Discriminator Of Generative Adversarial Network Research Articles

Consistency of Multiagent Distributed Generative Adversarial Networks.

The mathematical properties of generative adversarial networks (GANs) are presented via opinion dynamics, in which the discriminator is regarded as an agent while the generator is regarded as the principal (or another type of agent) in a GAN. In some cases, the overall convergence must be achieved through establishing a local information interaction between one agent and its neighbors via the multiagent system consistency theory and algorithms. So the goals of the multiagent consistency algorithm and the Nash equilibrium of GANs are virtually identical. Then, the existence of the Degroot model solution proves that the generator and discriminator in GANs can reach a consensus on the distribution function. Therefore, a new sufficient and necessary condition for the existence of a Nash equilibrium in GANs is obtained. Furthermore, a novel multiagent distributed GAN (MADGAN) is proposed to address the multiagent cognitive consistency problem in large-scale distributed network, based on the social group wisdom and the influence of the network structure on the agent. The nodes of a multiagent network are regarded as discriminators and generators, the discriminator with a relatively large degree of influence as a leader, and the generator as a follower, the lines with the degree denote the influence between the agents. The conditions of consensus are presented for a multigenerator and multidiscriminator (multiagent) distributed GAN by analyzing the existence of stationary distribution to the Markov chain of multiple agent states. Finally, theoretical results are verified with simulation based on the MNIST dataset.

Generating Optimized Guessing Candidates toward Better Password Cracking from Multi-Dictionaries Using Relativistic GAN

Despite their well-known weaknesses, passwords are still the de-facto authentication method for most online systems. Due to its importance, password cracking has been vibrantly researched both for offensive and defensive purposes. Hashcat and John the Ripper are the most popular cracking tools, allowing users to crack millions of passwords in a short time. However, their rule-based cracking has an explicit limitation of depending on password-cracking experts to come up with creative rules. To overcome this limitation, a recent trend has been to apply machine learning techniques to research on password cracking. For instance, state-of-the-art password guessing studies such as PassGAN and rPassGAN adopted a Generative Adversarial Network (GAN) and used it to generate high-quality password guesses without knowledge of password structures. However, compared with the probabilistic context-free grammar (PCFG), rPassGAN shows inferior password cracking performance in some cases. It was also observed that each password cracker has its own cracking space that does not overlap with other models. This observation led us to realize that an optimized candidate dictionary can be made by combining the password candidates generated by multiple password generation models. In this paper, we suggest a deep learning-based approach called REDPACK that addresses the weakness of the cutting-edge cracking tools based on GAN. To this end, REDPACK combines multiple password candidate generator models in an effective way. Our approach uses the discriminator of rPassGAN as the password selector. Then, by collecting passwords selectively, our model achieves a more realistic password candidate dictionary. Also, REDPACK improves password cracking performance by incorporating both the generator and the discriminator of GAN. We evaluated our system on various datasets with password candidates composed of symbols, digits, upper and lowercase letters. The results clearly show that our approach outperforms all existing approaches, including rule-based Hashcat, GAN-based PassGAN, and probability-based PCFG. The proposed model was also able to reduce the number of password candidates by up to 65%, with only 20% cracking performance loss compared to the union set of passwords cracked by multiple-generation models.

Image manipulation with natural language using Two-sided Attentive Conditional Generative Adversarial Network

Altering the content of an image with photo editing tools is a tedious task for an inexperienced user, especially, when modifying the visual attributes of a specific object in an image without affecting other constituents such as background etc. To simplify the process of image manipulation and to provide more control to users, it is better to utilize a simpler interface like natural language. It also enables to semantically modify parts of an image according to the given text. Therefore, in this paper, we address the challenge of manipulating images using natural language descriptions. We propose the Two-sidEd Attentive conditional Generative Adversarial Network (TEA-cGAN) to generate semantically manipulated images. TEA-cGAN’s contribution is seen as two-fold. The first contribution aims to attend locations that need to be modified during generation. It introduces two types of architectures that provide fine-grained attention both in the generator and discriminator of Generative Adversarial Network (GAN). To be specific, the first one i.e., the Single-scale architecture used in the generator focuses to modify only the text-relevant regions in an image and leaves other regions untouched. While the second one i.e., Multi-scale architecture further extended this idea by taking the different scales of image features into account. The second contribution purpose is to generate higher resolution images (e.g., 256 × 256) as they provide better quality and stability. Quantitative and qualitative experiments conducted on CUB and Oxford-102 datasets confirm that TEA-cGAN different scale architectures outperform existing methods while generating 128 × 128 resolution images including generating higher resolution image i.e., 256 × 256.

