Convolutional neural networks (CNN) have had unprecedented success in medical imaging and, in particular, in medical image segmentation. However, despite the fact that segmentation results are closer than ever to the ...
详细信息
Convolutional neural networks (CNN) have had unprecedented success in medical imaging and, in particular, in medical image segmentation. However, despite the fact that segmentation results are closer than ever to the inter-expert variability, CNNs are not immune to producing anatomically inaccurate segmentations, even when built upon a shape prior. In this paper, we present a framework for producing cardiac image segmentation maps that are guaranteed to respect pre-defined anatomical criteria, while remaining within the inter-expert variability. The idea behind our method is to use a well-trained CNN, have it process cardiac images, identify the anatomically implausible results and warp these results toward the closest anatomically valid cardiac shape. This warping procedure is carried out with a constrained variational autoencoder (cVAE) trained to learn a representation of valid cardiac shapes through a smooth, yet constrained, latent space. With this cVAE, we can project any implausible shape into the cardiac latent space and steer it toward the closest correct shape. We tested our framework on short-axis MRI as well as apical two and four-chamber view ultrasound images, two modalities for which cardiac shapes are drastically different. With our method, CNNs can now produce results that are both within the inter-expert variability and always anatomically plausible without having to rely on a shape prior.
Data linkage plays a crucial role in realizing big data's value but is often regarded as a threat to personal privacy. Regulations like GDPR requires users' consent on each specific use of data, which is not p...
详细信息
ISBN:
(纸本)9781665424769
Data linkage plays a crucial role in realizing big data's value but is often regarded as a threat to personal privacy. Regulations like GDPR requires users' consent on each specific use of data, which is not practical for data analyzers. In this study, we propose a way to address the problem by having a trustworthy third party collect data from two or more parties, then use the data to train one or more variational autoencoder (VAE) models to remove privacy and send them to the data providers. Using this model, the users express their consent to share data with a trustworthy party. The third party links data from various datasets together to build a variational autoencoder model that allows all parties to generate datasets with full attributes without revealing sensitive personal data. System architectures and machine learning accuracy of generated data sets are measured in this study.
We propose a new paradigm for maintaining speaker identity in dysarthric voice conversion (DVC). The poor quality of dysarthric speech can be greatly improved by statistical VC, but as the normal speech utterances of ...
详细信息
ISBN:
(纸本)9781713836902
We propose a new paradigm for maintaining speaker identity in dysarthric voice conversion (DVC). The poor quality of dysarthric speech can be greatly improved by statistical VC, but as the normal speech utterances of a dysarthria patient are nearly impossible to collect, previous work failed to recover the individuality of the patient. In light of this, we suggest a novel, two-stage approach for DVC, which is highly flexible in that no normal speech of the patient is required. First, a powerful parallel sequence-to-sequence model converts the input dysarthric speech into a normal speech of a reference speaker as an intermediate product, and a nonparallel, frame-wise VC model realized with a variational autoencoder then converts the speaker identity of the reference speech back to that of the patient while assumed to be capable of preserving the enhanced quality. We investigate several design options. Experimental evaluation results demonstrate the potential of our approach to improving the quality of the dysarthric speech while maintaining the speaker identity.
The demand of probabilistic time series forecasting has been recently raised in various dynamic system scenarios, for example, system identification and prognostic and health management of machines. To this end, we co...
详细信息
The demand of probabilistic time series forecasting has been recently raised in various dynamic system scenarios, for example, system identification and prognostic and health management of machines. To this end, we combine the advances in both deep generative models and state space model (SSM) to come up with a novel, data-driven deep probabilistic sequence model. Specifically, we follow the popular encoder-decoder generative structure to build the recurrent neural networks (RNN) assisted variational sequence model on an augmented recurrent input space, which could induce rich stochastic sequence dependency. Besides, in order to alleviate the inconsistency issue of the posterior between training and predicting as well as improving the mining of dynamic patterns, we (i) propose using a lagged hybrid output as input for the posterior at next time step, which brings training and predicting into alignment;and (ii) further devise a generalized auto-regressive strategy that encodes all the historical dependencies for the posterior. Thereafter, we first investigate the methodological characteristics of the proposed deep probabilistic sequence model on toy cases, and then comprehensively demonstrate the superiority of our model against existing deep probabilistic SSM models through extensive numerical experiments on eight system identification benchmarks from various dynamic systems. Finally, we apply our sequence model to a real-world centrifugal compressor forecasting problem, and again verify its outstanding performance by quantifying the time series predictive distribution.
