Realistic Samples Research Articles

Semiclassical Scattering by Edge Imperfections in Topological Insulators Under Magnetic Field

We study the scattering of edge states of 2D topological insulator in the uniform external magnetic field due to edge imperfections, common in realistic 2D topological insulator samples. The external magnetic field breaks time-reversal symmetry, opening the possibility of the scattering of otherwise topologically protected fermionic edge states. The scattering happens to be always an over-barrier event, irrespective of the shape of the edge deformation and magnitude of the magnetic field. We use the advanced Pokrovsky–Khalatnikov semiclassical approach, which allows us to obtain analytically both the main exponential and pre- exponential factors of the scattering amplitude for wide classes of analytic deformation profiles.

Classification and Comparison of Data Augmentation Techniques

Data augmentation is a technical method that generates additional data samples by applying different transformations and processes to the original training data. It aims to increase the size and diversity of the dataset, thus improving the models' generalization ability. This paper provides a detailed introduction to geometric transformations, data mixing, automated data augmentation methods, and generative adversarial methods. Geometric transformations manipulate the spatial configuration of data samples to create new variants that retain the original content. Data mixing techniques blend multiple data samples to produce novel training examples, promoting robustness and reducing overfitting. Automated data augmentation methods leverage search algorithms to discover optimal augmentation policies that improve model performance without manual intervention. Generative adversarial methods employ generative adversarial networks to synthesize realistic data samples that enrich the training dataset. These methods find broad application in domains including object detection, image classification, natural language processing, and speech recognition, significantly improving model performance. The paper explores the advantages and limitations of data augmentation techniques in practical applications and identifies future research directions and challenges. By systematically classifying and selecting appropriate data augmentation methods for specific application scenarios, this review provides a theoretical foundation and practical guidance for real-world applications.

Back analysis of geomechanical parameters based on a data augmentation algorithm and machine learning technique

Accurate geomechanical parameters are key factors for stability evaluation, disaster forecasting, structural design, and supporting optimization. The intelligent back analysis method based on the monitored information is widely recognized as the most efficient and cost-effective technique for inverting parameters. To address the low accuracy of measured data, and the scarcity of comprehensive datasets, this study proposes an innovative back analysis framework tailored for small sample sizes. We introduce a multi-faceted back analysis approach that combines data augmentation with advanced optimization and machine learning techniques. The auxiliary classifier generative adversarial network (ACGAN)-based data augmentation algorithm is first employed to generate synthetic yet realistic samples that adhere to the underlying probability distribution of the original data, thereby expanding the dataset and mitigating the impact of small sample sizes. Subsequently, we harness the power of optimized particle swarm optimization (OPSO) integrated with support vector machine (SVM) to mine the intricate nonlinear relationships between input and output variables. Then, relying on a case study, the validity of the augmented data and the performance of the developed OPSO-SVM algorithms based on two different sample sizes are studied. Results show that the new datasets generated by ACGAN almost coincide with the actual monitored convergences, exhibiting a correlation coefficient exceeding 0.86. Furthermore, the superiority of the OPSO-SVM algorithm is also demonstrated by comparing the displacement prediction capability of various algorithms through four indices. It is also indicated that the relative error of the predicted displacement values reduces from almost 20% to 5% for the OPSO-SVM model trained with 25 samples and that trained with 625 samples. Finally, the inversed parameters and corresponding convergences predicted by the two OPSO-SVM models trained with different samples are discussed, indicating the feasibility of the combination application of ACGAN and OPSO-SVM in back analysis of geomechanical parameters. This endeavor not only facilitates the progression of underground engineering analysis in scenarios with limited data, but also serves as a pivotal reference for both researchers and practitioners alike.

SCAWA scales: A new digital tool for wrinkles clinical grading based on AI.

