. 1440–1448 (2015)...
Existing image convolutional neural network based on image super-resolution algorithm has a problem of image texture blurred to improve. In this paper, we first analyze the factors of reconstructed image quality, then use the parametric rectified linear unit (PRLU) to solve the problem of over-co...
This work was supported in part by Intelligent recognition of open-pit mining based on deep learning and remote sensing big data—Taking the eastern Mongolian region as an example under Grant NJZY22278; in part by Research on Convolutional Neural Network Algorithm Based on Big Data Analysis and ...
We do not aim at providing an overview of the massive work on neural networks in the past decades. Some prominent examples include LeCun's LeNet convolutional networks [30] and Schmidhuber's long short-term memories [31], illustrating the fundamental work that leads to the deep learning revol...
Huang and others have developed a framework based on convolutional neural networks and LSTM to solve the problem of air pollutant index prediction, and use historical data to predict the air pollution index29. Because of its structural design, LSTM has very good performance in time-series data ...