Efficient and accurate real-time forecasting of national spatial ozone distribution is critical to the provision of effective early warning. Traditional numerical air quality models require a high computational cost a...
详细信息
Efficient and accurate real-time forecasting of national spatial ozone distribution is critical to the provision of effective early warning. Traditional numerical air quality models require a high computational cost associated with running large-scale numerical simulations. In this work, we introduce a hybrid model (VAE-GAN) combining a generative adversarial network (GAN) with a variational autoencoder (VAE) to learn the dynamic ozone distributions in spatial and temporal spaces. The VAE-GAN model can not only decipher the complex nonlinear relationship between the inputs (the past states/ozone and meteorological factors) and outputs (ozone), but also provide ozone forecasts for a long lead-time beyond the training period. The performance of VAE-GAN is demonstrated in hourly and daily spatio-temporal ozone forecasts over China. The training datasets from 2013 to 2017 and validation datasets from 2018 to 2019 are the collection of data from the air quality reanalysis datasets. With the use of VAE, large dataset sizes are decreased by three orders of magnitude, enabling hourly and daily forecasts to be computed in seconds. Results show that the VAE-GAN achieves a reasonable accuracy in the prediction of both the spatial and temporal evolution patterns of hourly and daily ozone fields, as compared to the Nested Air Quality Prediction Modeling System (commonly used in China), the reanalysis data and observations during the validation period. Thus, the VAE-GAN is a cost-effective tool for large data-driven predictions, which can potentially reinforce air pollution prediction efforts in providing risk assessment and management in a timely manner.
Deep generative models for graphs are promising for being able to sidestep expensive search procedures in the huge space of chemical compounds. However, incorporating complex and non-differentiable property metrics in...
详细信息
ISBN:
(数字)9783030368029
ISBN:
(纸本)9783030368029;9783030368012
Deep generative models for graphs are promising for being able to sidestep expensive search procedures in the huge space of chemical compounds. However, incorporating complex and non-differentiable property metrics into a generative model remains a challenge. In this work, we formulate a differentiable objective to regularize a variational autoencoder model that we design for graphs. Experiments demonstrate that the regularization performs excellently when used for generating molecules since it can not only improve the performance of objectives optimization task but also generate molecules with high quality in terms of validity and novelty.
Prompt identification of structural damage is essential for effective postdisaster responses. To this end, this paper proposes a deep neural network (DNN)-based framework to identify seismic damage based on structural...
详细信息
Prompt identification of structural damage is essential for effective postdisaster responses. To this end, this paper proposes a deep neural network (DNN)-based framework to identify seismic damage based on structural response data recorded during an earthquake event. The DNN in the proposed framework is constructed by variational autoencoder, which is one of the self-supervised DNNs that can construct the continuous latent space of the input data by learning probabilistic characteristics. The DNN is trained using the flexibility matrices obtained by operational modal analysis (OMA) of simulated structural responses of the target structure under the undamaged state. To consider the load-dependency of OMA results, the undamaged state of the structure is represented by the flexibility matrix, which is closest to that obtained from the measured seismic response in the latent space. The seismic damage of each member is then estimated based on the difference between the two matrices using the flexibility disassembly method. As a numerical example, the proposed method is applied to a 5-story, 5-bay steel frame structure for which structural analyses are first performed under artificial ground motions to create train and test datasets. The proposed framework is verified with the near-real-time simulation using ground motions of El Centro and Kobe earthquakes. The example demonstrates that the proposed DNN-based method can identify seismic damage accurately in near-real-time.
Collaborative filtering (CF) methods based on graph convolutional network (GCN) and autoencoder (AE) achieve outstanding performance. But the GCN-based CF methods suffer from information loss problems, which are cause...