Effective Visual Domain Adaptation via Generative Adversarial Distribution Matching.

In the field of computer vision, without sufficient labeled images, it is challenging to train an accurate model. However, through visual adaptation from source to target domains, a relevant labeled dataset can help solve such problem. Many methods apply adversarial learning to diminish cross-domain distribution difference. They are able to greatly enhance the performance on target classification tasks. Generative adversarial network (GAN) loss is widely used in adversarial adaptation learning methods to reduce an across-domain distribution difference. However, it becomes difficult to decline such distribution difference if generator or discriminator in GAN fails to work as expected and degrades its performance. To solve such cross-domain classification problems, we put forward a novel adaptation framework called generative adversarial distribution matching (GADM). In GADM, we improve the objective function by taking cross-domain discrepancy distance into consideration and further minimize the difference through the competition between a generator and discriminator, thereby greatly decreasing cross-domain distribution difference. Experimental results and comparison with several state-of-the-art methods verify GADM's superiority in image classification across domains.

Adaptive DropBlock-Enhanced Generative Adversarial Networks for Hyperspectral Image Classification

In recent years, hyperspectral image (HSI) classification based on generative adversarial networks (GAN) has achieved great progress. GAN-based classification methods can mitigate the limited training sample dilemma to some extent. However, several studies have pointed out that existing GAN-based HSI classification methods are heavily affected by the imbalanced training data problem. The discriminator in GAN always contradicts itself and tries to associate fake labels to the minority-class samples, and thus impair the classification performance. Another critical issue is the mode collapse in GAN-based methods. The generator is only capable of producing samples within a narrow scope of the data space, which severely hinders the advancement of GAN-based HSI classification methods. In this paper, we proposed an Adaptive DropBlock-enhanced Generative Adversarial Networks (ADGAN) for HSI classification. First, to solve the imbalanced training data problem, we adjust the discriminator to be a single classifier, and it will not contradict itself. Second, an adaptive DropBlock (AdapDrop) is proposed as a regularization method employed in the generator and discriminator to alleviate the mode collapse issue. The AdapDrop generated drop masks with adaptive shapes instead of a fixed size region, and it alleviates the limitations of DropBlock in dealing with ground objects with various shapes. Experimental results on three HSI datasets demonstrated that the proposed ADGAN achieved superior performance over state-of-the-art GAN-based methods. Our codes are available at https://github.com/summitgao/HC_ADGAN

Self-Supervised Representation Learning for Remote Sensing Image Change Detection Based on Temporal Prediction

Traditional change detection (CD) methods operate in the simple image domain or hand-crafted features, which has less robustness to the inconsistencies (e.g., brightness and noise distribution, etc.) between bitemporal satellite images. Recently, deep learning techniques have reported compelling performance on robust feature learning. However, generating accurate semantic supervision that reveals real change information in satellite images still remains challenging, especially for manual annotation. To solve this problem, we propose a novel self-supervised representation learning method based on temporal prediction for remote sensing image CD. The main idea of our algorithm is to transform two satellite images into more consistent feature representations through a self-supervised mechanism without semantic supervision and any additional computations. Based on the transformed feature representations, a better difference image (DI) can be obtained, which reduces the propagated error of DI on the final detection result. In the self-supervised mechanism, the network is asked to identify different sample patches between two temporal images, namely, temporal prediction. By designing the network for the temporal prediction task to imitate the discriminator of generative adversarial networks, the distribution-aware feature representations are automatically captured and the result with powerful robustness can be acquired. Experimental results on real remote sensing data sets show the effectiveness and superiority of our method, improving the detection precision up to 0.94–35.49%.