Clinical assessment of wrinkle depth is essential for efficacy evaluations of anti-ageing products. Standardized photographic scales, representative of different wrinkle depths are often used by experts to assign subjects reliable grades. These tools, based on real pictures, usually exist as hard copies (printed books or sheets) for invivo gradings. Our project aims at developing a methodology to create digital standardized computer-generated scales, allowing photograph and real-life gradings, and providing raters with greater comfort, accessibility, and flexibility in their construction, thanks to the artificial intelligence significative contribution. A completely new approach, based on machine learning, allows the creation of Standardized ColorFace® AI-based Wrinkle Assessment (SCAWA) scales. Instead of using real photographs, the scale images are computer-generated. A generative adversarial network (GAN) is trained to create realistic wrinkle samples that are finely controllable by exploring the GAN latent space. Finally, the scale images are selected among hundreds of artificial images depicting natural wrinkle appearances, such as the illustrated wrinkle evolution is well-detailed (small gaps between grades), morphologically stable, and mathematically linear according to a criterion of wrinkle conspicuous depth. The created 12-point scale for crow's feet wrinkle evaluation on ColorFace® pictures is proven to be realistic, linear, and robustly and accurately usable for photograph assessments. The scale coherence in terms of image ranking has been validated, as well as its reliability and acceptability in real conditions of use. Additionally, the wrinkle grades obtained by the SCAWA scale are well correlated (R = 0.94) with the ones obtained by the Skin Aging Atlas on the same pictures. The AI methodology and digital format brought also interesting side results, such as an enhanced harmonization between experts and a higher representativeness, that is, a decrease of out-of-range pictures. SCAWA scale makes the most of machine learning to provide an innovative digital tool to ease wrinkles in visual assessment of pictures, while optimizing linearity, homogeneity, and accuracy aspects. The experts' enthusiastic feedback about the scale format and quality is promising regarding the adaptation of the methodology to other signs and a larger distribution of this tool in the market of cosmetic product efficacy assessment.

Sustainable urea treatment by environmental-friendly and highly hydrophilic vesicle-like iron phosphate-based carbon

Despite the versatile potential applications of urea, its unfavorable characteristics for conventional treatment methods hinder its utilization. Therefore, this study developed vesicle-like iron phosphate-based carbon (IP@C400) as a breakthrough urea removal and recovery material for a wide range of urea-containing sources. IP@C400 rapidly exhibited an exceptional capacity (2242 mg/g in 1 h) across a wide range of pH, even in synthetic hemodialysis wastewater with high urea concentrations and diverse co-existing components, compared with the 60 prominent adsorbents. The adsorption process followed dual Pseudo-kinetic, Langmuir-isotherm models with the involvement of primary robust physical (i.e., H-bonding and electrostatic interaction) and chemical mechanisms (i.e., hydrolysis). Remarkably, IP@C400 can maintain high urea removal (90 %) or recovery efficiency (95 %) even after 10 cycles with minimal leakages of Fe and P (far below WHO and EUWFD standards)—a significant improvement over disposable options. IP@C400 could also perform efficiently on batch and a new approach integrating with a naturally accessible material based on the fixed-bed column using low-range urea realistic samples, achieving 65.2 L water over 10 cycles with undetected urea, neutral pH, and well-aligned water safety standards with a minimal adsorbent dose (0.1 g.L−1) and economical cost ($0.05 L-1). Lastly, its environmentally friendly nature, which contains essential nutrients for plant growth, further enhances its recyclability after release. Thus, IP@C400 offers a solution to environmental sustainability and the urgent ultrapure water issue that industries are facing.

Reliable generation of privacy-preserving synthetic electronic health record time series via diffusion models.

Electronic health records (EHRs) are rich sources of patient-level data, offering valuable resources for medical data analysis. However, privacy concerns often restrict access to EHRs, hindering downstream analysis. Current EHR deidentification methods are flawed and can lead to potential privacy leakage. Additionally, existing publicly available EHR databases are limited, preventing the advancement of medical research using EHR. This study aims to overcome these challenges by generating realistic and privacy-preserving synthetic EHRs time series efficiently. We introduce a new method for generating diverse and realistic synthetic EHR time series data using denoizing diffusion probabilistic models. We conducted experiments on 6 databases: Medical Information Mart for Intensive Care III and IV, the eICU Collaborative Research Database (eICU), and non-EHR datasets on Stocks and Energy. We compared our proposed method with 8 existing methods. Our results demonstrate that our approach significantly outperforms all existing methods in terms of data fidelity while requiring less training effort. Additionally, data generated by our method yield a lower discriminative accuracy compared to other baseline methods, indicating the proposed method can generate data with less privacy risk. The proposed model utilizes a mixed diffusion process to generate realistic synthetic EHR samples that protect patient privacy. This method could be useful in tackling data availability issues in the field of healthcare by reducing barrier to EHR access and supporting research in machine learning for health. The proposed diffusion model-based method can reliably and efficiently generate synthetic EHR time series, which facilitates the downstream medical data analysis. Our numerical results show the superiority of the proposed method over all other existing methods.

