Conditional Wasserstein Generative Adversarial Network Research Articles

Research on fault diagnosis of industrial robots based on generative adversarial network

Abstract-In today's industrial production, companies have to introduce various types of industrial robots in order to reduce labor costs, shorten product processing time, and obtain greater profits. This move has caused the installed capacity of industrial robots to increase year by year. However, this also makes industrial robots have more and more frequent failures and an increasing amount of failures in actual production. The occurrence of industrial robot failures is unavoidable, and the amount of data in the healthy state and the amount of data in the fault state are extremely unbalanced. Whether the fault diagnosis model is efficient and accurate is the key to providing reliable theoretical and technical support for the health management of industrial robots. Therefore, this paper constructs an imbalanced data fault diagnosis model based on Wasserstein Generative Adversarial Network (WGAN) for fault diagnosis of industrial robots in data balanced and unbalanced states. We evaluate the quality of generated data through the following three aspects: the change in loss values ​​of the discriminator and generator, the degree of overlap between the generated data and real data samples, and the classification accuracy of the classification algorithm. The results show that the data generated by the WGAN model is of high quality and can be used as extended data for the fault data set. Next, we used the deep residual network designed in this article to perform fault diagnosis on the balanced data set expanded with the generated data. The experimental results confirmed the high quality of the generated data and the outstanding expressive ability of the deep residual network. In order to measure the performance of the model in this paper, we used 7 types of generative adversarial networks to generate fault data. The quality of the generated data samples is still analyzed and evaluated through the three aspects mentioned above. The results show that only the conditional Wasserstein generative adversarial network model generates data of high quality, which can be used as extended data for fault data and has good results in actual applications.

Application of generative AI-based data augmentation technique in transformer winding deformation fault diagnosis

Accurately diagnosing various winding deformation faults is significant in power transformer maintenance. Among the different fault diagnosis methods, frequency response analysis (FRA) is widely used but still poses challenges. Artificial intelligence (AI)-based methods have recently been proposed to interpret FRA data. Nevertheless, these approaches are either complicated or exhibit limited generalization performance due to real-world FRA fault data scarcity. Inspired by AI-generated content (AIGC), this study proposes a data augmentation technique named conditional Wasserstein generative adversarial network with gradient penalty (Conditional-WGAN-GP) combined with fault diagnosis model. Numerous FRA-based data are automatically generated using the proposed data augmentation technique based on real FRA data obtained from a specially designed 10 kV transformer. The augmented dataset is then used to train fault diagnosis models to detect winding deformation faults. The trained fault diagnosis model is subsequently applied to assess two actual transformers. Experimental results demonstrate that when combined with the proposed method, even simpler fault diagnosis models can achieve high accuracy, exhibiting an improvement of approximately 5 % compared to the previous baseline model. The fault diagnosis models combined with the proposed data augmentation technique demonstrate improved generalization and robustness. (GitHub code: https://github.com/cy1034429432/Diagnosing-Transformer-Winding-Deformation-Fault-Types-from-FRA-Based-on-Conditional-WGAN-GP-/tree/main).

The identification of gas–liquid two-phase flow patterns in a centrifugal pump with an imbalanced sample and variable rotational speeds

Due to the effects of the working medium and external conditions, centrifugal pumps often produce complex gas–liquid two-phase flows during operation. The entrainment of gas can lead to a significant decrease in the pressure head of the centrifugal pump and a serious reduction in conveyance efficiency. Therefore, in the absence of visualization or noninvasive techniques, the question of how to predict and model the flow status inside the pump is a key issue that needs to be urgently addressed. Using experimental data for gas–liquid two-phase flows in centrifugal pumps, this paper proposes a flow pattern identification model based on one-dimensional pressure increment signals. This deep learning model consists of convolutional neural networks and bidirectional gated recurrent units. Bayesian optimization is used for automatic hyperparameter tuning, and under the premise of balanced data distribution, the model achieves an identification rate of 0.98 on the test set. Subsequently, addressing the issue of the uneven sample distribution in centrifugal pumps due to their poor gas-carrying capacity, a data augmentation model based on the conditional Wasserstein generative adversarial network with gradient penalty approach is proposed based on the predictive model. Finally, to establish a flow pattern identification method with good extrapolation capability, a transfer learning strategy that accounts for the domain adaptation problem is introduced to the dataset with variable speeds based on the previous two studies. This method achieves knowledge transfer from the source domain to the target domain with a recognition rate of over 0.988 without hyperparameter optimization.

