Hyperspectral unmixing is to decompose the mixed pixels into pure spectral signatures (endmembers) and their proportions (abundances). Recently, deep learning-based methods have been applied to ...enhance the representation ability of unmixing models by extracting joint spatial-spectral characteristics of the hyperspectral data. However, most deep learning based-unmixing methods usually conduct global smoothing by convolutions on the whole hyperspectral imagery, which may ignore the variations within the imagery and result in oversmoothing. In this article, we propose a deep network for hyperspectral unmixing based on a new global-local smoothing autoencoder (GLA). GLA is an unsupervised model, which aims at exploring the local homogeneity and the global self-similarity of hyperspectral imagery. The proposed GLA network mainly includes two modules: a Local Continuous conditional random field Smoothing (LCS) module and a global recurrent smoothing (GRS) module. In LCS, we propose a conditional random field-based smoothing strategy to describe the joint spatial-spectral information within a local homogeneity region, which also reduces the risk of abundance maps boundary blurry. In GRS, we follow the self-similarity assumption for hyperspectral imagery and develop a recurrent neural network structure to exploit potential long-distance dependency relationships among pixels. The GLA is compared with several state-of-the-art unmixing methods on both real and synthetic data, and the abundance estimation results indicate that our method is promising. We will publish the code of GLA if this article has the honor to be accepted.
Representation learning is a crucial and challenging task within multimodal sentiment analysis. Effective multimodal sentiment representations contain two key aspects: consistency and difference. ...However, the state-of-the-art multimodal sentiment analysis approaches failed to capture the difference and consistency of sentiment information across diverse modalities. To address the multimodal sentiment representation problem, we propose an autoencoder-based self-supervised learning framework. In the pre-training stage, an autoencoder is designed for each modality, leveraging unlabeled data to learn richer sentiment representations for each modality through sample reconstruction and modality consistency detection tasks. In the fine-tuning stage, the pre-trained autoencoder is injected into MulT (AE-MT) and enhance the model's ability to extract deep sentiment information by incorporating a contrastive learning auxiliary task. Our experiments on the popular Chinese sentiment analysis benchmark (CH-SIMS v2.0) and English sentiment analysis benchmark (MOSEI) demonstrate significant gains over baseline models.
•A novel pixel-level mask autoencoder (PLMAE) is proposed to build a high-performance image recovery mechanism.•The idea of channel coding is used in the data embedding mechanism to take full ...advantage of PLMAE.•Compared to state-of-the-art methods, the proposed method significantly improves the embedding rate.
In the study of vacating-room-after-encryption reversible data hiding in encrypted images (VRAE RDHEI), pixel prediction is an important mechanism to achieve reversibility, which has a crucial impact on the capacity and fidelity. In this paper, we propose a novel pixel-level masked autoencoders (PLMAE) as a high-performance pixel predictor for RDHEI. Unlike the original masked autoencoders (MAE), PLMAE focuses on pixel-level reconstruction rather than semantic patch-level reconstruction. The purpose of PLMAE is to spare more carrier pixels while maintaining relatively high prediction accuracy, thereby improving the RDHEI capacity. Based on PLMAE, a novel RDHEI method is proposed. In the proposed method, the data hider encodes the secret data using a polar code and then embeds the encoded data. After the image is decrypted, the receiver considers the carrier pixels as masked pixels, predicts the original states of the carrier pixels using PLMAE to extract the secret data, and then decodes the secret data and recovers the image based on the decoding results. The experimental results demonstrate that the proposed method in this paper can achieve better performance than the existing methods.
The Denoising Autoencoder (DAE) enhances the flexibility of data stream method in exploiting unlabeled samples. Nonetheless, the feasibility of DAE for data stream analytic deserves in-depth study ...because it characterizes a fixed network capacity which cannot adapt to rapidly changing environments. Deep evolving denoising autoencoder (DEVDAN), is proposed in this paper. It features an open structure in the generative phase and the discriminative phase where the hidden units can be automatically added and discarded on the fly. The generative phase refines the predictive performance of discriminative model exploiting unlabeled data. Furthermore, DEVDAN is free of the problem-specific threshold and works fully in the single-pass learning fashion. We show that DEVDAN can find competitive network architecture compared with state-of-the-art methods on the classification task using ten prominent datasets simulated under the prequential test-then-train protocol.
