Figure - available from: Soft Computing
This content is subject to copyright. Terms and conditions apply.
Source publication
Molecular biology studies on cancer, using gene expression datasets, have revealed that the datasets have a very small number of samples. Obtaining medical data is difficult and expensive due to privacy constraints. Accuracy of classifiers depends greatly on the quality and quantity of input data. The problem of small sample size or small data size...
Citations
... The Generative Adversarial Network (GAN) [2] is a machine learning framework that consists of two competing neural networks: the generator (G) and the discriminator (D). These networks engage in a zero-sum game [14][15][16]. ...
... CBiGAN [21] was developed to minimize BiGAN's generalization losses and complete the learning process at lower power consumption. It has been stated that WGAN [22] provides better quality output due to improved stability in the optimization process and the correlation between the convergence of the generator and the loss function compared to the standard GAN model [14]. Figure 1 shows examples of faces generated by InterfaceGAN, ProGAN, StyleGAN3, and a real image from the CelebA dataset. ...
The rapid advancement of synthetic media, while beneficial, has also spawned GAN-generated deepfakes, which pose risks, including misinformation and digital fraud. This paper investigates the detectability of GAN-generated static images, focusing on residual artifacts that are imperceptible to humans but detectable through digital analysis. Our approach introduces three key advancements: (1) a taxonomy for classifying GAN residues in deepfake detection; (2) a unique mixed dataset combining StyleGAN3, ProGAN, and InterfaceGAN to aid cross-model detection research; and (3) a combination of frequency space analysis and RGB color correlation methods to improve artifact detection. Covering three different transform methods, three GAN models, and twelve classification methods, ours is the most comprehensive study of detection of static deepfake face images produced by GANs. Our results demonstrate that artifact-based detection can achieve high accuracy, precision, recall, and F1 scores, challenging prior assumptions about the detectability of synthetic face images.
... Generative Adversarial Networks (GANs) have diverse applications, including image conversion [2] and data augmentation [3] [4]. Data augmentation addresses data scarcity by generating realistic synthetic samples, enhancing AI models' performance and generalization, as seen in improving classification algorithms [5]. Unlike traditional augmentation methods, which apply transformations like convolution or noise, GANs generate synthetic data that preserves the real data's distribution and intrinsic features, aiding in tasks that are challenging to capture, such as generating seismic signals [6]. ...
Human Activity Recognition (HAR) is crucial in health and sports, where integrating wearable sensors with Artificial Intelligence (AI) leads to innovative solutions. However, data collection can be labor-intensive and slow, hindering project timelines. Data augmentation presents a viable solution to this challenge. Our study enhanced our original dataset using a TimeGAN-based neural network for wearable device sensors. The experimental outcomes confirm that it's feasible to construct efficient classifiers using the synthetic data generated, which closely resembles real data. This approach saves significant time in data collection and accelerates the development of wearable technologies. Consequently, we developed an effective classifier for HAR using synthetic data.
... Given that these models typically require a large amount of data to be constructed and achieve high performance [1], Generative Adversarial Networks (GANs) emerge as a promising solution to the data availability problem [2] [3]. This data augmentation technique is particularly effective, generating realistic synthetic samples that can enhance the performance and generalization of AI models, such as improving the performance of classification algorithms [4]. ...
Human Activity Recognition (HAR) with artificial intelligence fosters the development of innovative solutions. However, building AI models often requires a substantial amount of data and can be time-consuming. In this context, our work adopted the TimeGAN technique for data augmentation, facilitating the construction of a more efficient model. We developed a classifier that integrates both synthetic and real data. This strategy significantly reduces the time required for data collection and may accelerate the development of new wearable technologies. This approach represents a promising step in optimizing development processes in AI applications for HAR, enhancing the speed and effectiveness of technological innovation.
... Chaudhari et al. [10] utilized a modified generative adversarial model to generate the data samples with multivariate noise and a Gaussian distribution. The MG-GAN network was used to reach the saddle point quicker since the created data is in feature space. ...