Data augmentation for Human Activity Recognition with Generative Adversarial Networks.

Currently, Human Activity Recognition (HAR) applications need a large volume of data to be able to generalize to new users and environments. However, the availability of labeled data is usually limited and the process of recording new data is costly and time-consuming. Synthetically increasing datasets using Generative Adversarial Networks (GANs) has been proposed, outperforming cropping, time-warping, and jittering techniques on raw signals. Incorporating GAN-generated synthetic data into datasets has been demonstrated to improve the accuracy of trained models. Regardless, currently, there is no optimal GAN architecture to generate accelerometry signals, neither a proper evaluation methodology to assess signal quality or accuracy using synthetic data. This work is the first to propose conditional Wasserstein Generative Adversarial Networks (cWGANs) to generate synthetic HAR accelerometry signals. Furthermore, we calculate quality metrics from the literature and study the impact of synthetic data on a large HAR dataset involving 395 users. Results show that i) cWGAN outperforms original Conditional Generative Adversarial Networks (cGANs), being 1D convolutional layers appropriate for generating accelerometry signals, ii) the performance improvement incorporating synthetic data is more significant as the dataset size is smaller, and iii) the quantity of synthetic data required is inversely proportional to the quantity of real data.

The efficacy and generalizability of conditional GANs for posterior inference in physics-based inverse problems

In this work, we train conditional Wasserstein generative adversarial networks to effectively sample from the posterior of physics-based Bayesian inference problems. The generator is constructed using a U-Net architecture, with the latent information injected using conditional instance normalization. The former facilitates a multiscale inverse map, while the latter enables the decoupling of the latent space dimension from the dimension of the measurement, and introduces stochasticity at all scales of the U-Net. We solve PDE-based inverse problems to demonstrate the performance of our approach in quantifying the uncertainty in the inferred field. Further, we show the generator can learn inverse maps which are local in nature, which in turn promotes generalizability when testing with out-of-distribution samples.

Imbalanced Fault Diagnosis Using Conditional Wasserstein Generative Adversarial Networks With Switchable Normalization

Imbalanced Fault Diagnosis Using Conditional Wasserstein Generative Adversarial Networks With Switchable Normalization

CWGAN-GP with residual network model for lithium-ion battery thermal image data expansion with quantitative metrics

Lithium batteries find extensive applications in energy storage. Temperature is a crucial indicator for assessing the state of lithium-ion batteries, and numerous experiments require thermal images of lithium-ion batteries for research purposes. However, acquiring thermal imaging samples of lithium-ion battery faults is challenging due to factors such as high experimental costs and associated risks. To address this, our study proposes the utilization of a Conditional Wasserstein Generative Adversarial Network with Gradient Penalty and Residual Network (CWGAN-GP with Residual Network) to augment the dataset of thermal images depicting lithium-ion battery faults. We employ various evaluation metrics to quantitatively analyze and compare the generated thermal images of lithium-ion batteries. Subsequently, the expanded dataset, comprising four types of thermal images depicting lithium-ion battery faults, is input into a Mask Region-based Convolutional Neural Network for training. The results demonstrate that the proposed model surpasses both traditional Generative Adversarial Network and Wasserstein Generative Adversarial Network in terms of the quality of generated thermal images of lithium-ion batteries. Moreover, the augmentation of the dataset leads to an improvement in the fault diagnosis accuracy of the Mask Region-based Convolutional Neural Network.

Intelligent fault diagnosis for air handing units based on improved generative adversarial network and deep reinforcement learning

Data-driven Automatic fault detection and diagnosis (AFDD) for air handling units (AHUs) is crucial for ensuring the stable operation and energy consumption of the heating ventilation air-conditioning (HVAC) system. However, traditional machine learning methods often underperform when confronted with insufficient training sample data, especially when lacking samples from the fault types. Based on the issues of insufficient samples from the fault types and imbalanced training dataset, this study proposes a novel AFDD approach using transformer integrated conditional Wasserstein generative adversarial network and deep reinforcement learning (TCWGAN-DRL) to synthesize the fault data and select high quality synthetic data samples. Firstly, we utilize the proposed TransCWGAN to synthesize fault samples. Then, reinforcement learning is utilized to select high quality synthetic samples. Finally, the filtered samples and the real fault samples are merged to form the training dataset for conventional supervised learning classifiers. Experimental results demonstrate that the enriched training dataset can effectively improve the AFDD results and outperforms recently published existing methods, for instance, compared to the suboptimal model, our method exhibits an increase in fault recognition accuracy of 4.9%, 3.66%, and 4.02% when the number of real fault samples is 15, 20, and 30, respectively.

