Neural networks are used in many tasks today. One of them is the images processing. autoencoder is very popular neural networks for such problems. Denoising autoencoder is an important autoencoder because some tasks w...
详细信息
ISBN:
(纸本)9781728199573
Neural networks are used in many tasks today. One of them is the images processing. autoencoder is very popular neural networks for such problems. Denoising autoencoder is an important autoencoder because some tasks we need a preprocessed image to get less noisy result. This research describes ways to analyze noisy images produced by a physically-based render engine and how to reduce that noise. The results showed that the algorithms are logarithmic.
A big challenge existing in genetic functionality prediction is that genetic datasets comprise few samples but massive unclear structured features, i.e., 'large p, small N' problem. To tackle this problem, we ...
详细信息
ISBN:
(纸本)9781450368599
A big challenge existing in genetic functionality prediction is that genetic datasets comprise few samples but massive unclear structured features, i.e., 'large p, small N' problem. To tackle this problem, we propose Non-local Self-attentive autoencoder (NSAE) which applies attention-driven genetic variant modelling. The backbone attention layer captures long-range dependency relationship among cells (i.e., features) and thus allocates weights to construct attention maps based on cell significance. Utilizing attention maps, NSAE can effectively seize and leverage significant features in a non-local way from numerous cells. Our proposed NSAE outperforms the state-of-the-art algorithms on two genomics datasets from Roadmap projects. The visualization of the attention layer also validates NSAE's ability to highlight important features.
Blind hyperspectral unmixing has become an important task for hyperspectral applications. In this paper, we propose a dual branch autoencoder with a novel sparse prior to simultaneously extract endmembers and abundanc...
详细信息
ISBN:
(纸本)9781509066315
Blind hyperspectral unmixing has become an important task for hyperspectral applications. In this paper, we propose a dual branch autoencoder with a novel sparse prior to simultaneously extract endmembers and abundances from the raw HSI. The dual branch structure extends the linear mixing model by only modeling linear mixtures of the endmembers and treating the bilinear interactions as error. In this way, the proposed model doesn't require the assumptions of explicit forms of bilinear interactions. The proposed sparse prior, named as orthogonal sparse prior, is based on the key observation that the abundance vector of one pixel is very sparse, there are often no more than two non-zero elements. Different from the conventional norm-based sparse prior which assumes the abundance maps are independent, the orthogonal sparse prior explores the orthogonality between the abundance maps. Extensive experiments on two real datasets show that the proposed method significantly and consistently outperforms the compared state-of-the-art methods, with up to 50% improvements.
Many ever-increasingly complex engineering optimization problems fall into the class of High-dimensional Expensive Problems (HEPs), where fitness evaluations are very time-consuming. It is extremely challenging and di...
详细信息
ISBN:
(纸本)9781728185262
Many ever-increasingly complex engineering optimization problems fall into the class of High-dimensional Expensive Problems (HEPs), where fitness evaluations are very time-consuming. It is extremely challenging and difficult to produce promising solutions in high-dimensional search space. In this paper, an autoencoder-embedded Evolutionary Optimization (AEO) framework is proposed for the first time. As an efficient dimension reduction tool, an autoencoder is used to compress high-dimensional landscape to informative low-dimensional space. The search operation in this low-dimensional space can facilitate the population converge towards the optima more efficiently. To balance the exploration and exploitation ability during optimization, two sub-populations coevolve in a distributed fashion, where one is assisted by an autoencoder and the other undergoes a regular evolutionary process. The information between these two sub-populations are dynamically exchanged. The proposed algorithm is validated by testing several 200 dimensional benchmark functions. Compared with the state-of-art algorithms for HEPs, AEO shows extraordinarily high efficiency for these challenging problems.
This research introduces least square adversarial autoencoder (LSAA)-an autoencoder that is able to reconstruct data and also generate data that has characteristics similar to data distribution from the prior distribu...
详细信息
ISBN:
(纸本)9781728192796
This research introduces least square adversarial autoencoder (LSAA)-an autoencoder that is able to reconstruct data and also generate data that has characteristics similar to data distribution from the prior distribution LSAA uses least square generative adversarial network loss function on its discriminator. LSAA minimizes Pearson ? 2 divergence between the latent variable distribution and the prior distribution. In this research, a Python program is developed to model LSAA by utilizing MNIST data set and FashionMNIST data set. The program is implemented using PyTarch. All of the programming activities are carried out in the cloud environment provided by the Tokopedia-Universitas Indonesia AI Center, using DGR-1 (GPU Tesla V100) as its computing resource. The experimental results show that the mean squared error of LSAA for MNIST data set and FasbionMNIST data set are 0.0080 and 0.0099, respectively. Furthermore, the Frechet Inception Distance score of LSAA for MNIST data set and FashionMNIST data set are 11.1280 and 27.5737, respectively. These results indicate that the least square adversarial autoencoder is able to reconstruct the image properly and also able to generate images similar to the training samples.
Maintenance is the process of preserving the good condition of a system to ensure its reliability and availability to perform specific operations. The way maintenance is nowadays performed in industry is changing than...