•A novel semi-supervised fault diagnosis method is proposed.•The model can be trained using both labeled and unlabeled data simultaneously.•The performance of the proposed method is experimentally ...validated on two kinds of facilities.
Accurate fault diagnosis is critical to the safe and reliable operation of rotating machinery. Intelligent fault diagnosis techniques based on deep learning have recently gained increasing attention due to their ability to rapidly and efficiently extract features from data and provide accurate diagnosis results. Most of the successes achieved by the state-of-the-art fault diagnosis methods are obtained through supervised learning, which requires a substantial set of labeled data. To reduce the dependence of the fault diagnosis method on labeled data and make full use of the more abundant unlabeled data, a semi-supervised fault diagnosis method called hybrid classification autoencoder is proposed in this paper. This newly designed model utilizes a softmax classifier to directly diagnose the health condition based on the encoded features from the autoencoder. The commonly used mean square error (MSE) of unsupervised autoencoder is also modified to adopt the labels of data, therefore the model can be trained using the labeled and unlabeled data simultaneously. The proposed method is validated by a motor bearing dataset and an industrial hydro turbine dataset. The results show that the proposed method can obtain fairly high diagnosis accuracies and surpass the existing methods on a very small fraction of labeled data.
Deep learning, in particular the deep convolutional neural networks, has received increasing interests in face recognition recently, and a number of deep learning methods have been proposed. This ...paper summarizes about 330 contributions in this area. It reviews major deep learning concepts pertinent to face image analysis and face recognition, and provides a concise overview of studies on specific face recognition problems, such as handling variations in pose, age, illumination, expression, and heterogeneous face matching. A summary of databases used for deep face recognition is given as well. Finally, some open challenges and directions are discussed for future research.
Display omitted
•Presents a comprehensive survey of deep learning based face recognition methods.•Categorizes various approaches systematically.•Focuses on architectures, loss functions and other related things.•Includes specific issues on pose, expression, 3D, heterogeneous matching, etc.•Contains a broad review of related face databases.
The sparsity-driven technique is a widely used tool to solve the synthetic aperture radar (SAR) imaging problem. However, it always encounters sensitivity to motion errors. To solve this problem, ...this article proposes a new deep neural network architecture, i.e., the sparse autoencoder network (SAE-Net). The proposed SAE-Net is designed to implement SAR imaging and autofocus simultaneously. In SAE-Net, the encoder transforms the SAR echo into an imaging result, and the decoder regenerates the SAR echo using the obtained imaging result. The encoder is designed by the unfolded alternating direction method of multipliers (ADMM), while the decoder is formulated into a linear mapping. The joint reconstruction loss and the entropy loss are utilized to guide the training of the SAE-Net. Notably, the algorithm operates in a totally self-supervised form and requires no other training dataset. The methodology was tested on both synthetic and real SAR data. These tests show that the proposed architecture outperforms other state-of-the-art autofocus methods in sparsity-driven SAR imaging applications.
In the past decades, recommendation systems have provided lots of valuable personalized suggestions for the users to address the problem of information over-loaded. Collaborative Filtering (CF) is ...one of the most commonly applied and successful recommendation approaches, which refers to using the preferences of groups with similar interests to recommend information to other users. Recently, in addition to the traditional matrix factorization techniques, deep learning methods have been proposed to learn more abstract and higher-level representations for recommendation. However, most previous deep recommendation methods learn the higher-level feature representations of users and items through an identical model structure, which ignores the different characteristics of the user-based and item-based data. In addition, the rating matrix is usually sparse which may result in a significant degradation of recommendation performance. To address these problems, we propose a representation learning method with Collaborative Autoencoder for Personalized Recommendation (CAPR for short). In this method, user-based and item-based feature representations are learned by two different autoencoders for capturing different features of the data. Meanwhile, items’ attributions are combined into the feature representations with semi-autoencoder for alleviating the sparsity problem. Extensive experimental results confirm the effectiveness of our proposed method compared to other state-of-the-art matrix factorization methods and deep recommendation methods.
