The clinical diagnosis of eye disorders including diabetic retinopathy relies heavily on retinal vessel segmentation. CNN-based methods are the preferred approaches for retina vessel segmentation in recent years, but ...
详细信息
ISBN:
(纸本)9781665429238
The clinical diagnosis of eye disorders including diabetic retinopathy relies heavily on retinal vessel segmentation. CNN-based methods are the preferred approaches for retina vessel segmentation in recent years, but they are data hungry and prone to overfitting on the training set and achieving suboptimal results on the validation set or the test set. Taking this into consideration, we propose to integrate a variational autoencoder reconstruction branch to pose extra regularization on the shared encoder and increase the generalization ability of networks. Furthermore, to deal with the unbalanced vessel scale distribution, a multi-scale context extractor is carefully designed, which employed the regular convolution and dilated convolution to extract multi-scale context and utilized different fusion method to obtain better complementary features. Extensive experiment results demonstrate that our proposed method achieves comparable state-of-the-art performance on the popular datasets: DRIVE and CHASEDB1.
variational inference provides a way to approximate probability densities. It does so by optimizing an upper or a lower bound on the likelihood of the observed data (the evidence). The classic variational inference ap...
详细信息
ISBN:
(纸本)9781665462839
variational inference provides a way to approximate probability densities. It does so by optimizing an upper or a lower bound on the likelihood of the observed data (the evidence). The classic variational inference approach suggests to maximize the Evidence Lower BOund (ELBO). Recent proposals suggest to optimize the variational R ' enyi bound (VR) and. upper bound. However, these estimates are either biased or difficult to approximate, due to a high variance. In this paper we introduce a new upper bound (termed VRLU) which is based on the existing variational R ' enyi bound. In contrast to the existing VR bound, the Monte Carlo (MC) approximation of the VRLU bound is unbiased. Furthermore, we devise a (sandwiched) upper-lower bound variational inference method (termed VRS) to jointly optimize the upper and lower bounds. We present a set of experiments, designed to evaluate the new VRLU bound, and to compare the VRS method with the classic VAE and the VR methods over a set of digit recognition tasks. The experiments and results demonstrate the VRLU bound advantage, and the wide applicability of the VRS method.
We propose a VAE-GAN network with a two-channel decoder for addressing multiple image-to-video translation tasks, i.e., generating multiple videos of different categories by a single model. We consider this image-to-v...
详细信息
ISBN:
(数字)9783031138706
ISBN:
(纸本)9783031138706;9783031138690
We propose a VAE-GAN network with a two-channel decoder for addressing multiple image-to-video translation tasks, i.e., generating multiple videos of different categories by a single model. We consider this image-to-video translation as a video generation task rather than a video prediction that needs multiple frames as input. After training, the model only requires the first frame of the video and its corresponding attribute to generate the required video. The advantage of combining the variational autoencoder (VAE) and Generative Adversarial Network (GAN) is to avoid the shortcomings of both: VAE components can give rise to blur, and unstable gradients caused by the GAN. Extensive qualitative and quantitative experiments are conducted on the MUG [1] dataset. We draw the following conclusions from this empirical study: compared with state-of-the-art approaches, our approach (VAE-GAN) exhibits significant improvements in generative capability.
Dimensionality reduction is a common tool for visualization and inference of population structure from genotypes, but popular methods either return too many dimensions for easy plotting (PCA) or fail to preserve globa...
详细信息
Dimensionality reduction is a common tool for visualization and inference of population structure from genotypes, but popular methods either return too many dimensions for easy plotting (PCA) or fail to preserve global geometry (t-SNE and UMAP). Here we explore the utility of variational autoencoders (VAEs)-generative machine learning models in which a pair of neural networks seek to first compress and then recreate the input data-for visualizing population genetic variation. VAEs incorporate nonlinear relationships, allow users to define the dimensionality of the latent space, and in our tests preserve global geometry better than t-SNE and UMAP. Our implementation, which we call popvae, is available as a command-line python program at ***/kr-colab/popvae. The approach yields latent embeddings that capture subtle aspects of population structure in humans and Anopheles mosquitoes, and can generate artificial genotypes characteristic of a given sample or population.
Estimating the gradients of stochastic nodes in stochastic computational graphs is one of the crucial research questions in the deep generative modeling community, which enables gradient descent optimization on neural...
详细信息
Estimating the gradients of stochastic nodes in stochastic computational graphs is one of the crucial research questions in the deep generative modeling community, which enables gradient descent optimization on neural network parameters. Stochastic gradient estimators of discrete random variables, such as the Gumbel-Softmax reparameterization trick for Bernoulli and categorical distributions, are widely explored. Meanwhile, other discrete distribution cases, such as the Poisson, geometric, binomial, multinomial, negative binomial, etc., have not been explored. This paper proposes a generalized version of the Gumbel-Softmax stochastic gradient estimator. The proposed method is able to reparameterize generic discrete distributions, not restricted to the Bernoulli and the categorical, and it enables learning on large-scale stochastic computational graphs with discrete random nodes. Our experiments consist of (1) synthetic examples and applications on variational autoencoders, which show the efficacy of our methods; and (2) topic models, which demonstrate the value of the proposed estimation in practice.
