Denoising sound is essential for improving signal quality in various applications such as speech processing, sound event classification, and machine failure detection systems. This paper proposes an autoencoder method...
详细信息
ISBN:
(纸本)9798350323078
Denoising sound is essential for improving signal quality in various applications such as speech processing, sound event classification, and machine failure detection systems. This paper proposes an autoencoder method to remove two types of noise, Gaussian white noise, and environmental noise from water flow, from induction motor sounds. The method is trained and evaluated on a dataset of 246 sounds from the Machinery Fault Database (MAFAULDA). The denoising effectiveness is measured using the mean square error (MSE), which indicates that both noise types can be significantly reduced with the proposed method. The MSE is below or equal to 0.15 for normal operation sounds and misalignment sounds. This improvement in signal quality can facilitate further processing, such as induction motor operation classification. Overall, this work presents a promising approach for denoising machine sounds using an autoencoder, with potential for application in other industrial settings.
During the past decades, significant progress has been made in the field of artificial neural networks to process images (Convolutional Neural Networks), audio signals (Temporal Convolutional Networks), or textual inf...
详细信息
Due to the escalating utilization of communication networks and the prevailing occurrence of cyber attacks, intrusion detection systems (IDSs) have emerged as imperative components in network security. Machine learnin...
详细信息
ISBN:
(纸本)9798350310900
Due to the escalating utilization of communication networks and the prevailing occurrence of cyber attacks, intrusion detection systems (IDSs) have emerged as imperative components in network security. Machine learning (ML) and deep learning (DL) based IDSs have gained popularity due to their detection capability and adaptability. However, this type of schemes are susceptible to adversarial attacks, which involve minor perturbations to attack features causing misclassification. autoencoders (AEs) have proven effective in mitigating adversarial attacks in computer vision, but their capacity for enhancing IDSs remains relatively unexplored. In this paper, we focus on the use of AEs to detect adversarial network flows. Specifically, we propose an AE-enhanced IDS (AE-IDS) that leverages the power of AEs to improve the robustness of IDSs against adversarial attacks. Our experimental results indicate that AE-IDS outperforms the baseline schemes under investigation in terms of accuracy and detection rate. We believe that AE-IDS showcases the potential of using AEs to enhance the robustness of IDSs, providing improved security against sophisticated and evolving cyber threats.
Smart packaging machines incorporate various components (blades, motors, films) to accomplish the packaging process and are involved in almost all types of the manufacturing industry. Proper maintenance and monitoring...
详细信息
ISBN:
(纸本)9798350311259
Smart packaging machines incorporate various components (blades, motors, films) to accomplish the packaging process and are involved in almost all types of the manufacturing industry. Proper maintenance and monitoring of the components over time can help industries to maintain a sustainable production environment. On the contrary, a faulty system may degrade production efficiency and increase the cost. Smart packaging machines comprising several sensors can generate time series data and leverage data driven condition monitoring models to overcome faulty conditions. In this work, we have studied the application of autoencoder as a data driven condition monitoring tool for the predictive maintenance of packaging machines. The trained autoencoder on the new system's data can detect worn or degraded components over time. We have also used the Bayesian optimization algorithm to tune the hyper-parameters of the autoencoder for better predictive performance. Moreover, the reconstruction error is analyzed to identify the worn components in the packaging machine.
We propose a new latent factor conditional asset pricing model. Like Kelly, Pruitt, and Su (KPS, 2019), our model allows for latent factors and factor exposures that depend on covariates such as asset characteristics....
详细信息
We propose a new latent factor conditional asset pricing model. Like Kelly, Pruitt, and Su (KPS, 2019), our model allows for latent factors and factor exposures that depend on covariates such as asset characteristics. But, unlike the linearity assumption of KPS, we model factor exposures as a flexible nonlinear function of covariates. Our model retrofits the workhorse unsupervised dimension reduction device from the machine learning literature - autoencoder neural networks - to incorporate information from covariates along with returns themselves. This delivers estimates of nonlinear conditional exposures and the associated latent factors. Furthermore, our machine learning framework imposes the economic restriction of no-arbitrage. Our autoencoder asset pricing model delivers out-of-sample pricing errors that are far smaller (and generally insignificant) compared to other leading factor models. (c) 2020 Elsevier B.V. All rights reserved.
In recent advancements in text summarization, BERT has gained popularity for encoding documents. However, sentence-based extractive models often lead to redundant or uninformative phrases in the generated summaries. A...
详细信息
ISBN:
(纸本)9798350342734
In recent advancements in text summarization, BERT has gained popularity for encoding documents. However, sentence-based extractive models often lead to redundant or uninformative phrases in the generated summaries. Additionally, BERT, which is pretrained on sentence pairs rather than full documents, struggles to capture long-range dependencies present within a document. To overcome these challenges, we introduce DBERT-ELVA, a discourse-aware neural summarization model. DBERT-ELVA extracts sub-sentential discourse units, offering a more refined granularity for extractive selection, in contrast to traditional sentence-based approaches. To learn a compressed representation of these discourse units while capturing the interdependencies among them, an autoencoder is designed, utilizing Extreme Learning for improved generalization performance. Experimental evaluations on popular summarization benchmarks demonstrate that the proposed model significantly outperforms state-of-the-art methods, including other BERT-based models, by a substantial margin.