... The WT-GAN model achieved a value of around 97% accuracy for the colon, leukemia, and prostate datasets. The proposed model accuracy results are compared with the existing paper [10], the colon, leukemia, and prostate cancer achieved 91.7%, 88.4%, and 93.6%, respectively. The proposed WT-GAN augmentation strategy improves the classification accuracy of 5.3%, 8.6%, and 3.4% against traditional tabular GAN. ...
Several diverse fields including the healthcare system and drug development sectors have benefited immensely through the adoption of deep learning (DL), which is a subset of artificial intelligence (AI) and machine learning (ML). Cancer makes up a significant percentage of the illnesses that cause early human mortality across the globe, and this situation is likely to rise in the coming years, especially when non-communicable illnesses are not considered. As a result, cancer patients would greatly benefit from precise and timely diagnosis and prediction. Deep learning (DL) has become a common technique in healthcare due to the abundance of computational power. Gene expression datasets are frequently used in major DL-based applications for illness detection, notably in cancer therapy. The quantity of medical data, on the other hand, is often insufficient to fulfill deep learning requirements. Microarray gene expression datasets are used for training procedures despite their extreme dimensionality, limited volume of data samples, and sparsely available information. Data augmentation is commonly used to expand the training sample size for gene data. The Wasserstein Tabular Generative Adversarial Network (WT-GAN) model is used for the data augmentation process for generating synthetic data in this proposed work. The correlation-based feature selection technique selects the most relevant characteristics based on threshold values. Deep FNN and ML algorithms train and classify the gene expression samples. The augmented data give better classification results (> 97%) when using WT-GAN for cancer diagnosis.
... Convolutional networks in particular have achieved remarkable performances across diverse tasks from classification [32,33], segmentation [34], reconstruction [35] and registration [36]. In certain applications, deep learning systems have surpassed human experts, fueling enthusiasm for a broader technology-powered transformation in imaging diagnostics [37][38][39][40][41]. ...
Medical imaging has become integral to modern healthcare, enabling non-invasivevisualization and assessment of anatomical structures. However, medical imagingdatasets are often limited in size and diversity, constraining development of robust analysis algorithms. Meanwhile, generative adversarial networks (GANs) haveachieved remarkable synthetic image generation capabilities. This paper comprehensively reviews contemporary GAN techniques and evaluates their effectiveness producing synthetic medical images to augment scarce training data. Six prevalent GAN architectures were trained on diverse medical imaging datasets. A systematic hyperparameter optimization strategy coupled with quantitative imageanalysis reveal substantial variability in output fidelity and diversity. Downstreamsegmentation task performance provides further domain-specific assessments on theutility of the generated datasets. The study reveals that while select advanced GANscan produce seemingly realistic medical images, the synthetic data consistentlyunderperforms real datasets on specialized tasks. The results caution against indiscriminate use of GAN-produced medical images but highlight paths for developing tailored GAN solutions for enhanced training. Keywords deep learning; generative adversarial networks; medical imaging; synthetic data
... Mode collapse occurs when the generator does not learn to generalize and instead fixates on a single class, or components from many classes. The generated data may lack diversity, causing unbalanced generated data and furthering the problem that data augmentation aims to solve [39,40]. Thereby, the trained generator and models with different architectures require tertiary means of evaluation [37,41,42,43,44]. ...
... Data augmentation can increase the dataset size by generating synthetic data points. GANs are capable of generating synthetic data, and a modified generator GAN (MG-GAN) generates synthetic data that conform to a Gaussian distribution [37]. In comparison to traditional data augmentation methods, a MG-GAN significantly improved cancer type classification accuracy for a breast cancer patient gene expression dataset. ...