Deep learning based on Transformer architecture for power system short-term voltage stability assessment with class imbalance

Most existing data-driven power system short-term voltage stability assessment (STVSA) approaches presume class-balanced input data. However, in practical applications, the occurrence of short-term voltage instability following a disturbance is minimal, leading to a significant class imbalance problem and a consequent decline in classifier performance. This work proposes a Transformer-based STVSA method to address this challenge. By utilizing the basic Transformer architecture, a stability assessment Transformer (StaaT) is developed as a classification model to reflect the correlation between the operational states of the system and the resulting stability outcomes. To combat the negative impact of imbalanced datasets, this work employs a conditional Wasserstein generative adversarial network with gradient penalty (CWGAN-GP) for synthetic data generation, aiding in the creation of a balanced, representative training set for the classifier. Semi-supervised clustering learning is implemented to enhance clustering quality, addressing the lack of a unified quantitative criterion for short-term voltage stability. Numerical tests on the IEEE 39-bus test system extensively demonstrate that the proposed method exhibits robust performance under class imbalances up to 100:1 and noisy environments, and maintains consistent effectiveness even with an increased penetration of renewable energy. Comparative results reveal that the CWGAN-GP generates more balanced datasets than traditional oversampling methods and that the StaaT outperforms other deep learning algorithms. This study presents a compelling solution for real-world STVSA applications that often face class imbalance and data noise challenges.

Multi-modal generative adversarial networks for synthesizing time-series structural impact responses

The process of validating newly-defined state observers can potentially require a significant amount of data gathered from instrumentation. However, collecting data for high-rate dynamic events (short time-scale impact and shock) can be very expensive. Additionally, experiments performed for collecting data can provide highly variable results while the high-energy impacts introduce damage into the structure being tested, consequently resulting in different results for subsequent tests. This paper proposes the use of Generative Adversarial Networks (GANs) to generate data that supplements the experimental datasets required for the validation of state observers. GANs are a class of deep learning models used for generating data statistically comparable to that on which it was trained. They are an ideal candidate for validation due to their consistency, speed, and portability during inference. In this work, data collected from an electronics package under shock is used to examine the generative ability of GANs. This paper proposes a conditional Wasserstein GAN (CWGAN) implementation for the production of synthetic high-rate dynamic vibrations, and introduces the conditional input to the critic at the layer towards the end as opposed to the first layers. Results suggest the generative model proposed is capable of producing data statistically similar to the provided training data. The generated data is compared to the training data, and the advantages and limitations of the model are explored. The model and its artifacts are provided as supplemental material to this article and shared through a public repository.

Solution of physics-based inverse problems using conditional generative adversarial networks with full gradient penalty

The solution of probabilistic inverse problems for which the corresponding forward problem is constrained by physical principles is challenging. This is especially true if the dimension of the inferred vector is large and the prior information about it is in the form of a collection of samples. In this work, a novel deep learning based approach is developed and applied to solving these types of problems. The approach utilizes samples of the inferred vector drawn from the prior distribution and a physics-based forward model to generate training data for a conditional Wasserstein generative adversarial network (cWGAN). The cWGAN learns the probability distribution for the inferred vector conditioned on the measurement and produces samples from this distribution. The cWGAN developed in this work differs from earlier versions in that its critic is required to be 1-Lipschitz with respect to both the inferred and the measurement vectors and not just the former. This leads to a loss term with the full (and not partial) gradient penalty. It is shown that this rather simple change leads to a stronger notion of convergence for the conditional density learned by the cWGAN and a more robust and accurate sampling strategy. Through numerical examples it is shown that this change also translates to better accuracy when solving inverse problems. The numerical examples considered include illustrative problems where the true distribution and/or statistics are known, and a more complex inverse problem motivated by applications in biomechanics.

