We investigate the effect of variational autoencoder (VAE) based data anonymization and its ability to preserve anomalous subgroup properties. We present a Utility Guaranteed Deep Privacy (UGDP) system which casts exi...
详细信息
ISBN:
(纸本)9781509066315
We investigate the effect of variational autoencoder (VAE) based data anonymization and its ability to preserve anomalous subgroup properties. We present a Utility Guaranteed Deep Privacy (UGDP) system which casts existing anomalous pattern detection methods as a new utility measure for data synthesis. UGDP's approach shows that properties of an anomalous subset of records, identified in the original data set, are preserved through the anonymization of a VAE. This is despite the newly generated records being completely synthetic. More specifically, the Bias-Scan algorithm identifies a subgroup of records that are consistently over- (or under-) risked by a black-box classifier as an area of 'poor fit'. This scanning process is applied on both pre- and post- VAE synthesized data. The areas of poor fit (i.e. anomalous records) persist in both settings. We evaluate our approach using publicly available datasets from the financial industry. Our evaluation confirmed that the approach is able to produce synthetic datasets that preserved a high level of subgroup differentiation as identified initially in the original dataset. Such a distinction was maintained while having distinctly different records between the synthetic and original dataset.
Grasp planning and most specifically the grasp space exploration is still an open issue in robotics. This article presents a data-driven oriented methodology to model the grasp space of a multi-fingered adaptive gripp...
详细信息
Grasp planning and most specifically the grasp space exploration is still an open issue in robotics. This article presents a data-driven oriented methodology to model the grasp space of a multi-fingered adaptive gripper for known objects. This method relies on a limited dataset of manually specified expert grasps, and uses variational autoencoder to learn grasp intrinsic features in a compact way from a computational point of view. The learnt model can then be used to generate new non-learnt gripper configurations to explore the grasp space.
In this paper an optimization algorithm for time synchronization in telecommunication network is proposed based on VAE(variational Auto Encoder)framework. Firstly features are represented in latent space under propose...
详细信息
ISBN:
(纸本)9781728160429
In this paper an optimization algorithm for time synchronization in telecommunication network is proposed based on VAE(variational Auto Encoder)framework. Firstly features are represented in latent space under proposed framework while performance of synchronization network is measured and evaluated. Secondly optimization algorithm is further designed with which feature of abnormal samples and benchmark are adaptively merged for smooth adjustment with low risk in practical network operation. Meanwhile considering the characteristics as domain knowledge of synchronization network, a novel metric is adopted to reduce the fluctuation of adjustment. The simulation results verified that performance of synchronization network is significantly improved by optimization templates reconstructed through decoding part of VAE model. It is implied that prior knowledge of synchronization in latent space is introduced with certain interpret-ability for assessment of monitoring performance while optimization adjustment can be properly operated through novel metric proposed in this algorithm.
Speech separation plays an important role in a speech-related system since it can denoise, extract, and enhance speech signals. In recent years, many methods are proposed to separate the human voice of noise and other...
详细信息
ISBN:
(纸本)9783030630065;9783030630072
Speech separation plays an important role in a speech-related system since it can denoise, extract, and enhance speech signals. In recent years, many methods are proposed to separate the human voice of noise and other sounds. To separate the speech from a complicated signal, we propose a more powerful method by using a VAE model and then postprocessing with a bandpass filter. This combination can use to extract the original human speech in the mixture with not only high-frequency noise but also many different sounds. Our approach can be flexibly applied for the new background sounds.
The missing data issue is often found in real-world datasets and it is usually handled with imputation strategies that replace the missing values with new data. Recently, generative models such as variational Autoenco...
详细信息
ISBN:
(纸本)9781728169262
The missing data issue is often found in real-world datasets and it is usually handled with imputation strategies that replace the missing values with new data. Recently, generative models such as variational autoencoders have been applied for this imputation task. However, they were always used to perform the entire imputation, which has presented limited results when comparing to other state-of-the-art methods. In this work, a new approach called variational autoencoder Filter for Bayesian Ridge Imputation is introduced. It uses a variational autoencoder at the beginning of the imputation pipeline to filter the instances that are later fitted to a Bayesian ridge regression used to predict the new values. The approach was compared to four state-of-the-art imputation methods using 10 datasets from the healthcare context covering clinical trials, all injected with missing values under different rates. The proposed approach significantly outperformed the remaining methods in all settings, achieving an overall improvement between 26% and 67%.
Uncertainty in observations about the state of affairs is unavoidable, and generally undesirable, so we are motivated to try to minimize its effect on data analysis. Detection of anomalies in data has become an import...
