\ Industrial control network is a direct interface between information system and physical control process. Due to the lack of authentication, encryption, and other necessary security protection designs, it has become...
详细信息
\ Industrial control network is a direct interface between information system and physical control process. Due to the lack of authentication, encryption, and other necessary security protection designs, it has become the main target of malicious attacks under the trend of increasing openness. In order to protect the industrial control systems, we examine the detection of abnormal traffic in industrial control network and propose a method of detecting abnormal traffic in industrial control network based on autoencoder technology. What is more, a new deep autoencoder model was designed to reduce the dimensionality of traffic data in industrial control network. In this article, the Kullback-Leibler divergence was added to the loss function to improve the ability of feature extraction and the ability to recover raw data. Finally, this model was compared with the traditional data dimensionality reduction method (principal component analysis (PCA), independent component analysis, and singular value decomposition) on gas pipeline dataset. The results show that the approach designed in this article outperforms the three methods in different scenes in terms of f(1) score.
We compare standard autoencoder topologies' performances for timbre generation. We demonstrate how different activation functions used in the autoencoder's bottleneck distributes a training corpus's embedd...
详细信息
ISBN:
(纸本)9781728169262
We compare standard autoencoder topologies' performances for timbre generation. We demonstrate how different activation functions used in the autoencoder's bottleneck distributes a training corpus's embedding. We show that the choice of sigmoid activation in the bottleneck produces a more bounded and uniformly distributed embedding than a leaky rectified linear unit activation. We propose a one-hot encoded chroma feature vector for use in both input augmentation and latent space conditioning. We measure the performance of these networks, and characterize the latent embeddings that arise from the use of this chroma conditioning vector. An open source, real-time timbre synthesis algorithm in Python is outlined and shared.
autoencoders are commonly trained using element-wise loss. However, element-wise loss disregards high-level structures in the image which can lead to embeddings that disregard them as well. A recent improvement to aut...
详细信息
ISBN:
(纸本)9781728169262
autoencoders are commonly trained using element-wise loss. However, element-wise loss disregards high-level structures in the image which can lead to embeddings that disregard them as well. A recent improvement to autoencoders that helps alleviate this problem is the use of perceptual loss. This work investigates perceptual loss from the perspective of encoder embeddings themselves. autoencoders are trained to embed images from three different computer vision datasets using perceptual loss based on a pretrained model as well as pixel-wise loss. A host of different predictors are trained to perform object positioning and classification on the datasets given the embedded images as input. The two kinds of losses are evaluated by comparing how the predictors performed with embeddings from the differently trained autoencoders. The results show that, in the image domain, the embeddings generated by autoencoders trained with perceptual loss enable more accurate predictions than those trained with element-wise loss. Furthermore, the results show that, on the task of object positioning of a smallscale feature, perceptual loss can improve the results by a factor 10. The experimental setup is available online.(1)
Recent years have witnessed tremendous interest in understanding and predicting information spread on social media platforms such as Twitter, Facebook, etc. Existing diffusion prediction methods primarily exploit the ...
详细信息
ISBN:
(纸本)9781450368223
Recent years have witnessed tremendous interest in understanding and predicting information spread on social media platforms such as Twitter, Facebook, etc. Existing diffusion prediction methods primarily exploit the sequential order of influenced users by projecting diffusion cascades onto their local social neighborhoods. However, this fails to capture global social structures that do not explicitly manifest in any of the cascades, resulting in poor performance for inactive users with limited historical activities. In this paper, we present a novel variational autoencoder framework (Inf-VAE) to jointly embed homophily and influence through proximity-preserving social and position-encoded temporal latent variables. To model social homophily, Inf-VAE utilizes powerful graph neural network architectures to learn social variables that selectively exploit the social connections of users. Given a sequence of seed user activations, Inf-VAE uses a novel expressive co-attentive fusion network that jointly attends over their social and temporal variables to predict the set of all influenced users. Our experimental results on multiple real-world social network datasets, including Digg, Weibo, and Stack-Exchanges demonstrate significant gains (22% MAP@10) for Inf-VAE over state-of-the-art diffusion prediction models;we achieve massive gains for users with sparse activities, and users who lack direct social neighbors in seed sets.
autoencoder-based communication systems show great potential due to the fast development of Deep Learning. In this letter, we propose an end-to-end communication structure for transmitting spatial-time symbol blocks. ...
详细信息
ISBN:
(纸本)9781728172361
autoencoder-based communication systems show great potential due to the fast development of Deep Learning. In this letter, we propose an end-to-end communication structure for transmitting spatial-time symbol blocks. For simulation, we build up three schemes of multi-input multi-output(MIMO) systems respectively. The results show that the proposed autoencoder structures are able to learn a regular constellation alphabet and attain a lower bit error rate than conventional communication systems in low signal-to-noise(SNR) situation.
Deep autoencoders (AEs) have recently been applied to the blind hyperspectral unmixing task to estimate endmembers and their corresponding abundances simultaneously. The objective of an autoencoder is to reconstruct a...