Review of imbalanced fault diagnosis technology based on generative adversarial networks

Abstract In the field of industrial production, machine failures not only negatively affect productivity and product quality, but also lead to safety accidents, so it is crucial to accurately diagnose machine failures in time and take appropriate measures. However, machines cannot operate with faults for extended periods, and the diversity of fault modes results in limited data collection, posing challenges to building accurate fault prediction models. Despite recent advancements, intelligent fault diagnosis methods based on traditional sampling and machine learning have shown notable progress. Nonetheless, these methods heavily rely on human expertise, making it challenging to extract comprehensive feature information. To address these challenges, numerous imbalance fault diagnosis methods based on generative adversarial networks (GANs) have emerged, GANs can generate realistic samples that conform to the distribution of the original data, showing promising results in diagnosing imbalances in critical components such as bearings and gears, despite their great potential, GAN methods also face challenges, including difficulties in training and generating abnormal samples. However, whether it is GAN-based resampling technology or traditional sampling technology, there are fewer reviews on noise-containing imbalance, intra- and inter-class dual imbalance, multi-class imbalance, time series imbalance and other problems in small samples, and there is a lack of a more comprehensive summary of the solutions to the above imbalance problems. Therefore, the purpose of this paper is to deeply explore the imbalance problems under various failure modes, and review and analyze the research methods and results based on GANs on this basis. By suggesting future research directions, this paper aims to provide guidance and reference for research in the field of industrial production maintenance.

Discriminative Hamiltonian variational autoencoder for accurate tumor segmentation in data-scarce regimes

Deep learning has gained significant attention in medical image segmentation. However, the limited availability of annotated training data presents a challenge to achieving accurate results. In efforts to overcome this challenge, data augmentation techniques have been proposed. However, the majority of these approaches primarily focus on image generation. For segmentation tasks, providing both images and their corresponding target masks is crucial, and the generation of diverse and realistic samples remains a complex task, especially when working with limited training datasets. To this end, we propose a new end-to-end hybrid architecture based on Hamiltonian Variational Autoencoders (HVAE) and a discriminative regularization to improve the quality of generated images. Our method provides an accurate estimation of the joint distribution of the images and masks, resulting in the generation of realistic medical images with reduced artifacts and off-distribution instances. As generating 3D volumes requires substantial time and memory, our architecture operates on a slice-by-slice basis to segment 3D volumes, capitalizing on the richly augmented dataset. Experiments conducted on two public datasets, BRATS (MRI modality) and HECKTOR (PET modality), demonstrate the efficacy of our proposed method on different medical imaging modalities with limited data.

Conditional image-to-image translation generative adversarial network (cGAN) for fabric defect data augmentation

The availability of comprehensive datasets is a crucial challenge for developing artificial intelligence (AI) models in various applications and fields. The lack of large and diverse public fabric defect datasets forms a major obstacle to properly and accurately developing and training AI models for detecting and classifying fabric defects in real-life applications. Models trained on limited datasets struggle to identify underrepresented defects, reducing their practicality. To address these issues, this study suggests using a conditional generative adversarial network (cGAN) for fabric defect data augmentation. The proposed image-to-image translator GAN features a conditional U-Net generator and a 6-layered PatchGAN discriminator. The conditional U-Network (U-Net) generator can produce highly realistic synthetic defective samples and offers the ability to control various characteristics of the generated samples by taking two input images: a segmented defect mask and a clean fabric image. The segmented defect mask provides information about various aspects of the defects to be added to the clean fabric sample, including their type, shape, size, and location. By augmenting the training dataset with diverse and realistic synthetic samples, the AI models can learn to identify a broader range of defects more accurately. This technique helps overcome the limitations of small or unvaried datasets, leading to improved defect detection accuracy and generalizability. Moreover, this proposed augmentation method can find applications in other challenging fields, such as generating synthetic samples for medical imaging datasets related to brain and lung tumors.