详细信息
Collaborative filtering (CF) methods based on graph convolutional network (GCN) and autoencoder (AE) achieve outstanding performance. But the GCN-based CF methods suffer from information loss problems, which are caused by information lossy initialization and using low-order Chebyshev Polynomial to fit the graph convolution kernel. And the AE-based CF methods obtain the prediction results by reconstructing the user-item interaction matrix, which does not conduct deep excavation of the behavior patterns, resulting in the limited-expression ability. To solve the above problems, we propose variational autoencoder-Enhanced Graph Convolutional Network (VE-GCN) for CF. Specifically, we use a variational autoencoder (VAE) to compress the interactive behavior patterns as the prior information of GCN to achieve sufficient learning, thus alleviating the information lossy initialization problem. And then the generalized graph Laplacian convolution kernel is proposed in GCN to handle the high-frequency information loss problem caused by Chebyshev Polynomial fitting in the GCN-based CF. To the best of our knowledge, VE-GCN is a feasible method to handle the information loss problems mentioned above in GCN-based CF for the first time. Meanwhile, the structure of GCN is optimized by removing redundant feature transformation and nonlinear activation function, and using DenseGCN to complete multi-level information interaction. Experiments on four real-world datasets show that the VE-GCN achieves state-of-the-art performance. (c) 2022 Elsevier Ltd. All rights reserved.
Airspace design is subject to a multitude of constraints, which are mainly driven by the concern to keep the risk of mid-air collision below a target level of safety. For that purpose, Monte Carlo simulation methods c...
详细信息
Airspace design is subject to a multitude of constraints, which are mainly driven by the concern to keep the risk of mid-air collision below a target level of safety. For that purpose, Monte Carlo simulation methods can be applied to estimate aircraft conflict probability but require the accurate generation of artificial trajectories. Generative models allow to generate an infinite number of trajectories for air traffic procedures where only few observations are available. The generated trajectories must not only resemble observed trajectories in terms of statistical distributions but they should stay flyable and consider uncertainty due to weather, air traffic control, aircraft performances, or human factors. This paper focuses on the generation problem, and its main contribution lies in the adaptation of the variational autoencoder structure to the problem of 4 -dimensional aircraft trajectories modelling using Temporal Convolutional Networks and a prior distribution composed of a variational Mixture of Posteriors (VampPrior). The proposed model has been trained on trajectories in the Terminal Manoeuvre Area of Zurich airport, which have a particularly high degree of variability as air traffic controllers often take actions that deviate aircraft from the nominal approach procedure. The model has demonstrated great abilities to take into account such amount of uncertainty. Regarding metrics that evaluate the estimation of the statistical distribution of the observed trajectories, and the flyability of the generated ones, the proposed method outperforms traditional statistical methods by being able to generate more complex and realistic trajectories.
作者:
Hu, CongSong, Xiao-NingJiangnan Univ
Sch Artificial Intelligence & Comp Sci Wuxi 214122 Jiangsu Peoples R China Jiangnan Univ
Jiangsu Prov Engn Lab Pattern Recognit & Computat Wuxi 214122 Jiangsu Peoples R China Minjiang Univ
Fujian Prov Key Lab Informat Proc & Intelligent C Fuzhou 350121 Peoples R China
To tackle the problem of semi-supervised learning (SSL), we propose a new autoencoder-based deep model. Ladder networks (LN) is an autoencoder-based method for representation learning which has been successfully appli...
详细信息
To tackle the problem of semi-supervised learning (SSL), we propose a new autoencoder-based deep model. Ladder networks (LN) is an autoencoder-based method for representation learning which has been successfully applied on unsupervised learning and semi-supervised learning. However, It ignores the manifold information of high-dimensional data and usually achieves unmeaning features which are very difficult to use in the subsequent tasks, such as prediction and recognition. To these issues, we proposed Graph Regularized variational Ladder Networks (GRVLN), which explicitly and implicitly employs the manifold structure of data. Our contributions can be summarized as two folds: (1) Graph regularization is used to build all decoder layers, which explicitly promotes the manifold learning via graph laplacian matrixs;(2) variational autoencoder is used as the backbone instead of traditional autoencoder in the encoder layers for implicitly learning the manifold structure of data distribution. Compared with ladder networks and other autoencoder-based methods, GRVLN achieves superior performance in semi-supervised classification tasks. Experimental results show that our method also has a comparable performance with state-of-the-art methods on several benchmark data sets.
暂无评论