The industry is shifting from a selection between a few ready-made models by the customer towards highly customizable, individualized, made-on-demand final products. This also encompasses the footwear industry, where ...
详细信息
The industry is shifting from a selection between a few ready-made models by the customer towards highly customizable, individualized, made-on-demand final products. This also encompasses the footwear industry, where 3D printing presents a modern solution for customizable shoes. While fashionable shoes can already be produced by this method, this specific manufacturing process is still in an early stage, and process parameters and material behavior are not fully known. Therefore, the reject rate is currently higher as it is for conventionally made shoes and needs to be reduced. To achieve industrial scaling of this manufacturing process, an automatic detection of faulty printed shoes is needed, which can be later integrated into a self-learning and self-adjusting manufacturing process. In this contribution, we present a novel, two-step, deep learning-based method for detecting faults on printed shoes by using RGB images of the shoe. In the first step, the region of the shoe in the image is marked. Then, in the second step, this region is analyzed for anomalies using an autoencoder approach. Since the method uses regular RGB images, the integration effort into existing and ongoing process flows is low. The approach is validated using two different printed shoe parts - a white partial shoe with defects added artificially and an orange whole shoe, available with and without real defects. Promising results were achieved, showing a good detection rate for the orange shoe and a decent detection rate for the white shoe.
A novel big data-predictive control approach for nonlinear multi-timescale processes is presented in this paper. Multiple Dynamical Latent Variable autoencoders (DLVAEs) are employed to approximate multi-timescale dyn...
详细信息
A novel big data-predictive control approach for nonlinear multi-timescale processes is presented in this paper. Multiple Dynamical Latent Variable autoencoders (DLVAEs) are employed to approximate multi-timescale dynamics, utilizing timescale-based low-pass filtering and resampling of historical input-output data. The encoder in each DLVAE projects the nonlinear physical variable space onto a linear latent variable space, represented by a kernel space in behavioral system theory. During training, we not only impose kernel spaces and reconstruct data but also establish connections among latent variables from different DLVAEs at matching time-steps. Collectively, these multi-level latent variables span a wide prediction time horizon with limited (non-uniformly spaced) steps encompassing the current, near, and distant future. In online tracking control, we guide the latent variables from each DLVAE to their respective setpoints (derived from physical variable setpoints) while maintaining consistent physical variable values at matching time-steps, all within a linear framework.
Gas metal arc welding is widely used in industrial series production for joining aluminum. A lot of factors, such as instabilities and complex dependencies, influence the quality of the resulting welding seams. It is ...
详细信息
Gas metal arc welding is widely used in industrial series production for joining aluminum. A lot of factors, such as instabilities and complex dependencies, influence the quality of the resulting welding seams. It is challenging to identify the causes of welding defects, and the real reason is not always well understood. Ensuring the process stability helps production workers to increase the overall production efficiency. The process stability increases the process repeatability, so the welding performance is optimized and rejects are avoided. This paper presents a technique to detect process instabilities within the multivariate process variables automatically. An autoencoder architecture is implemented. The latent space of the autoencoder and reconstruction of the time series are used to detect process instabilities. Detected issues are visualized in a heatmap, including supportive metrics to describe deviations from the expected behavior. As a result, the proposed architecture supports process optimization and leads to an increase in production transparency.
With improved machine learning models, studies on bankruptcy prediction show improved accuracy. This paper proposes three relatively newly-developed methods for predicting bankruptcy based on real-life data. The resul...
详细信息
With improved machine learning models, studies on bankruptcy prediction show improved accuracy. This paper proposes three relatively newly-developed methods for predicting bankruptcy based on real-life data. The result shows among the methods (support vector machine, neural network with dropout, autoencoder), neural network with added layers with dropout has the highest accuracy. And a comparison with the former methods (logistic regression, genetic algorithm, inductive learning) shows higher accuracy.
The article presents the results of applying machine learning techniques to detect fraudulent banking transactions. The market of antifraud systems was studied. Ensemble methods for solving classification problem as w...
详细信息
The article presents the results of applying machine learning techniques to detect fraudulent banking transactions. The market of antifraud systems was studied. Ensemble methods for solving classification problem as well as dimensionality reduction techniques were examined. The proposed analysis procedure is based on the selection of the best machine learning model and the identification of the most significant features for detecting fraud. Results-based recommendations can be used in financial institutions as well as in other organizations, where it is required to identify and prevent entities’ fraudulent actions that pose a threat to the functioning of business processes and electronic systems. The proposed fraud detection methodology was implemented on the cloud-based analytical platform Statistical Analysis System (SAS) Viya.
Energy and moisture contents are important food chemical attributes. In the current study, a nondestructive Terahertz (THz) time-domain imaging system was first time used for evaluating the energy and moisture distrib...