A Survey on Surface Defect Inspection Based on Generative Models in Manufacturing

Surface defect inspection based on deep learning has demonstrated outstanding performance in improving detection accuracy and model generalization. However, the small scale of defect datasets always limits the application of deep models in industry. Generative models can obtain realistic samples in a very cheap way, which can effectively solve this problem and thus has received widespread attention in recent years. This paper provides a comprehensive analysis and summary of the current studies of surface defect inspection methods proposed between 2022 and 2024. First, according to the use of generative models, these methods are classified into four categories: Variational Auto-Encoders (VAEs), Generative Adversarial Networks (GANs), Diffusion Models (DMs), and multi-models. Second, the research status of surface defect inspection based on generative models in recent years is discussed from four aspects: sample generation, detection objective, inspection task, and learning model. Then, the public datasets and evaluation metrics that are commonly used for surface defect inspection are discussed, and a comparative evaluation of defect inspection methods based on generative models is provided. Finally, this study discusses the existing challenges for the defect inspection methods based on generative models, providing insights for future research.

A systematic approach to the analysis of illicit drugs for DNA with an overview of the problems encountered

Due to the restricted nature of illicit drugs, it is difficult to conduct research surrounding the analysis of this drug material for any potential DNA in sufficient quantities acceptable for high numbers of replicates. Therefore, the current research available in peer reviewed journals thus far regarding analysing illicit drugs for DNA has been performed under varying experimental conditions, often using surrogate chemicals in place of illicit drugs. The data presented within this study originated from the analysis of genuine illicit drugs prepared both in controlled environments and those seized at the Australian border (and therefore from an uncontrolled environment) to determine if DNA can be obtained from this type of material. This study has been separated into three main parts (total n=114 samples): firstly, methamphetamine synthesised within a controlled environment was spiked with both saliva and trace DNA to determine the yield following DNA extraction; secondly, methamphetamine also synthesised in a controlled environment but on a larger scale was tested for the amount of DNA added incidentally throughout the synthesis, including the additional steps of recrystallising, homogenising and “cutting” the drug material to simulate preparation for distribution; and thirdly, the detection of human DNA within samples of cocaine and heroin seized at the Australian border. The DNA Fast Flow Microcon Device was utilised to concentrate all replicates from the same source into one combined extract to improve the DNA profiles for the samples where no DNA spiking occurred. Full STR profiles were successfully obtained from drug samples spiked with both saliva and trace DNA. Methamphetamine was present in the final DNA extracts and caused incompatibilities with the quantification of DNA using Qubit. The yields of DNA from drugs not spiked with DNA sources were much lower, resulting in 36 % of samples yielding alleles where all others did not. These results were not unexpected given these were realistic drug samples where the history of the drug material was unknown. This is the first study to obtain DNA profiles from genuine illicit drug material in both controlled and uncontrolled environments and indicates that the analysis of illicit drugs for DNA is an avenue worth pursuing to provide information which can in turn assist with disrupting the supply of these drugs. Given that DNA profiling is carried out worldwide using essentially the same systems as described within this study, the potential for impact is on a national and international scale.

Superluminal light propagation in a three-level ladder system

Superluminal light propagation is typically accompanied by significant absorption that might prevent its observation in realistic samples. We propose an all-optical implementation exploiting the two-photon resonance in three-level media to overcome this problem. With several computational methods, we analyze three possible configurations of optically-dressed systems and identify an optimal configuration for superluminal propagation. Due to the far-detuned operating regime with low absorption, this scenario avoids the usual need for population inversion, gain assistance or nonlinear optical response. Our analysis covers a broad parameter space and aims for the identification of conditions where significant pulse advancement can be achieved at high transmission levels. In this context, a figure of merit is introduced accounting for a trade-off between the desired group-index values and transmission level. This quantity helps to identify the optimal characteristics of the dressing beam.

