摘要 :
Abstract Image generation is a hot topic in the field of machine learning and computer vision. As a representative of its algorithm, the Generative Adversarial Network (GAN) has the problem of mode collapse in practice. The propos...
展开
Abstract Image generation is a hot topic in the field of machine learning and computer vision. As a representative of its algorithm, the Generative Adversarial Network (GAN) has the problem of mode collapse in practice. The proposed Dual Discriminator Weighted Mixture Generative Adversarial Network (D2WMGAN) approach can cope with this problem. On the one hand, the D2WMGAN uses the mixed distribution of multiple generators to approximate the real distribution, in order to prevent the extreme situation that multiple generators learn the same distribution and generate the same class of samples, with a classifier to play games with generators to make different generators learn different distributions. On the other hand, the objective function of D2WMGAN weights the Kullback–Leibler (KL) divergence and the reverse KL divergence, and uses their complementary characteristics to improve the quality and diversity of samples from the generators. Then, the theoretical conditional optimality of the D2WMGAN is proved theoretically, which shows that multiple generators can learn the real data distribution in the case of the optimal discriminator and classifier. Finally, extensive experiments are conducted on a large amount of synthetic data and real-world large-scale datasets (such as, CIFAR-10 and MNIST), and the commonly used GAN evaluation indicators (Wasserstein distance, JS divergence, Inception score, and Frechet Inception Distance) are introduced for comparative analysis. Experimental results show that the proposed D2WMGAN approach can better learn multiple mode data, generate rich realistic samples, and effectively solve the problem of mode collapse.
收起
摘要 :
Automatic line art colorization plays an important role in anime and comic industry. While existing methods for line art colorization are able to generate plausible colorized results, they tend to suffer from the color bleeding is...
展开
Automatic line art colorization plays an important role in anime and comic industry. While existing methods for line art colorization are able to generate plausible colorized results, they tend to suffer from the color bleeding issue. We introduce an explicit segmentation fusion mechanism to aid colorization frameworks in avoiding color bleeding artifacts. This mechanism is able to provide region segmentation information for the colorization process explicitly so that the colorization model can learn to avoid assigning the same color across regions with different semantics or inconsistent colors inside an individual region. The proposed mechanism is designed in a plug-and-play manner, so it can be applied to a diversity of line art colorization frameworks with various kinds of user guidances. We evaluate this mechanism in tag-based and reference-based line art colorization tasks by incorporating it into the state-of-the-art models. Comparisons with these existing models corroborate the effectiveness of our method which largely alleviates the color bleeding artifacts.
收起
摘要 :
Conditional Generative Adversarial Network (cGAN) is a general purpose approach for many image-to-image translation tasks, which aims to translate images from one form to another resulting in high-quality translated images. In thi...
展开
Conditional Generative Adversarial Network (cGAN) is a general purpose approach for many image-to-image translation tasks, which aims to translate images from one form to another resulting in high-quality translated images. In this paper, the loss function of the cGAN model is modified by combining the adversarial loss of state-of-the-art Generative Adversarial Network (GAN) models with a new combination of non-adversarial loss functions to enhance model performance and generate more realistic images. Specifically, the effect of the Wasserstein GAN (WGAN), the WGAN with Gradient Penalty (WGAN-GP), and least Squared GAN (lsGAN) adversarial loss functions are explored. Several comparisons are performed to select an optimized combination of L1<math><mrow is="true"><msub is="true"><mrow is="true"><mi is="true">L</mi></mrow><mrow is="true"><mn is="true">1</mn></mrow></msub></mrow></math> with structure, gradient, content-based, Kullback-Leibler divergence, and softmax non-adversarial loss functions. For experimentation purposes, the Facades dataset is used in case of image-to-image translation task. Peak-signal-to-noise-ratio (PSNR), Structural Similarity Index (SSIM), Universal Quality Index (UQI), and Visual Information Fidelity (VIF) are used to quantitatively evaluate the translated images. Based on our experimental results, the best combination of the loss functions for image-to-image translation on facade dataset is (WGAN) adversarial loss with (L1<math><mrow is="true"><msub is="true"><mrow is="true"><mi is="true">L</mi></mrow><mrow is="true"><mn is="true">1</mn></mrow></msub></mrow></math> and content) non-adversarial loss functions. The model generates fine structure images, and captures both high and low frequency details of translated images. Image in-painting and lesion segmentation is investigated to demonstrate practicality of proposed work.