详细信息
ISBN:
(纸本)9781510638808
Deep autoencoders (AEs) have recently been applied to the blind hyperspectral unmixing task to estimate endmembers and their corresponding abundances simultaneously. The objective of an autoencoder is to reconstruct an input data matrix unsupervisedly with an encoder network and a decoder network. Applied to spectral unmixing, an appropriately constrained (e.g. abundance non-negativity and abundance sum-to-one) autoencoder can be trained such that the activations of the final layer of the encoder and the weights of the decoder form abundances and endmember signatures, respectively. In this paper, we present a novel regularization technique for autoencoder-based hyperspectral unmixing. We propose the inclusion of a generative adversarial network (GAN) joint training objective to condition the decoder to generalize to unseen abundance mixtures. In addition to regularizing the endmember weights of the decoder, this approach has the benefit of explicitly modeling the prior distribution of hyperspectral pixels for a given scene as the abundance output of the generator. The benefit of the proposed strategy as compared to a baseline autoencoder and a Gaussian-dropout-regularized autoencoder is evaluated on synthetic and real data sets, demonstrating that it can produce endmember estimates closer to the ground truth.
Road safety has always been a major concern, where a variety of competences is involved, ranging from government and local authorities, medical caregivers and other service provides. Prompt intervention in emergency c...
详细信息
ISBN:
(纸本)9781728152004
Road safety has always been a major concern, where a variety of competences is involved, ranging from government and local authorities, medical caregivers and other service provides. Prompt intervention in emergency cases is one of the key factors to minimize damages. Therefore, real-time surveillance is proposed as an efficient means to detect problems on roads. Video surveillance alone is not enough to detect serious accidents, since any hazardous behavior on the road may be confused with an accident, which may lead to many wrong alarms. Instead, audio processing has the potential to recognize sounds coming from different sources, such as crashes, tire skidding, harsh braking, etc. Since a few years, deep learning has become the state of the art for audio events detection. However, the usual dominance of absence of events in road surveillance would make a bias in the training process. Therefore, a novel method to initialize the neural network's weights using an autoencoder trained only on event-related data is used to balance the data distribution.
Video summarization is an important tool considering the amount of data to analyze. Techniques in this area aim to yield synthetic and useful visual abstraction of the videos contents. Hence, in this paper we present ...
详细信息
ISBN:
(纸本)9789897584022
Video summarization is an important tool considering the amount of data to analyze. Techniques in this area aim to yield synthetic and useful visual abstraction of the videos contents. Hence, in this paper we present a new summarization algorithm, based on image features, which is composed by the following steps: (i) Query video processing using cosine similarity metric and total variation smoothing to identify classes in the query sequence;(ii) With this result, build a labeled training set of frames;(iii) Generate the unlabeled training set composed by samples of the video database;(iv) Training a deep semi-supervised autoencoder;(v) Compute the K-means for each video separately, in the encoder space, with the number of clusters set as a percentage of the video size;(vi) Select key-frames in the K-means clusters to define the summaries. In this methodology, the query video is used to incorporate prior knowledge in the whole process through the obtained labeled data. The step (iii) aims to include unknown patterns useful for the summarization process. We evaluate the methodology using some videos from OPV video database. We compare the performance of our algorithm with the VSum. The results indicate that the pipeline was well succeed in the summarization presenting a F-score value superior to VSum.
This paper proposes a strategy for the stock market closing price prediction-by-prediction using the autoencoder long short-term memory (AE-LSTM) networks. To integrate technical analysis with deep learning methods, t...
详细信息
ISBN:
(纸本)9781728172965
This paper proposes a strategy for the stock market closing price prediction-by-prediction using the autoencoder long short-term memory (AE-LSTM) networks. To integrate technical analysis with deep learning methods, technical indicators and oscillators are added to the raw dataset as features. The wavelet transformation is used as a noise-removal technique in the stock index. Anomaly detection in dataset is also performed through the z-score method. First, the autoencoder is trained to represent the data. Then, the encoder extracts feature and puts them into the LSTM network for predicting the closing price of the stock index. Afterwards, the system predicts subsequently based on the previous predictions. To evaluate the theoretical results, the proposed method is experimented on the standard and poor's 500 (S&P 500) stock market index through several simulation studies. To analyze the results, several performance criteria are used to compare the results with the generative adversarial network (GAN). The simulation studies are conducted to show the effectiveness of the proposed method in the Python environment, and the results show that the proposed prediction-by-prediction method outperforms GAN in terms of daily adjusted closing price prediction.
Deep learning (DL) based autoencoder (AE) has been proposed recently as a promising, and potentially disruptive Physical Layer (PHY) design for beyond-5G communication systems. Compared to a traditional communication ...
详细信息
ISBN:
(纸本)9783030457785;9783030457778
Deep learning (DL) based autoencoder (AE) has been proposed recently as a promising, and potentially disruptive Physical Layer (PHY) design for beyond-5G communication systems. Compared to a traditional communication system with a multiple-block structure, the DL based AE provides a new PHY paradigm with a pure data-driven and end-to-end learning based solution. However, significant challenges are to be overcome before this approach becomes a serious contender for practical beyond-5G systems. One of such challenges is the robustness of AE under interference channels. In this paper, we first evaluate the performance and robustness of an AE in the presence of an interference channel. Our results show that AE performs well under weak and moderate interference condition, while its performance degrades substantially under strong and very strong interference condition. We further propose a novel online adaptive deep learning (ADL) algorithm to tackle the performance issue of AE under strong and very strong interference, where level of interference can be predicted in real time for the decoding process. The performance of the proposed algorithm for different interference scenarios is studied and compared to the existing system using a conventional DL-assist AE through an offline learning method. Our results demonstrate the robustness of the proposed ADL-assist AE over the entire range of interference levels, while existing AE fail to perform in the presence of strong and very strong interference. The work proposed in this paper is an important step towards enabling AE for practical 5G and beyond communication systems with dynamic and heterogeneous interference.
暂无评论