详细信息
Maintenance is the process of preserving the good condition of a system to ensure its reliability and availability to perform specific operations. The way maintenance is nowadays performed in industry is changing thanks to the increasing availability of data and condition assessment methods. Soft sensors have been widely used over last years to monitor industrial processes and to predict process variables that are difficult to measured. The main objective of this study is to monitor and evaluate the condition of the compressor in a particular industrial gas turbine by developing a soft sensor following an autoencoder architecture. The data used to monitor and analyze its condition were captured by several sensors located along the compressor for around five years. The condition assessment of an industrial gas turbine compressor reveals significant changes over time, as well as a drift in its performance. These results lead to a qualitative indicator of the compressor behavior in long-term performance.
Melanocytes are skin cells that give color to the skin and form melanin color pigments. The unbalanced division and proliferation of these cells result in skin cancer. The early diagnosis and proper treatment of skin ...
详细信息
Melanocytes are skin cells that give color to the skin and form melanin color pigments. The unbalanced division and proliferation of these cells result in skin cancer. The early diagnosis and proper treatment of skin cancer are so important. In this scope, a novel model that relies upon the autoencoder, spiking, and convolutional neural networks is proposed to ensure a useful decision support tool in this study. The experiments were carried out on an open-access dataset called the ISIC skin cancer consisting of 1800 being and 1497 malignant tumor images. In the proposed approach, the dataset is reconstructed using the autoencoder model. The original dataset and structured dataset were trained and classified by the MobileNetV2 model that consists of residual blocks, and the spiking networks. The classification success rate of the study was 95.27%. As a result, it was seen that the autoencoder model and spiking networks contributed to enhancing the performance of the MobileNetV2 model. Thanks to the proposed model, a novel fully automated decision support tool with high sensitivity was ensured for skin cancer detection. (c) 2021 Elsevier Ltd. All rights reserved.
Sensing networks and their environments are essential in intelligent building systems because of their increasing dependency on operational data. Virtual sensing technology has been applied in building energy systems ...
详细信息
Sensing networks and their environments are essential in intelligent building systems because of their increasing dependency on operational data. Virtual sensing technology has been applied in building energy systems to provide the more reliable and informative sensing environments. However, conventional virtual sensors still have structural and practical limitations under the physical sensor absences and limited datasets. Existing virtual sensors are separately developed by modeling multiple input variables and a single target (Xs to 1'), which is the variable-level virtual sensor (VLVS);therefore, these virtual sensors cannot benefit by either using their target variable (1') or by considering other virtual sensors when developing the models. This can result in insufficient accuracy, particularly in the limited sensors. Herein, to overcome these limitations, a novel virtual sensing framework, system-level virtual sensing (SLVS), is proposed for building energy systems using an autoencoder. Two strategies are also proposed. The autoencoder-based SLVS with the two strategies was applied in a real operational district heating system. The first strategy showed an improved accuracy using a new assistance virtual sensor, which is derived by additional information and knowledge regarding system design, control, and devices. It could also overcome the training data dependency in the limited datasets. The second strategy provided a replacement function for the SLVS specialized for backup and a calibration effect for the existing VLVS. Thus, the results showed that the suggested SLVS can achieve multifunctional high-accuracy virtual sensing;the accuracies of 99.89%, 99.68%, and 97.91% were shown respectively for temperatures, pressures, and control signals.
The reciprocity of the wireless channel is the basis of the secret key generation on the physical layer as an emerging method for information security enhancement. Due to the duplex mode of communication systems, the ...
详细信息
ISBN:
(纸本)9781728173276
The reciprocity of the wireless channel is the basis of the secret key generation on the physical layer as an emerging method for information security enhancement. Due to the duplex mode of communication systems, the channel responses detected by two connected legitimate nodes are not completely reciprocal. In this paper, an efficient physical layer secret key generation scheme is proposed based on an autoencoder to extract reciprocal features from weakly correlated channel estimates. The encoder part after training is equipped at both nodes to generate features for quantization in various channel environments. Simulation results show lower mean squared error (MSE) and better mutual information of the extracted features comparing to the most efficient linear generation method, the principal component analysis (PCA) based one. It reveals the great potential to generate symmetric secret keys with better performance, particularly in low SNR scenarios.
Recently, autoencoder (AE)-based embedding approaches have achieved state-of-the-art performance in many tasks, especially in top-k recommendation with user embedding or node classification with node embedding. Howeve...
详细信息
ISBN:
(纸本)9781450370233
Recently, autoencoder (AE)-based embedding approaches have achieved state-of-the-art performance in many tasks, especially in top-k recommendation with user embedding or node classification with node embedding. However, we find that many real-world data follow the power-law distribution with respect to the data object sparsity. When learning AE-based embeddings of these data, dense inputs move away from sparse inputs in an embedding space even when they are highly correlated. This phenomenon, which we call polarization, obviously distorts the embedding. In this paper, we propose TRAP that leverages two-level regularizers to effectively alleviate the polarization problem. The macroscopic regularizer generally prevents dense input objects from being distant from other sparse input objects, and the microscopic regularizer individually attracts each object to correlated neighbor objects rather than uncorrelated ones. Importantly, TRAP is a meta-algorithm that can be easily coupled with existing AE-based embedding methods with a simple modification. In extensive experiments on two representative embedding tasks using six-real world datasets, TRAP boosted the performance of the state-of-the-art algorithms by up to 31.53% and 94.99% respectively.
暂无评论