收起
摘要 :
Generative Adversarial Networks (GANs) are one of the most popular and powerful models to learn the complex high dimensional distributions. However, they usually suffer from instability and generalization issues which may lead to ...
展开
Generative Adversarial Networks (GANs) are one of the most popular and powerful models to learn the complex high dimensional distributions. However, they usually suffer from instability and generalization issues which may lead to poor generations. Most existing works focus on stabilizing the training for the discriminators of GANs while ignoring their generalization issue. In this work, we aim to improve the generalization capability of GANs by promoting the local robustness within the small neighborhood of the training samples. We prove that the robustness in the small neighborhood of the training sets can lead to better generalization. Particularly, we design a new robust method called Robust Generative Adversarial Network (RGAN) in which the generator and discriminator compete with each other in a worst-case setting within a small Wasserstein ball. The generator tries to map the worst input distribution (rather than a Gaussian distribution used in most GANs) to the real data distribution, while the discriminator attempts to distinguish the real and fake distributions with the worst perturbations. Intuitively, the proposed RGAN can learn a good generator and discriminator that can even perform well on the worst-case input points. Strictly, we have proved that RGAN can obtain a tighter generalization upper bound than the traditional GANs under mild assumptions, ensuring a theoretical superiority of RGAN over GANs. We conduct our proposed method on five different baselines (five popular GAN models). And a series of experiments on CIFAR-10, STL-10 and CelebA datasets indicate that our proposed robust frameworks outperform five baseline models substantially and consistently.
收起
摘要 :
Sequence generative adversarial networks (SeqGAN) have been used to improve conditional sequence generation tasks, for example, chit-chat dialogue generation. To stabilize the training of SeqGAN, Monte Carlo tree search (MCTS) or ...
展开
Sequence generative adversarial networks (SeqGAN) have been used to improve conditional sequence generation tasks, for example, chit-chat dialogue generation. To stabilize the training of SeqGAN, Monte Carlo tree search (MCTS) or reward at every generation step (REGS) is used to evaluate the goodness of a generated subsequence. MCTS is computationally intensive, but the performance of REGS is worse than MCTS. In this paper, we propose stepwise GAN (StepGAN), in which the discriminator is modified to automatically assign scores quantifying the goodness of each subsequence at every generation step. StepGAN has significantly less computational costs than MCTS. We demonstrate that StepGAN outperforms previous GAN-based methods on both synthetic experiment and chit-chat dialogue generation.
收起
摘要 :
Generative Adversarial Networks (GAN) is a framework for estimating generative models by training a generative model G which captures the data distribution and a discriminative model D which is trained to discriminate a real examp...
展开
Generative Adversarial Networks (GAN) is a framework for estimating generative models by training a generative model G which captures the data distribution and a discriminative model D which is trained to discriminate a real example sampled from the training dataset from a fake sample generated by G simultaneously. G and D are usually defined by neural networks, so that GAN is trained with backpropagation. In this article, how to train GAN and recent methods for image generation using GAN are briefly introduced.
收起
摘要 :
Generative adversarial network (GAN) is one of the most promising methods for unsupervised learning in recent years. GAN works via adversarial training concept and has shown excellent performance in the fields image synthesis, ima...
展开
Generative adversarial network (GAN) is one of the most promising methods for unsupervised learning in recent years. GAN works via adversarial training concept and has shown excellent performance in the fields image synthesis, image super-resolution, video generation, image translation, etc. Compared with classical algorithms, quantum algorithms have their unique advantages in dealing with complex tasks, quantum machine learning (QML) is one of the most promising quantum algorithms with the rapid development of quantum technology. Specifically, Quantum generative adversarial network (QGAN) has shown the potential exponential quantum speedups in terms of performance. Meanwhile, QGAN also exhibits some problems, such as barren plateaus, unstable gradient, model collapse, absent complete scientific evaluation system, etc. How to improve the theory of QGAN and apply it that have attracted some researcher. In this paper, we comprehensively and deeply review recently proposed GAN and QAGN models and their applications, and we discuss the existing problems and future research trends of QGAN.