This paper presents a transformative methodology that harnesses the power of digital twin (DT) technology for the advanced condition monitoring of lithium-ion batteries (LIBs) in electric vehicles (EVs). In contrast to conventional solutions, our approach eliminates the need to calibrate sensors or add additional hardware circuits. The digital replica works seamlessly alongside the embedded battery management system (BMS) in an EV, delivering real-time signals for monitoring. Our system is a significant step forward in ensuring the efficiency and sustainability of EVs, which play an essential role in reducing carbon emissions. A core innovation lies in the integration of the digital twin into the battery monitoring process, reshaping the landscape of energy storage and alternative power sources such as lithium-ion batteries. Our comprehensive system leverages a cloud-based IoT network and combines both physical and digital components to provide a holistic solution. The physical side encompasses offline modeling, where a long short-term memory (LSTM) algorithm trained with various learning rates (LRs) and optimized by three types of optimizers ensures precise state-of-charge (SOC) predictions. On the digital side, the digital twin takes center stage, enabling the real-time monitoring and prediction of battery activity. A particularly innovative aspect of our approach is the utilization of a time-series generative adversarial network (TS-GAN) to generate synthetic data that seamlessly complement the monitoring process. This pioneering use of a TS-GAN offers an effective solution to the challenge of limited real-time data availability, thus enhancing the system’s predictive capabilities. By seamlessly integrating these physical and digital elements, our system enables the precise analysis and prediction of battery behavior. This innovation—particularly the application of a TS-GAN for data generation—significantly contributes to optimizing battery performance, enhancing safety, and extending the longevity of lithium-ion batteries in EVs. Furthermore, the model developed in this research serves as a benchmark for future digital energy storage in lithium-ion batteries and comprehensive energy utilization. According to statistical tests, the model has a high level of precision. Its exceptional safety performance and reduced energy consumption offer promising prospects for sustainable and efficient energy solutions. This paper signifies a pivotal step towards realizing a cleaner and more sustainable future through advanced EV battery management.
... In 2020, Chaudhari et al. [18] firstly proposed modified generator GAN (MG-GAN), which is fed with original data along with minimalistic multivariate noise to generate data with Gaussian distribution. In 2021, Kwon et al. [19] indicated that GANs are not effective with whole genes, and expanded RNA expression data for selected significant genes using GANs. ...
Background
Although gene expression data play significant roles in biological and medical studies, their applications are hampered due to the difficulty and high expenses of gathering them through biological experiments. It is an urgent problem to generate high quality gene expression data with computational methods. WGAN-GP, a generative adversarial network-based method, has been successfully applied in augmenting gene expression data. However, mode collapse or over-fitting may take place for small training samples due to just one discriminator is adopted in the method.
Results
In this study, an improved data augmentation approach MDWGAN-GP, a generative adversarial network model with multiple discriminators, is proposed. In addition, a novel method is devised for enriching training samples based on linear graph convolutional network. Extensive experiments were implemented on real biological data.
Conclusions
The experimental results have demonstrated that compared with other state-of-the-art methods, the MDWGAN-GP method can produce higher quality generated gene expression data in most cases.
... Artificial intelligence algorithms are of essential use in the prognosis and diagnosis of medical conditions. Some diseases (e.g., cancers) are complicated, showing much variation in staging, duration, location, treatment response, cell differentiation and origin, and understanding of their cause [1]. Microarray gene expression data were recently used to determine the kind of cancer a person has [2], and the rough weakness of the employed datasets is the curse of dimensionality [3]. ...
... Chaudhari et al. 11 and Viñas et al. 12 proposed GAN-based data augmentation studies for gene expression data. Moreover, Moreno-Barea et al. 13 developed a conditional GAN method using gene expression data, and Ahmen et al. 14 proposed a GAN architecture that integrates two omics datasets to generate omics data from the other omics dataset. ...
The accurate prediction of patients with complex diseases, such as Alzheimer's disease (AD), as well as disease stages, including early- and late-stage cancer, is challenging owing to substantial variability among patients and limited availability of clinical data. Deep metric learning has emerged as a promising approach for addressing these challenges by improving data representation. In this study, we propose a joint triplet loss model with a semi-hard constraint (JTSC) to represent data in a small number of samples. JTSC strictly selects semi-hard samples by switching anchors and positive samples during the learning process in triplet embedding and combines a triplet loss function with an angular loss function. Our results indicate that JTSC significantly improves the number of appropriately represented samples during training when applied to the gene expression data of AD and to cancer stage prediction tasks. Furthermore, we demonstrate that using an embedding vector from JTSC as an input to the classifiers for AD and cancer stage prediction significantly improves classification performance by extracting more accurate features. In conclusion, we show that feature embedding through JTSC can aid in classification when there are a small number of samples compared to a larger number of features.