AttentionFGAN: Infrared and Visible Image Fusion Using Attention-Based Generative Adversarial Networks

Infrared and visible image fusion aims to describe the same scene from different aspects by combining complementary information of multi-modality images. The existing Generative adversarial networks (GAN) based infrared and visible image fusion methods cannot perceive the most discriminative regions, and hence fail to highlight the typical parts existing in infrared and visible images. To this end, we integrate multi-scale attention mechanism into both generator and discriminator of GAN to fuse infrared and visible images (AttentionFGAN). The multi-scale attention mechanism aims to not only capture comprehensive spatial information to help generator focus on the foreground target information of infrared image and background detail information of visible image, but also constrain the discriminators focus more on the attention regions rather than the whole input image. The generator of AttentionFGAN consists of two multi-scale attention networks and an image fusion network. Two multi-scale attention networks capture the attention maps of infrared and visible images respectively, so that the fusion network can reconstruct the fused image by paying more attention to the typical regions of source images. Besides, two discriminators are adopted to force the fused result keep more intensity and texture information from infrared and visible image respectively. Moreover, to keep more information of attention region from source images, an attention loss function is designed. Finally, the ablation experiments illustrate the effectiveness of the key parts of our method, and extensive qualitative and quantitative experiments on three public datasets demonstrate the advantages and effectiveness of AttentionFGAN compared with the other state-of-the-art methods.

Generative attention adversarial classification network for unsupervised domain adaptation

Domain adaptation is a significant and popular issue of solving distribution discrepancy among different domains in computer vision. Generally, previous works proposed are mainly devoted to reducing domain shift between source domain with labeled data and target domain without labels. Adversarial learning in deep networks has already been widely applied to learn disentangled and transferable features between two different domains to minimize domains distribution discrepancy. However, these methods rarely consider class distributions among source data during adversarial learning, and they pay little attention to these transferable regions among source and target domains images. In this paper, we propose a Generative Attention Adversarial Classification Network (GAACN) model for unsupervised domain adaptation. To learn a joint feature distribution between source and target domains, we present an improved generative adversarial network (GAN) following the feature extractor. Firstly, the discriminator of GAN discriminates the distribution of domains and the classes distribution among source data during adversarial learning, so that our feature extractor can learn a joint feature distribution between source and target domains and maintain the classes consistent simultaneously. Secondly, we present an attention module embedded in GAN, which allows the discriminator to discriminate the transferable regions among the images of source and target domains. Lastly, we propose a simple and efficient method which allocates pseudo-labels for unlabeled target data, and it can improve the performance of our model GAACN while mitigating negative transfer. Extensive experiments demonstrate that our proposed model achieves perfect results on several standard domain adaptation datasets.

Deep Learning Designs of Anisotropic Metasurfaces in Ultrawideband Based on Generative Adversarial Networks

Metasurfaces have been developed as a promising approach for manipulating electromagnetic waves. Recently, deep learning algorithms have been introduced to design metasurfaces, but the network can only output one solution for each desired input and suffers from nonunique issue. To overcome the aforementioned challenges, a deep neural network model for inverse designs of anisotropic metasurfaces with full phase properties in ultrawideband is proposed. Given the target reflection spectra as inputs, the candidate metasurface patterns are generated through a generative adversarial network (GAN), and the corresponding predictions are simply achieved by the accurate forward neural network model to match the target spectra in the whole band with high fidelity. By training the generator and discriminator in GAN in an alternating order combined with setting a threshold of discriminator loss to trigger the phase prediction, the proposed method is much more efficient and consumes less time in the training process. Numerical simulations and experimental results demonstrate that the reflection phases of the generated meta‐atoms have excellent agreements with the given targets, providing an efficient way in automatically designing metasurfaces. The most important advantage of this approach over the previous schemes is to improve the design speed significantly with very good accuracy.