详细信息
ISBN:
(纸本)9781728185262
Uncertainty in observations about the state of affairs is unavoidable, and generally undesirable, so we are motivated to try to minimize its effect on data analysis. Detection of anomalies in data has become an important research area. In this paper, we propose a novel approach to anomaly detection based on the variational autoencoder method with a Mish activation function and a Negative Log-Likelihood loss function. The proposed method is validated with ten standard datasets, comparing performance on each of the various activation functions and loss functions. Experimental results show that our proposed method offers an improvement over existing methods. Statistical properties (i.e., F1 score, AUC, and ROC) of the method are also examined in light of the experimental results.
Many approaches to training generative models by distinct training objectives have been proposed in the past. variational autoencoder (VAE) is an outstanding model of them based on log-likelihood. In this paper, we pr...
详细信息
ISBN:
(纸本)9783030638351;9783030638368
Many approaches to training generative models by distinct training objectives have been proposed in the past. variational autoencoder (VAE) is an outstanding model of them based on log-likelihood. In this paper, we propose a novel learnable prior, Pull-back Prior, for VAEs by adjusting the density of the prior through a discriminator that can assess the quality of data. It involves the discriminator from the theory of GANs to enrich the prior in VAEs. Based on it, we propose a more general framework, VAE with a Pull-back Prior (VAEPP), which uses existing techniques of VAEs and WGANs, to improve the log-likelihood, quality of sampling and stability of training. In MNIST and CIFAR-10, the log-likelihood of VAEPP outperforms models without autoregressive components and is comparable to autoregressive models. In MNIST, Fashion-MNIST, CIFAR-10 and CelebA, the FID of VAEPP is comparable to GANs and SOTA of VAEs.
Due to the complexity of emotional features, there has been limited success in emotional voice conversion. One major challenge is that conversion between more than two kinds of emotions often accompanies distortion of...
详细信息
ISBN:
(纸本)9781509066315
Due to the complexity of emotional features, there has been limited success in emotional voice conversion. One major challenge is that conversion between more than two kinds of emotions often accompanies distortion of voice signal. The factorized hierarchical variational autoencoder (FHVAE) [1] was previously shown to have an ability, called sequence-level regularization, to generate disentangled representations of both sequence-level (such as speaker identity) and segment-level features. This study exploits the FHVAE pipeline to produce disentangled representations of emotion, making it possible to greatly facilitate emotional voice conversion. We propose three versions of algorithms for improving the quality of disentangled representation and audio synthesis. We conducted three mean opinion score (MOS) surveys to assess the performance of our models in terms of 1) speaker's voice preservation, 2) emotion conversion, and 3) audio naturalness.
Fault detection is important for improving the reliability of spacecraft, ensuring the long-term stable operation, and reducing the economic loss caused by failure. In order to solve the problems such as the large amo...
详细信息
ISBN:
(纸本)9781728176871
Fault detection is important for improving the reliability of spacecraft, ensuring the long-term stable operation, and reducing the economic loss caused by failure. In order to solve the problems such as the large amount of test data, the scarcity of fault data samples and the real-time requirements in the field of spacecraft fault detection, an improved unsuperyised deep learning algorithm based on variational autoencoder (VAE) is proposed. The algorithm adopts Gated Recurrent Unit (GRU) based recurrent neural networks as encoder to automatically extract features of input data, and then uses VAE to learn the correlation features of multiple test data. The proposed network, trained only on the normal training dataset, is a typical unsupervised method which could learn features and reconstruct the data on the training set with a small loss. Once the reconstruction loss of the input data is larger than the pre-set threshold, the corresponding input data is considered as fault data. Experiments show that the proposed method is feasible and can effectively detect faults.
Laughter is one of the most famous non verbal sounds that human produce since birth, it conveys messages about our emotional state. These characteristics make it an important sound that should be studied in order to i...
详细信息
ISBN:
(数字)9781728175133
ISBN:
(纸本)9781728175133
Laughter is one of the most famous non verbal sounds that human produce since birth, it conveys messages about our emotional state. These characteristics make it an important sound that should be studied in order to improve the human-machine interactions. In this paper we investigate the audio laughter generation process from its acoustic features. This suggested process is considered as an analysis-transformation-synthesis benchmark based on unsupervised dimensionality reduction techniques: The standard autoencoder (AE) and the variational autoencoder (VAE). Therefore, the laughter synthesis methodology consists of transforming the extracted high-dimensional log magnitude spectrogram into a low-dimensional latent vector. This latent vector contains the most valuable information used to reconstruct a synthetic magnitude spectrogram that will be passed through a specific vocoder to generate the laughter waveform. We systematically, exploit the VAE to create new sound (speech-laugh) based on the interpolation process. To evaluate the performance of these models two evaluation metrics were conducted: objective and subjective evaluations.
暂无评论