A residential load forecasting method for multi-attribute adversarial learning considering multi-source uncertainties

The rapid development of the Internet of Things and device-level meters provides accurate energy consumption data for various household devices, and making full use of these data to achieve higher accuracy of residential load forecasting is an open problem. Moreover, the uncertainties and insufficient feature extraction of variables make it challenging for individual residential load forecasts. In this regard, this paper proposes a novel short-term residential load forecasting method. It has the following aspects: (1) Considering the multi-source uncertainties, a multi-source uncertainties divide-and-conquer mechanism is proposed, which classifies the residential load from the device level according to different sources of uncertainties and forecasts separately. (2) Considering the multiple attributes of influencing variables, a multi-attribute adversarial learning mechanism based on the conditional Wasserstein generative adversarial network with gradient penalty (cWGAN-GP) is designed. In cWGAN-GP, this paper uses a long short-term memory network to implement the generator and uses a convolutional neural network to implement the discriminator. The former is used to extract temporal features, while the latter is used to extract spatial features, and they realize the fusion of multiple attributes through adversarial training to improve prediction accuracy. Compared with existing load forecasting methods, our proposed method’s mean absolute percentage error decreased by 1.5–38.9%, mean square error decreased by 3.5–35.9%, normalized root MSE decreased by 1.7–19.9%, and mean absolute error decreased by 3.7–27.0%.

A tropical cyclone intensity prediction model using conditional generative adversarial network

This study proposes a model for predicting tropical cyclone (TC) intensity evolution by utilizing the conditional Wasserstein generative adversarial network with gradient penalty (CWGAN-GP) for TC-related risk assessment. The TC intensity evolution is treated as a random variable, which is the output of a nonlinear system conditioned on a series of input random variables that characterize both the TC state (e.g., translation speed and potential intensity) and environment (e.g., vertical wind shear and ocean mixed layer depth). To represent the high-dimensional non-Gaussian probabilistic characteristics, the system is modeled by the CWGAN-GP and calibrated concerning the dataset of 1010 historical TCs from the western North Pacific basin. The over-ocean and overland branches of the model are calibrated separately and then merged. The suitability of the established model is validated by comparing the predicted TC intensity to the observations and the results of existing models such as those based on linear regression. Numerical results indicate that the proposed model successfully replicates the probabilistic properties such as the non-Gaussian marginal distribution of the intensity change and the input-output joint distribution and moments between the TC intensity and predictors, which the linear regression model cannot provide. Further application examples are also carried out by using the model to simulate the evolution of historical TC events and the extreme TC intensities in Southern China, demonstrating its potential role in the TC hazard assessment.

Crossline Reconstruction of 3D Seismic Data Using 3D cWGAN: A Comparative Study on Sleipner Seismic Survey Data

In seismic data acquisition, data loss can occur, particularly with the use of streamer systems in marine seismic exploration. These systems often cause spatial aliasing problems by having close inline intervals and wide crossline intervals to maximize the exploration range. To improve the resolution of seismic data in the crossline direction, various machine learning techniques have been employed for crossline data reconstruction. In this study, we introduce a 3D cWGAN (conditional Wasserstein generative adversarial network) for interpolating 3D seismic data. We evaluate the model’s performance by comparing it with 2D cWGAN and 3D U-Net. In this study, two interpolation strategies are employed to reconstruct missing data in the crossline direction. The first strategy uses a 2D network, which trains a model using inline data and applies the trained model to the crossline direction via 2D cWGAN. The second strategy employs a 3D network, which uses the 3D volume of the seismic data directly via 3D cWGAN and 3D U-Net. We demonstrate the effectiveness of the proposed method using the Sleipner CO2 4D seismic survey dataset. Our results show that the 3D cWGAN is more efficient in enhancing resolution and computation compared to the 2D cWGAN or 3D U-Net.

TS-GAN: Time-series GAN for Sensor-based Health Data Augmentation

Deep learning has achieved significant success on intelligent medical treatments, such as automatic diagnosis and analysis of medical data. To train an automatic diagnosis system with high accuracy and strong robustness in healthcare, sufficient training data are required when using deep learning-based methods. However, given that the data collected by sensors that are embedded in medical or mobile devices are inadequate, it is challenging to train an effective and efficient classification model with state-of-the-art performance. Inspired by generative adversarial networks (GANs), we propose TS-GAN, a Time-series GAN architecture based on long short-term memory (LSTM) networks for sensor-based health data augmentation, thereby improving the performance of deep learning-based classification models. TS-GAN aims to learn a generative model that creates time-series data with the same space and time dependence as the real data. Specifically, we design an LSTM-based generator for creating realistic data and an LSTM-based discriminator for determining how similar the generated data are to real data. In particular, we design a sequential-squeeze-and-excitation module in the LSTM-based discriminator to better understand space dependence of real data, and apply the gradient penalty originated from Wasserstein GANs in the training process to stabilize the optimization. We conduct comparative experiments to evaluate the performance of TS-GAN with TimeGAN, C-RNN-GAN and Conditional Wasserstein GANs through discriminator loss, maximum mean discrepancy, visualization methods and classification accuracy on health datasets of ECG_200, NonInvasiveFatalECG_Thorax1, and mHealth, respectively. The experimental results show that TS-GAN exceeds other state-of-the-art time-series GANs in almost all the evaluation metrics, and the classifier trained on synthetic datasets generated by TS-GAN achieves the highest classification accuracy of 97.50% on ECG_200, 94.12% on NonInvasiveFatalECG_Thorax1, and 98.12% on mHealth, respectively.