In smart and intelligent health care, smartphone sensor-based automatic recognition of human activities has evolved as an emerging field of research. In many application domains, deep learning (DL) strategies are more...
详细信息
ISBN:
(纸本)9798350334586
In smart and intelligent health care, smartphone sensor-based automatic recognition of human activities has evolved as an emerging field of research. In many application domains, deep learning (DL) strategies are more effective than conventional machine learning (ML) models, and human activity recognition (HAR) is no exception. In this paper, we propose a novel framework (CAEL-HAR), that combines CNN, autoencoder and LSTM architectures for efficient smartphone-based HAR operation. There is a natural synergy between the modeling abilities of LSTMs, autoencoders, and CNNs. While AEs are used for dimensionality reduction and CNNs are the best at automating feature extraction, LSTMs excel at modeling time series. Taking advantage of their complementarity, the proposed methodology combines CNNs, AEs, and LSTMs into a single architecture. We evaluated the proposed architecture using the UCI, WISDM public benchmark datasets. The simulation and experimental results certify the merits of the proposed method and indicate that it outperforms computing time, F1-score, precision, accuracy, and recall in comparison to the current state-of-the-art methods.
Orchard tree inventory has been an essential step to obtain up-to-date information for effective tree treatments and crop insurance purposes. Inventorying trees is often performed manually through fieldwork surveys, w...
详细信息
ISBN:
(纸本)9798350341393
Orchard tree inventory has been an essential step to obtain up-to-date information for effective tree treatments and crop insurance purposes. Inventorying trees is often performed manually through fieldwork surveys, which are generally time-consuming, costly, and subject to errors. Motivated by the latest advances in UAV imagery and deep learning, we propose a new framework for individual tree detection and health assessment. We adopt a divide-and-conquer approach to address the problem of orchard trees' health assessment in two stages. First, we build a tree detection model based on a hard negative mining strategy to improve object detection. In the second stage, we address the health classification problem using a new convolutional autoencoder architecture mainly designed to extract relevant features. The performed experiments demonstrate the robustness of the proposed framework for orchard tree health assessment from UAV images. In particular, our framework achieves an F1-score of 86.24% for tree detection and an overall accuracy of 98.06% for tree health assessment. Moreover, our work could be generalized for a wide range of UAV applications involving a detection/classification process.
Federated Learning (FL) is a promising collaborative training paradigm that utilizes decentralized on-device data. Using supervised learning approaches in FL-based network intrusion detection systems often leads to po...
详细信息
ISBN:
(纸本)9798350321814
Federated Learning (FL) is a promising collaborative training paradigm that utilizes decentralized on-device data. Using supervised learning approaches in FL-based network intrusion detection systems often leads to poor classification performance because of the highly imbalanced data with limited labeled network traffic anomalies from data collected by edge devices. Furthermore, detecting zero-day anomalies/attacks without a priori knowledge is difficult. Due to these constraints, unsupervised learning-based methods, such as autoencoders, which only use benign traffic to build the detection model, appear to be the desired choice to identify anomalous network traffic. In this work, we propose a Compressed autoencoder-based Federated Network (CAFNet) framework for network anomaly detection to deal with the labeled data scarcity issue while preserving data owner's privacy and reducing communication overhead. Our framework leverages the latent representation of autoencoders to capture important information in the input features of the distributed network devices and eliminate the transmission of redundant information (weights) during federated training. Our extensive experimental results with three publicly available network intrusion detection datasets show that our proposed framework can significantly lower communication cost up to 65% of the state-of-the-art model compression strategies used in traditional FL as well as achieves attack detection performance comparable to conventional FL framework.
Single-cell RNA sequencing (scRNA-seq) permits researchers to study the complex mechanisms of cell heterogeneity and diversity. Unsupervised clustering is of central importance for the analysis of the scRNA-seq data, ...
详细信息
Single-cell RNA sequencing (scRNA-seq) permits researchers to study the complex mechanisms of cell heterogeneity and diversity. Unsupervised clustering is of central importance for the analysis of the scRNA-seq data, as it can be used to identify putative cell types. However, due to noise impacts, high dimensionality and pervasive dropout events, clustering analysis of scRNA-seq data remains a computational challenge. Here, we propose a new deep structural clustering method for scRNA-seq data, named scDSC, which integrate the structural information into deep clustering of single cells. The proposed scDSC consists of a Zero-Inflated Negative Binomial (ZINB) model-based autoencoder, a graph neural network (GNN) module and a mutual-supervised module. To learn the data representation from the sparse and zero-inflated scRNA-seq data, we add a ZINB model to the basic autoencoder. The GNN module is introduced to capture the structural information among cells. By joining the ZINB-based autoencoder with the GNN module, the model transfers the data representation learned by autoencoder to the corresponding GNN layer. Furthermore, we adopt a mutual supervised strategy to unify these two different deep neural architectures and to guide the clustering task. Extensive experimental results on six real scRNA-seq datasets demonstrate that scDSC outperforms state-of-the-art methods in terms of clustering accuracy and scalability. Our method scDSC is implemented in Python using the Pytorch machine-learning library, and it is freely available at .
暂无评论