Counterpropagating optical solitary waves in orientation-modulated nematic liquid crystals

We investigate counterpropagating optical solitary waves – nematicons – in non-homogeneously oriented nematic liquid crystals. A non-symmetric angular distribution of the optic axis versus beam propagation provides the solitons with direction-dependent trajectories. By performing numerical experiments in realistic planar samples, we launch nematicons in opposite directions and examine the resulting non-specular transmission in terms of optical nonreciprocity and isolation.

Novel pyrazine Schiff-base derivative functionalized water-soluble polymer as a highly selective fluorescent chemosensor for Al3+ ions in pure aqueous solution

A novel water-soluble polymer P1 functionalized with pyrazine Schiff-base derivative was successfully synthesized. As an efficient “turn-on” fluorescent chemosensor, P1 showed high selectivity and anti-interference ability for the detection of Al3+ ions in pure aqueous solutions. The detection limit of P1 for Al3+ ions was as low as 3.68 × 10−8 M. P1 displayed good detectability for Al3+ ions over a wide pH range of 5–10. The detection of Al3+ ions by P1 has been proven to have excellent reversibility by alternately adding Al3+ ions and EDTA to P1 aqueous solution. Furthermore, P1 could be applied to quantitative determination of Al3+ ions in real water samples. P1-based test strips were also facilely prepared for convenient on-site qualitative detection of Al3+ ions in realistic aqueous samples.

Star Generative Adversarial VGG Network-Based Sample Augmentation for Insulator Defect Detection

Deep learning-based automated detection of insulator defects in electric power systems is a critical technological challenge, pivotal for ensuring reliability and efficiency in the global energy infrastructure. However, the effectiveness of the deep learning model is severely compromised by the scarcity of defective insulator samples. To tackle this problem, the present study proposes a style transfer approach utilizing an improved Star Generative Adversarial Network 2 (StarGAN2) model to generate artificial samples of faulty insulators, which adeptly synthesizes artificial faulty insulator samples on a one-to-many basis, markedly diminishing the necessity for extensive empirical data collection. Through the integration of identity loss, the proposed model ensures the fidelity of content and the preservation of critical defect semantics. Additionally, the proposed model incorporates a pre-trained Visual Geometry Group (VGG) network and perceptual loss, thus improving the quality of generated samples without additional artificial labeling. Finally, various experiments are conducted to assess the quality and authenticity of the generated samples and their impact on the detection model. The results demonstrate that StarGAN2 could generate realistic insulator defect samples and improve the performance of defect detection models.

Validation of the VUV-reflective coating for next-generation liquid xenon detectors

Coating detector materials with films highly reflective in the vacuum ultraviolet region improves sensitivity of the next-generation rare-event detectors that use liquid xenon. In this work, we investigate the MgF2-Al-MgF2 coating designed to achieve high reflectance at 175 nm, the mean wavelength of liquid xenon (LXe) scintillation. The coating was applied to an unpolished, passivated copper substrate mimicking a realistic detector component of the proposed nEXO experiment, as well as to two unpassivated substrates with “high” and “average” levels of polishing. After confirming the composition and morphology of the thin-film coating using TEM and EDS, the samples underwent reflectance measurements in LXe and gaseous nitrogen (GN2). Measurements in LXe exposed the coated samples to -100°C for several hours. No peeling of the coatings was observed after several thermal cycles. Polishing is found to strongly correlate with the measured specular reflectance (R spec). In particular, 5.8(5)% specular spike reflectance in LXe was measured for the realistic sample at 20° of incidence, while the values for similar angles of incidence on the high and average polish samples are 62.3(1.3)% and 27.4(7)%, respectively. At large angles (66°–75°), the R spec in LXe for the three samples increases to 23(5)%, 80(8)%, and 84(18)%, respectively. The R spec at around 45° was measured in both GN2 and LXe for average polish sample and shows a reasonable agreement. Importantly, the total reflectance of the samples is comparable and estimated to be 92(8)%, 85(8)%, and 83(8)% in GN2 for the realistic, average, and high polish samples, respectively. This is considered satisfactory for the next-generation LXe experiments that could benefit from using reflective films, such as nEXO and DARWIN, thus validating the design of the coating.