Improving Adversarial Neural Machine Translation for Morphologically Rich Language

Generative adversarial networks (GAN) have great successes on natural language processing (NLP) and neural machine translation (NMT). However, the existing discriminator in GAN for NMT only combines two words as one query to train the translation models, which restrict the discriminator to be more meaningful and fail to apply rich monolingual information. Recent studies only consider one single reference translation during model training, this limit the GAN model to learn sufficient information about the representation of source sentence. These situations are even worse when languages are morphologically rich. In this article, an extended version of GAN model for neural machine translation is proposed to optimize the performance of morphologically rich language translation. In particular, we use the morphological word embedding instead of word embedding as input in GAN model to enrich the representation of words and overcome the data sparsity problem during model training. Moreover, multiple references are integrated into discriminator to make the model consider more context information and adapt to the diversity of different languages. Experimental results on German $\leftrightarrow$ English, French $\leftrightarrow$ English, Czech $\leftrightarrow$ English, Finnish $\leftrightarrow$ English, Turkish $\leftrightarrow$ English, Chinese $\leftrightarrow$ English, Finnish $\leftrightarrow$ Turkish and Turkish $\leftrightarrow$ Czech translation tasks demonstrate that our method achieves significant improvements over baseline systems.

Generative Adversarial Regularized Mutual Information Policy Gradient Framework for Automatic Diagnosis

Automatic diagnosis systems have attracted increasing attention in recent years. The reinforcement learning (RL) is an attractive technique for building an automatic diagnosis system due to its advantages for handling sequential decision making problem. However, the RL method still cannot achieve good enough prediction accuracy. In this paper, we propose a Generative Adversarial regularized Mutual information Policy gradient framework (GAMP) for automatic diagnosis which aims to make a diagnosis rapidly and accurately. We first propose a new policy gradient framework based on the Generative Adversarial Network (GAN) to optimize the RL model for automatic diagnosis. In our framework, we take the generator of GAN as a policy network, and also use the discriminator of GAN as a part of the reward function. This generative adversarial regularized policy gradient framework can try to avoid generating randomized trials of symptom inquires deviated from the common diagnosis paradigm. In addition, we add mutual information to enhance the reward function to encourage the model to select the most discriminative symptoms to make a diagnosis. Experiment evaluations on two public datasets show that our method beats the state-of-art methods, not only can achieve higher diagnosis accuracy, but also can use a smaller number of inquires to make diagnosis decision.

Discriminator Feature-Based Inference by Recycling the Discriminator of GANs

Generative adversarial networks (GANs)successfully generate high quality data by learning amapping from a latent vector to the data. Various studies assert that the latent space of a GAN is semanticallymeaningful and can be utilized for advanced data analysis and manipulation. To analyze the real data in thelatent space of a GAN, it is necessary to build an inference mapping from the data to the latent vector. Thispaper proposes an effective algorithm to accurately infer the latent vector by utilizing GAN discriminator features. Our primary goal is to increase inference mappingaccuracy with minimal training overhead. Furthermore,using the proposed algorithm, we suggest a conditionalimage generation algorithm, namely a spatially conditioned GAN. Extensive evaluations confirmed that theproposed inference algorithm achieved more semantically accurate inference mapping than existing methodsand can be successfully applied to advanced conditionalimage generation tasks.

Retinal Vessel Segmentation Combined With Generative Adversarial Networks and Dense U-Net

Retinal blood vessels are the basis for clinical diagnosis of some diseases. Achieving automatic retinal vessel segmentation from fundus images is an important and challenging work. In this paper, a neural network architecture based on the Dense U-net using Inception module is proposed for retinal vessel segmentation. First, the skip connections in traditional U-net are replaced with Dense Block to achieve full fusion of features from shallow layers to deep layers. Then, the Inception module is applied to supersede the traditional convolution operation. Thus, vessel features corresponding to the convolution kernels of different sizes can be extracted. Finally, Generative Adversarial Networks (GAN) are adopted in the training phase. The Dense U-net using Inception module is treated as the generator of GAN, and a multilayer neural network is created as the discriminator of GAN. The generator and discriminator are trained alternately. The loss function is a combination of segmentation loss and GAN loss. So that segmentation results can be fitted to the ground truth from both pixel value and pixel distribution. The algorithm proposed in this paper is verified on the public Digital Retinal Images for Vessel Extraction (DRIVE) dataset, where the Dice rate reaches 82.15%, and the AU-ROC and AU-PR reach 0.9772 and 0.9058, respectively. Experiments show that the proposed algorithm is effective in realizing automatic retinal vessel segmentation.