In the sporting world, baseball has been quicker to embrace the use of data analytics than any other sport, as detailed baseball statistics have become readily available in large and diverse quantities to the general ...
详细信息
ISBN:
(纸本)9781643680194;9781643680187
In the sporting world, baseball has been quicker to embrace the use of data analytics than any other sport, as detailed baseball statistics have become readily available in large and diverse quantities to the general public. Professional baseball teams use this data to develop game plans and evaluate players. In this work, we explore the latter by using a variational autoencoder (VAE), a special class of artificial neural networks. Specifically, we wish to relate a player's season-long batting statistics with the latent skills that a professional athlete needs to succeed in the MLB. In the growing field of sports analytics, we find this work incredibly important as it provides a novel, flexible, and powerful method to predict specific athletic skills based on years of recorded statistics.
The raw data utilized in training machine learning models faces a potential threat from membership inference attacks. To mitigate this risk, employing synthetic data instead of real data is proved effective in desensi...
详细信息
The raw data utilized in training machine learning models faces a potential threat from membership inference attacks. To mitigate this risk, employing synthetic data instead of real data is proved effective in desensitizing the information. We introduce a novel generative model, combining variational autoencoder and Generative Adversarial Network, to enhance privacy protection by generating synthetic data. In our approach, discrete variables are encoded by conditional generators, and sampling training is employed to ensure the distribution of synthetic data closely aligning with the real data. The modification of the model structure prompts a refinement of the loss function. We leverage Wasserstein distance with gradient penalty and SNorm to keep the stability of the model training process. Experimental results demonstrate that the efficacy of our model surpasses existing state-of-the-art models in terms of data utility metrics. Notably, in the face of membership inference attacks, the similarity from the results indicates the difficulty when distinguish the real data from synthetic data. It means our model have highlighting capabilities for the privacy protection.
The task of relation extraction in natural language processing is to identify the relation between two specified entities in a sentence. However, the existing model methods do not fully utilize the word feature inform...
详细信息
The task of relation extraction in natural language processing is to identify the relation between two specified entities in a sentence. However, the existing model methods do not fully utilize the word feature information and pay little attention to the influence degree of the relative relation extraction results of each word. In order to address the aforementioned issues, we propose a relation extraction method based on self-attention mechanism (SPCNN-VAE) to solve the above problems. First, we use a multi-head self-attention mechanism to process word vectors and generate sentence feature vector representations, which can be used to extract semantic dependencies between words in sentences. Then, we introduce the word position to combine the sentence feature representation with the position feature representation of words to form the input representation of piecewise convolutional neural network (PCNN). Furthermore, to identify the word feature information that is most useful for relation extraction, an attention-based pooling operation is employed to capture key convolutional features and classify the feature vectors. Finally, regularization is performed by a variational autoencoder (VAE) to enhance the encoding ability of model word information features. The performance analysis is performed on SemEval 2010 task 8, and the experimental results show that the proposed relation extraction model is effective and outperforms some competitive baselines.
The dynamics and stability of the semi-submersible offshore platforms are significantly impacted by the degradation of the mooring system. Identifying structural integrity issues in mooring systems through a data-driv...
详细信息
The dynamics and stability of the semi-submersible offshore platforms are significantly impacted by the degradation of the mooring system. Identifying structural integrity issues in mooring systems through a data-driven approach is challenging due to the infrequency of damage events and the difficulties in recording them. To address these challenges, this study proposes the Time-Series variational Semi-Supervised Learning (TSVSSL) framework, which effectively bridges the gap between supervised and unsupervised learning by leveraging unlabelled data for damage detection. The proposed framework features a distinctive training procedure in which the encoder-decoder and classifier components are trained concurrently. This process produces a well-clustered latent representation that enhances damage detection and supports class-specific artificial data generation. A numerical study using simulated responses of a 5 MW semi-submersible FOWT under varying metocean conditions demonstrated that the proposed framework outperformed existing deep learning methods in damage detection, achieving superior accuracy, precision, recall, and F1 score. Further, a rejection sampling technique is also introduced to effectively generates artificial data that closely aligns with actual time series displacement response. The novelty of the proposed framework lies in its dual focus on damage detection and artificial data generation marking a significant advancement in the data-driven assessment of mooring systems.
Attribute graph clustering is a fundamental and challenging task in graph data mining, requiring the adequate utilization of both node attributes and graph structure. Recently, a series of graph clustering methods hav...
详细信息
暂无评论