MRD‐GAN: Multi‐representation discrimination GAN for enhancing the diversity of the generated data

AbstractThe generative adversarial network (GAN) is a highly effective member of the generative models category and is extensively employed for generating realistic samples across various domains. The fundamental concept behind GAN involves two networks, a generator and a discriminator, competing against each other. During the training process, generator and discriminator networks encounter several issues that can potentially affect the quality and diversity of the generated samples. One such critical issue is mode collapse, where the generator fails to create varied samples. To tackle this issue, this article introduces a GAN approach called the multi‐representation discrimination GAN (MRD‐GAN). In this approach, the discriminator supports concurrent network discrimination flows to manage different representations of the data through various transformation functions, such as dimension rescaling, brightness adjustment, and gamma correction applied to the input data of the discriminator. We use a fusion function to aggregate the output of all flows and return a consolidated loss value to update the generator's weights. Hence, the discriminator conveys diverse feedback to the generator. The proposed approach has been evaluated on four distinct benchmarks, namely CelebA, Cifar‐10, Fashion‐Mnist, and Mnist. The experimental results demonstrate that the proposed approach surpasses the existing state‐of‐the‐art GAN models in terms of FID metric that measures the diversity of the generated samples. Significantly, the proposed approach demonstrates remarkable FID scores of 14.02, 30.19, 9.42, and 3.14 on the CelebA, Cifar‐10, Fashion‐Mnist, and Mnist datasets, respectively.

Evaluation of the SERS performances of Tabun and VX label-free detection in complex and multicomponent fluids

Nerve agents represent a serious threat to security worldwide. Chemical terrorism has become an alarming danger since the technological progresses have simplified the production of nerve agents. Therefore, there is an immediate demand for a fast and precise detection of these compounds on-site and real-time. In this perspective, Surface-Enhanced Raman Scattering (SERS) has emerged as a well-suited alternative for on-field detection.SERS performances of unfunctionalized SERS substrates were evaluated in realistic samples. Two nerve agents, Tabun and VX, were diluted in two matrix models: a contact lens solution, and a caffeine-based eye serum. The performance two research-grade instruments and two portable devices were compared. Despite the use of a small sampling volume of complex matrices without any sample pre-treatment, we achieved Tabun detection in both media, with a practical limit of detection (LOD) in the range of 7–9 ppm in contact lens liquid, and of 10.2 ppm in eye serum. VX detection turned out to be more challenging and was achieved only in contact lens solution, with a practical LOD in the range of 0.6–5 ppm.These results demonstrate the feasibility of on-field detection of nerve agents with SERS, that could be implemented when there is suspicion of chemical threat.

Velocity reconstruction in the era of DESI and Rubin/LSST. II. Realistic samples on the light cone

Velocity reconstruction in the era of DESI and Rubin/LSST. II. Realistic samples on the light cone

Uncertainty-Aware Surrogate Models for Airfoil Flow Simulations with Denoising Diffusion Probabilistic Models

Leveraging neural networks as surrogate models for turbulence simulation is a topic of growing interest. At the same time, embodying the inherent uncertainty of simulations in the predictions of surrogate models remains very challenging. The present study makes a first attempt to use denoising diffusion probabilistic models (DDPMs) to train an uncertainty-aware surrogate model for turbulence simulations. Due to its prevalence, the simulation of flows around airfoils with various shapes, Reynolds numbers, and angles of attack is chosen as the learning objective. Our results show that DDPMs can successfully capture the whole distribution of solutions and, as a consequence, accurately estimate the uncertainty of the simulations. The performance of DDPMs is also compared with varying baselines in the form of Bayesian neural networks and heteroscedastic models. Experiments demonstrate that DDPMs outperform the other methods regarding a variety of accuracy metrics. Besides, it offers the advantage of providing access to the complete distributions of uncertainties rather than providing a set of parameters. As such, it can yield realistic and detailed samples from the distribution of solutions.