详细信息
Energy and moisture contents are important food chemical attributes. In the current study, a nondestructive Terahertz (THz) time-domain imaging system was first time used for evaluating the energy and moisture distributions of sunflower seed kernels inside shells. For this task, a dual autoencoders (AE)-generative adversarial nets (GAN) spectral dehulling semi-supervised model was developed. The model could automatically learn the kernel information from the latent representations of the spectra of the intact seeds through adversarial learning to achieve feature disentanglement. Results indicated that the generated kernel images had similar features to the original kernel images and high-quality chemical distribution maps for energy and moisture contents of sunflower seed kernels inside shells were successfully obtained. As the current method took the advantage of the characteristics of THz imaging and selected a suitable deep learning algorithm, it has the potential to generalize for imaging other chemical substances of other dry shelled seeds or biological samples (moisture content and thickness below 15% and 5 mm, respectively).
Recurrent neural networks and exceedingly Long short-term memory (LSTM) have been investigated intensively in recent years due to their ability to model and predict nonlinear time-variant system dynamics. The present ...
详细信息
Recurrent neural networks and exceedingly Long short-term memory (LSTM) have been investigated intensively in recent years due to their ability to model and predict nonlinear time-variant system dynamics. The present paper delivers a comprehensive overview of existing LSTM cell derivatives and network architectures for time series prediction. A categorization in LSTM with optimized cell state representations and LSTM with interacting cell states is proposed. The investigated approaches are evaluated against defined requirements being relevant for an accurate time series prediction. These include short-term and long-term memory behavior, the ability for multimodal and multi-step ahead predictions and the according error propagation. Sequence-to-sequence networks with partially conditioning outperform the other approaches, such as bidirectional or associative networks, and are best suited to fulfill the requirements.
Scene classification of high-resolution remote sensing (HRRS) image is an important research topic and has been applied broadly in many fields. Deep learning method has shown its high potential to in this domain, owin...
详细信息
Scene classification of high-resolution remote sensing (HRRS) image is an important research topic and has been applied broadly in many fields. Deep learning method has shown its high potential to in this domain, owing to its powerful learning ability of characterizing complex patterns. However the deep learning methods omit some global and local information of the HRRS image. To this end, in this article we show efforts to adopt explicit global and local information to provide complementary information to deep models. Specifically, we use a patch based MS-CLBP method to acquire global and local representations, and then we consider a pretrained CNN model as a feature extractor and extract deep hierarchical features from full-connection layers. After fisher vector (FV) encoding, we obtain the holistic visual representation of the scene image. We view the scene classification as a reconstruction procedure and train several class-specific stack denoising autoencoders (SDAEs) of corresponding class, i.e., one SDAE per class, and classify the test image according to the reconstruction error. Experimental results show that our combination method outperforms the state-of-the-art deep learning classification methods without employing fine-tuning.
In this study, we propose a novel method for network-anomaly detection and failure-scale estimation using autoencoders, which are a type of neural network. The proposed method first divides the network into several gr...
详细信息
In this study, we propose a novel method for network-anomaly detection and failure-scale estimation using autoencoders, which are a type of neural network. The proposed method first divides the network into several groups. Subsequently, anomalies are detected using an autoencoder for each intergroup traffic, and the failure-scale is estimated from the number of autoencoders that have detected anomalies. We experimentally investigated anomaly detection during communication through a virtual network built using the network emulator Mininet and confirmed that the proposed method can successfully detect anomalies and estimate the failure scale.
Electric power SCADA (Supervisory Control and Data Acquisition) system gradually transforming from a separate private network to an open public network, seriously increases the vulnerability risk in electric power SCA...
详细信息
Electric power SCADA (Supervisory Control and Data Acquisition) system gradually transforming from a separate private network to an open public network, seriously increases the vulnerability risk in electric power SCADA. In order to assess the vulnerability risk in electric power SCADA system, the paper firstly uses Delphi method and AHP (Analytic Hierarchy Process) to build an index system of vulnerability risk assessment, to fully represent the vulnerability of electric power SCADA system. As index data of vulnerability risk assessment in power SCADA is characterized by strong relation and high dimensionality, the method of autoencoder is proposed to reduce dimensionality of index data by representing high-dimensional data in a low dimensional space. Auto encoder method can obtain the optimal initial weight in pre-training and then back-propagate error derivatives adjusting weights with the initial weights to minimize the reconstruction error finally getting the best reconstructed results. The paper conducts simulation experiments about reconstruction error in pre-training and fine-tuning process in MATLAB experimental platform, and the experimental results show that dimensional code received by reducing dimensionality of data can basically fully represent high-dimensional data. The lowdimensional code as input can significantly reduce the complexity in the construction of model of vulnerability risk assessment in Electric power SCADA system in later work.
暂无评论