VIRTUAL TRAINING SAMPLE GENERATION BY GENERATIVE ADVERSARIAL NETWORKS FOR HYPERSPECTRAL IMAGES CLASSIFICATION

Abstract. Convolutional Neural Networks (CNNs) as a well-known deep learning technique has shown a remarkable performance in visual recognition applications. However, using such networks in the area of hyperspectral image classification is a challenging and time-consuming process due to the high dimensionality and the insufficient training samples. In addition, Generative Adversarial Networks (GANs) has attracted a lot of attentions in order to generate virtual training samples. In this paper, we present a new classification framework based on integration of multi-channel CNNs and new architecture for generator and discriminator of GANs to overcome Small Sample Size (SSS) problem in hyperspectral image classification. Further, in order to reduce the computational cost, the methods related to the reduction of subspace dimension were proposed to obtain the dominant feature around the training sample to generate meaningful training samples from the original one. The proposed framework overcomes SSS and overfitting problem in classifying hyperspectral images. Based on the experimental results on real and well-known hyperspectral benchmark images, our proposed strategy improves the performance compared to standard CNNs and conventional data augmentation strategy. The overall classification accuracy in Pavia University and Indian Pines datasets was 99.8% and 94.9%, respectively.

E-WACGAN: Enhanced Generative Model of Signaling Data Based on WGAN-GP and ACGAN

In recent years, the generative adversarial network (GAN) has achieved outstanding performance in the image field and the derivatives of GAN, namely auxiliary classifier GAN (ACGAN) and Wasserstein GAN with gradient penalty (WGAN-GP) have also been widely used, but the GAN applications of nonimage domain are not wide. At the time when the telecommunication fraud is rampant, the signaling data of telephone contain a lot of useful information, which is helpful for distinguishing fraudulent and nonfraudulent calls. In this article, aiming at the problem of limited amount of data and information leakage in the research of telephone signaling data, we adopt WGAN-GP and ACGAN to generate analog data, which confirms distribution of true data. In order to solve the problem of category accuracy of analog data and to enhance the stability and speed of training, we proposed a new network structure for discriminator of GAN based on WGAN-GP and ACGAN. The experiments on telecommunication fraud dataset found that our method obtains better performance on signaling data than base model.

Deep learning fault diagnosis method based on global optimization GAN for unbalanced data

Deep learning can be applied to the field of fault diagnosis for its powerful feature representation capabilities. When a certain class fault samples available are very limited, it is inevitably to be unbalanced. The fault feature extracted from unbalanced data via deep learning is inaccurate, which can lead to high misclassification rate. To solve this problem, new generator and discriminator of Generative Adversarial Network (GAN) are designed in this paper to generate more discriminant fault samples using a scheme of global optimization. The generator is designed to generate those fault feature extracted from a few fault samples via Auto Encoder (AE) instead of fault data sample. The training of the generator is guided by fault feature and fault diagnosis error instead of the statistical coincidence of traditional GAN. The discriminator is designed to filter the unqualified generated samples in the sense that qualified samples are helpful for more accurate fault diagnosis. The experimental results of rolling bearings verify the effectiveness of the proposed algorithm.

Electrocardiogram generation with a bidirectional LSTM-CNN generative adversarial network

Heart disease is a malignant threat to human health. Electrocardiogram (ECG) tests are used to help diagnose heart disease by recording the heart’s activity. However, automated medical-aided diagnosis with computers usually requires a large volume of labeled clinical data without patients' privacy to train the model, which is an empirical problem that still needs to be solved. To address this problem, we propose a generative adversarial network (GAN), which is composed of a bidirectional long short-term memory(LSTM) and convolutional neural network(CNN), referred as BiLSTM-CNN,to generate synthetic ECG data that agree with existing clinical data so that the features of patients with heart disease can be retained. The model includes a generator and a discriminator, where the generator employs the two layers of the BiLSTM networks and the discriminator is based on convolutional neural networks. The 48 ECG records from individuals of the MIT-BIH database were used to train the model. We compared the performance of our model with two other generative models, the recurrent neural network autoencoder(RNN-AE) and the recurrent neural network variational autoencoder (RNN-VAE). The results showed that the loss function of our model converged to zero the fastest. We also evaluated the loss of the discriminator of GANs with different combinations of generator and discriminator. The results indicated that BiLSTM-CNN GAN could generate ECG data with high morphological similarity to real ECG recordings.