收起
摘要 :
Generative adversarial network (GAN) is one of the most promising methods for unsupervised learning in recent years. GAN works via adversarial training concept and has shown excellent performance in the fields image synthesis, ima...
展开
Generative adversarial network (GAN) is one of the most promising methods for unsupervised learning in recent years. GAN works via adversarial training concept and has shown excellent performance in the fields image synthesis, image super-resolution, video generation, image translation, etc. Compared with classical algorithms, quantum algorithms have their unique advantages in dealing with complex tasks, quantum machine learning (QML) is one of the most promising quantum algorithms with the rapid development of quantum technology. Specifically, Quantum generative adversarial network (QGAN) has shown the potential exponential quantum speedups in terms of performance. Meanwhile, QGAN also exhibits some problems, such as barren plateaus, unstable gradient, model collapse, absent complete scientific evaluation system, etc. How to improve the theory of QGAN and apply it that have attracted some researcher. In this paper, we comprehensively and deeply review recently proposed GAN and QAGN models and their applications, and we discuss the existing problems and future research trends of QGAN.
收起
摘要 :
Machine learning algorithms represent the intelligence that controls many information systems and applications around us. As such, they are targeted by attackers to impact their decisions. Text created by machine learning algorith...
展开
Machine learning algorithms represent the intelligence that controls many information systems and applications around us. As such, they are targeted by attackers to impact their decisions. Text created by machine learning algorithms has many types of applications, some of which can be considered malicious especially if there is an intention to present machine-generated text as human-generated. In this paper, we surveyed major subjects in adversarial machine learning for text processing applications. Unlike adversarial machine learning in images, text problems and applications are heterogeneous. Thus, each problem can have its own challenges. We focused on some of the evolving research areas such as: malicious versus genuine text generation metrics, defense against adversarial attacks, and text generation models and algorithms. Our study showed that as applications of text generation will continue to grow in the near future, the type and nature of attacks on those applications and their machine learning algorithms will continue to grow as well. Literature survey indicated an increasing trend in using pre-trained models in machine learning. Word/sentence embedding models and transformers are examples of those pre-trained models. Adversarial models may utilize same or similar pre-trained models as well. In another trend related to text generation models, literature showed effort to develop universal text perturbations to be used in both black-and white-box attack settings. Literature showed also using conditional GANs to create latent representation for writing types. This usage will allow for a seamless lexical and grammatical transition between various writing styles. In text generation metrics, research trends showed developing successful automated or semi-automated assessment metrics that may include human judgement. Literature showed also research trends of designing and developing new memory models that increase performance and memory utilization efficiency without validating real-time constraints. Many research efforts evaluate different defense model approaches and algorithms. Researchers evaluated different types of targeted attacks, and methods to distinguish human versus machine generated text.
收起
摘要 :
Missing data is a common problem faced with real-world datasets. Imputation is a widely used technique to estimate the missing data. State-of-the-art imputation approaches model the distribution of observed data to approximate the...
展开
Missing data is a common problem faced with real-world datasets. Imputation is a widely used technique to estimate the missing data. State-of-the-art imputation approaches model the distribution of observed data to approximate the missing values. Such an approach usually models a single distribution for the entire dataset, which overlooks the class-specific characteristics of the data. Class-specific characteristics are especially useful when there is a class imbalance. We propose a new method for imputing missing data based on its class-specific characteristics by adapting the popular Conditional Generative Adversarial Networks (CGAN). Our Conditional Generative Adversarial Imputation Network (CGAIN) imputes the missing data using class-specific distributions, which can produce the best estimates for the missing values. We tested our approach on baseline datasets and achieved superior performance compared with the state-of-the-art and popular imputation approaches.(c) 2021 Elsevier B.V. All rights reserved.
收起