Deep learning prediction of post-SBRT liver function changes and NTCP modeling in hepatocellular carcinoma based on DGAE-MRI.

Stereotactic body radiation therapy (SBRT) produces excellent local control for patients with hepatocellular carcinoma (HCC). However, the risk of toxicity for normal liver tissue is still a limiting factor. Normal tissue complication probability (NTCP) models have been proposed to estimate the toxicity with the assumption of uniform liver function distribution, which is not optimal. With more accurate regional liver functional imaging available for individual patient, we can improve the estimation and be more patient-specific. To develop normal tissue complication probability (NTCP) models using pre-/during-treatment (RT) dynamic Gadoxetic Acid-enhanced (DGAE) MRI for adaptation of RT in a patient-specific manner in hepatocellular cancer (HCC) patients who receive SBRT. 24 of 146 HCC patients who received SBRT underwent DGAE MRI. Physical doses were converted into EQD2 for analysis. Voxel-by-voxel quantification of the contrast uptake rate (k1) from DGAE-MRI was used to quantify liver function. A logistic dose-response model was used to estimate the fraction of liver functional loss, and NTCP was estimated using the cumulative functional reserve model for changes in Child-Pugh (C-P) scores. Model parameters were calculated using maximum-likelihood estimations. During-RT liver functional maps were predicted from dose distributions and pre-RT k1 maps with a conditional Wasserstein generative adversarial network (cWGAN). Imaging prediction quality was assessed using root-mean-square error (RMSE) and structural similarity (SSIM) metrics. The dose-response and NTCP were fit on both original and cWGAN predicted images and compared using a Wilcoxon signed-rank test. Logistic dose response models for changes in k1 yielded D50 of 35.2 (95% CI: 26.7-47.5) Gy and k of 0.62 (0.49-0.75) for the whole population. The high baseline ALBI (poor liver function) subgroup showed a significantly smaller D50 of 11.7 (CI: 9.06-15.4) Gy and larger k of 0.96 (CI: 0.74-1.22) compared to a low baseline ALBI (good liver function) subgroup of 54.8 (CI: 38.3-79.1) Gy and 0.59 (CI: 0.48-0.74), with p-values of<0.001 and=0.008, respectively, which indicates higher radiosensitivity for the worse baseline liver function cohort. Subset analyses were also performed for high/low baseline CP subgroups. The corresponding NTCP models showed good agreement for the fit parameters between cWGAN predicted and the ground-truth during-RT images with no statistical differences for low ALBI subgroup. NTCP models which incorporate voxel-wise functional information from DGAE-MRI k1 maps were successfully developed and feasibility was demonstrated in a small patient cohort. cWGAN predicted functional maps show promise for estimating localized patient-specific response to RT and warrant further validation in a larger patient cohort.

3D Point Cloud Classification with ACGAN-3D and VACWGAN-GP

Machine learning and deep learning techniques are widely used to make sense of 3D point cloud data which became ubiquitous and important due to the recent advances in 3D scanning technologies and other sensors. In this work, we propose two networks to predict the class of the input 3D point cloud: 3D Auxiliary Classifier Generative Adversarial Network (ACGAN-3D) and Versatile Auxiliary Conditional Wasserstein Generative Adversarial Network with Gradient Penalty (VACWGAN-GP). Unlike other classifiers, we are able to enlarge the limited data set with the data produced by generative models. We consequently aim to increase the success of the model by training it with more data. As suggested by the conventional ACGAN models, in addition to the real dataset, we train the Discriminator with synthetic data generated by the Generator using the class label. By doing so, we ensure that Discriminator can discriminate between the real data and the synthetic data. Thus, as the training evolves, the Generator is trained to produce more realistic synthetic data, which in turn forces Discriminator to classify or discriminate better. Defined originally on 2D images, our ACGAN-3D modifies this conventional ACGAN model in order to classify 3D point clouds by updating the neural network layers. Our second model VACWGAN-GP, on the other hand, demonstrates similar abilities with more stable training by replacing its Discriminator with Critic and by modifying its loss function. In this model, we managed to merge Wasserstein GAN-GP with conditional GAN in order to improve the classifier?s performance. The proposed models ACGAN-3D and VACWGAN-GP were tested extensively on 3D datasets and comparisons with the other state-of-the-art studies have revealed our clear advantages on various aspects. While ACGAN-3D can be preferred with its compact design, our second method VACWGAN-GP stands out for higher performance.