Semi-Supervised Deep Transfer Learning-Based on Adversarial Feature Learning for Label Limited SAR Target Recognition

The data-driven convolutional neural networks (CNNs) have achieved great progress in Synthetic Aperture Radar automatic target recognition (SAR-ATR) after being trained in a large scale of labeled samples. However, the insufficiency of labeled SAR data always leads to over-fitting, causing significant performance degradation. To solve the mentioned problem, a semi-supervised transfer learning method based on generative adversarial networks (GANs) is presented in the present paper. The discriminator of GAN with an encoder and a discriminative layer is redesigned to make it capable of learning the feature representation of input data with unsupervised settings. Instead of training a deep neural network with the insufficient labeled data set, we first train a GAN with varieties of unlabeled samples to learn generic features of SAR images. Subsequently, the learned parameters are readopted to initialize the target network to transfer the generic knowledge to specific SAR target recognition task. Lastly, the target network is fine-tuned using both the labeled and unlabeled training samples by a semi-supervised loss function. We evaluate the proposed method on the MSTAR and OpenSARShip data set with 80%, 60%, 40%, and 20% of the training set labeled, respectively. The results suggest that the proposed method achieves up to 23.58% accuracy enhancement over the random-initialized model.

Deep Segmentation Domain Adaptation Network With Weighted Boundary Constraint

Semantic segmentation domain adaptation is used to deal with segmentation problems in a new domain even without pixel-level labels. Highly precise boundaries are the major indicator of segmentation performance, but the previous methods mainly have focused on global representation rather than on local representation, leading to an inferior performance of object boundaries in domain adaptation. In this paper, we propose weighted boundary constraint to refine those segmentation predictions and incorporate it into a generative adversarial network (GAN)-based network for domain adaption to achieve further significant improvement. The boundary constraint loss is designed as the cross-entropy between the intermediate result and the refined result. In addition, confidence from discriminator of GAN is used to constrain the boundary constraint loss to reduce the negative impact from inaccurate object boundaries. The entire network can be learned in an end-to-end manner. Both quantitative and qualitative experiments demonstrate the benefits of our approach which shows the competitive performance with the state-of-the-art methods.

Cellular structure image classification with small targeted training samples.

Cell shapes provide crucial biological information on complex tissues. Different cell types often have distinct cell shapes, and collective shape changes usually indicate morphogenetic events and mechanisms. The identification and detection of collective cell shape changes in an extensive collection of 3D time-lapse images of complex tissues is an important step in assaying such mechanisms but is a tedious and time-consuming task. Machine learning provides new opportunities to automatically detect cell shape changes. However, it is challenging to generate sufficient training samples for pattern identification through deep learning because of a limited amount of images and annotations. We present a deep learning approach with minimal well-annotated training samples and apply it to identify multicellular rosettes from 3D live images of the Caenorhabditis elegans embryo with fluorescently labeled cell membranes. Our strategy is to combine two approaches, namely, feature transfer and generative adversarial networks (GANs), to boost image classification with small training samples. Specifically, we use a GAN framework and conduct an unsupervised training to capture the general characteristics of cell membrane images with 11,250 unlabelled images. We then transfer the structure of the GAN discriminator into a new Alex-style neural network for further learning with several dozen labeled samples. Our experiments showed that with 10-15 well-labeled rosette images and 30-40 randomly selected nonrosette images our approach can identify rosettes with more than 80% accuracy and capture more than 90% of the model accuracy achieved with a training data et that is five times larger. We also established a public benchmark dataset for rosette detection. This GAN-based transfer approach can be applied to the study of other cellular structures with minimal training samples.