•Two different autoencoders are used to capture characteristics for users and items.•Manifold regularization is integrated into autoencoder for user’s features learning.•The comprehensive experiments evaluate the effectiveness of our proposed method.
•A semi-supervised learning method named CAE-SGAN is proposed to classify surface defects of steels.•CAE-SGAN improves the performance of SGAN with limited training samples.•When training the ...discriminator of SGAN, the decoder network of CAE is not truncated.•CAE-SGAN is tested with sample images collected from three different steel production lines.•CAE-SGAN provides a better way to apply deep learning methods to some industrial scenes.
Defect inspection is extremely crucial to ensure the quality of steel surface. It affects not only the subsequent production, but also the quality of the end-products. However, due to the rare occurrence and appearance variations of defects, surface defect identification of steels has always been a challenging task. Recently, deep learning methods have shown outstanding performance in image classification, especially when there are enough training samples. Since most sample images of steel surface are unlabeled, a new semi-supervised learning method is proposed to classify surface defects of steels. The new method is named CAE-SGAN, as it is based on Convolutional Autoencoder (CAE) and semi-supervised Generative Adversarial Networks (SGAN). CAE-SGAN first trains a stacked CAE through massive unlabeled data. Considering the appearance variations of defects, the passthrough layer is used to help CAE extract fine-grained features. After CAE is trained, the encoder network of CAE is reserved as the feature extractor and fed into a softmax layer to form a new classifier. SGAN is introduced for semi-supervised learning to further improve the generalization ability of the new method. The classifier is trained with images collected from real production lines and images randomly generated by SGAN. Extensive experiments are carried out with samples captured from different steel production lines, and the results indicate that CAE-SGAN had yielded best performances compared with traditional methods. Especially for hot rolled plates, the classification rate is improved by around 16%.
In the chemical plants, data-driven process monitoring serves as a vital tool to ensure product quality and maintain production line safety. However, the accuracy of monitoring hinges directly upon ...the quality of process data. Given the inherently slow and complex nature of chemical processes, coupled with the potential for gross errors in process data leading to inaccuracies in model predictions, this paper proposes a method called Conditional Dynamic Variational Autoencoder combined with a Particle Filter (CDVAE-PF) for data reconciliation and subsequent process monitoring. CDVAE-PF leverages the capabilities of Conditional Dynamic Variational Autoencoder (CDVAE) to effectively model chemical process data in the presence of noise. This probabilistic model serves as the foundation for the Particle Filter (PF), which is employed for data reconciliation. Moreover, CDVAE-PF incorporates mechanisms to detect and rectify gross errors in process data, further enhancing its efficacy in data reconciliation. Subsequently, monitoring indices based on CDVAE are established to facilitate process monitoring. Through numerical simulations of a two-to-one variables Continuous Stirred Tank Reactor (CSTR) example and a fifteen-to-one variables dichloroethane distillation process from an actual chemical plant, CDVAE-PF demonstrates its effectiveness by reducing mean absolute error to 7.8 % and 12.8 % respectively in gross error data reconciliation. Moreover, in terms of monitoring performance, CDVAE-PF successfully mitigates misjudgments caused by gross errors, thereby significantly enhancing the reliability of process monitoring in chemical plants.
•Conditional dynamic variational autoencoder (CDVAE) is proposed.•CDVAE, a probabilistic model, is more robust for processes data with noise.•Gross errors based on CDVAE predictions are effectively detected and reconciled.•CDVAE combined with particle filters forms a dynamic data reconciliation scheme.•Reconciliation and monitoring results are presented via real chemical processes.