Image Reconstruction for Electrical Impedance Tomography (EIT) With Improved Wasserstein Generative Adversarial Network (WGAN)

The image reconstruction of electrical impedance tomography (EIT) is highly ill-posed and nonlinear. Because of the poor nonlinear fitting ability of analytical algorithms, reconstructed images of these algorithms are blurry and lack detailed features. Although high-quality EIT images can be obtained by applying deep-learning networks to image reconstruction, the interpretability and generalization ability of the network are difficult to guarantee. A deep-learning structure, namely conditional Wasserstein generative adversarial network with attention mechanism (CWGAN-AM), is proposed for EIT image reconstruction. CWGAN-AM consists of an imaging module, a generator, and a discriminator. The initial conductivity image obtained by the imaging module is added to both the generator and the discriminator as a constraint to improve the stability of reconstruction. The enhanced residual blocks (ERBs), the structure of residual in residual (RIR), and the attention unit are used in the generator to further improve the reconstruction accuracy for the inclusion boundary. The imaging results indicate that CWGAN-AM can accurately recover the irregular boundaries of inclusions, and effective reconstruction can be accomplished for the new conductivity distribution (inclusions with size/shape variations) and noisy samples.

Combing data-driven and model-driven methods for high proportion renewable energy distribution network reliability evaluation

To accurately evaluate the reliability of the renewable energy distribution network, a combined data-driven and model-driven method of distribution network reliability evaluation are proposed. The approach improves the traditional reliability evaluation method in the uncertainty scenario generation and fault consequence analysis stages. In the scenario generation stage, an interpretable network architecture capable of adaptively learning the inherent randomness, volatility, and time variability properties of renewable energy is designed. The proposed DG output uncertainty generation method integrates the Conditional Wasserstein Generative Adversarial Network (CWGAN) based scenario generation and clustering-based scenario reduction. It accurately portrays uncertainty in distributed generation (DG) output power. In the fault consequence analysis stage, an optimization model for the reliability evaluation of distribution networks based on mixed integer linear programming is proposed. The model fully accounts for substation protection, branch line protection, network reconfiguration, DG, and energy storage system (ESS) components. Time-series analytical expressions for the corresponding reliability indices are established. Finally, with the combination of data-driven and model-driven methods, a distribution network reliability evaluation framework considering the uncertainty of renewable energy is proposed. Simulation tests are performed on a modified IEEE 123 bus system. The results verify the effectiveness and feasibility of the method. The proposed simulation analysis method can effectively quantify the distribution networks’ reliability compared to the traditional method. It has higher calculation accuracy.

SFCWGAN-BiTCN with Sequential Features for Malware Detection

In the field of adversarial attacks, the generative adversarial network (GAN) has shown better performance. There have been few studies applying it to malware sample supplementation, due to the complexity of handling discrete data. More importantly, unbalanced malware family samples interfere with the analytical power of malware detection models and mislead malware classification. To address the problem of the impact of malware family imbalance on accuracy, a selection feature conditional Wasserstein generative adversarial network (SFCWGAN) and bidirectional temporal convolutional network (BiTCN) are proposed. First, we extract the features of malware Opcode and API sequences and use Word2Vec to represent features, emphasizing the semantic logic between API tuning and Opcode calling sequences. Second, the Spearman correlation coefficient and the whale optimization algorithm extreme gradient boosting (WOA-XGBoost) algorithm are combined to select features, filter out invalid features, and simplify structure. Finally, we propose a GAN-based sequence feature generation algorithm. Samples were generated using the conditional Wasserstein generative adversarial network (CWGAN) on the imbalanced malware family dataset, added to the trainset to supplement the samples, and trained on BiTCN. In comparison, in tests on the Kaggle and DataCon datasets, the model achieved detection accuracies of 99.56% and 96.93%, respectively, which were 0.18% and 2.98% higher than the models of other methods.