Sign in to use this feature.

Years

Between: -

Subjects

remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline

Journals

remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline

Article Types

Countries / Regions

remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline
remove_circle_outline

Search Results (945)

Search Parameters:
Keywords = adversarial samples

Order results
Result details
Results per page
Select all
Export citation of selected articles as:
24 pages, 3030 KB  
Article
Fire Resistance Prediction in FRP-Strengthened Structural Elements: Application of Advanced Modeling and Data Augmentation Techniques
by Ümit Işıkdağ, Yaren Aydın, Gebrail Bekdaş, Celal Cakiroglu and Zong Woo Geem
Processes 2025, 13(10), 3053; https://doi.org/10.3390/pr13103053 - 24 Sep 2025
Viewed by 11
Abstract
In order to ensure the earthquake safety of existing buildings, retrofitting applications come to the fore in terms of being fast and cost-effective. Among these applications, fiber-reinforced polymer (FRP) composites are widely preferred thanks to their advantages such as high strength, corrosion resistance, [...] Read more.
In order to ensure the earthquake safety of existing buildings, retrofitting applications come to the fore in terms of being fast and cost-effective. Among these applications, fiber-reinforced polymer (FRP) composites are widely preferred thanks to their advantages such as high strength, corrosion resistance, applicability without changing the cross-section and easy assembly. This study presents a data augmentation, modeling, and comparison-based approach to predict the fire resistance (FR) of FRP-strengthened reinforced concrete beams. The aim of this study was to explore the role of data augmentation in enhancing prediction accuracy and to find out which augmentation method provides the best prediction performance. The study utilizes an experimental dataset taken from the existing literature. The dataset contains inputs such as varying geometric dimensions and FRP-strengthening levels. Since the original dataset used in the study consisted of 49 rows, the data size was increased using augmentation methods to enhance accuracy in model training. In this study, Gaussian noise, Regression Mixup, SMOGN, Residual-based, Polynomial + Noise, PCA-based, Adversarial-like, Quantile-based, Feature Mixup, and Conditional Sampling data augmentation methods were applied to the original dataset. Using each of them, individual augmented datasets were generated. Each augmented dataset was firstly trained using eXtreme Gradient Boosting (XGBoost) with 10-fold cross-validation. After selecting the best-performing augmentation method (Adversarial-like) based on XGBoost results, the best-performing augmented dataset was later evaluated in HyperNetExplorer, a more advanced NAS tool that can find the best performing hyperparameter optimized ANN for the dataset. ANNs achieving R2 = 0.99, MSE = 22.6 on the holdout set were discovered in this stage. This whole process is unique for the FR prediction of structural elements in terms of the data augmentation and training pipeline introduced in this study. Full article
(This article belongs to the Special Issue Machine Learning Models for Sustainable Composite Materials)
Show Figures

Figure 1

20 pages, 3123 KB  
Article
Enhancing Melanoma Diagnosis in Histopathology with Deep Learning and Synthetic Data Augmentation
by Alex Rodriguez Alonso, Ana Sanchez Diez, Goikoane Cancho Galán, Rafael Ibarrola Altuna, Gonzalo Irigoyen Miró, Cristina Penas Lago, Mª Dolores Boyano López and Rosa Izu Belloso
Bioengineering 2025, 12(9), 1001; https://doi.org/10.3390/bioengineering12091001 - 21 Sep 2025
Viewed by 298
Abstract
Accurate diagnosis of melanoma using hematoxylin and eosin (H&E)-stained histological images is often challenged by the scarcity and imbalance of biomedical datasets, limiting the performance of deep learning models. This study investigates the impact of synthetic image generation, via generative adversarial networks (GAN), [...] Read more.
Accurate diagnosis of melanoma using hematoxylin and eosin (H&E)-stained histological images is often challenged by the scarcity and imbalance of biomedical datasets, limiting the performance of deep learning models. This study investigates the impact of synthetic image generation, via generative adversarial networks (GAN), on training automatic classifiers based on the ResNet-18 architecture. Two experimental setups were designed: one using only real images and another combining real images with synthetic ones of the melanocytic nevus class to balance the dataset. Models were trained and evaluated at resolutions up to 1024 × 1024 pixels, employing standard classification metrics and the Fréchet Inception Distance (FID) to assess the quality of the generated images. The results suggest that although mixed models do not consistently outperform those trained exclusively on real data, they achieve competitive performance, particularly in terms of specificity and reduction in false negatives. This study supports the use of synthetic data as a complementary tool in scenarios where the acquisition of new samples is limited and lays the groundwork for future research in conditional generation and synthesis of malignant samples. In our best-performing model (1024 × 1024 px, 50 epochs, mixed dataset), we achieved an accuracy of 96.00%, a specificity of 97.00%, and a reduction in false negatives from 80 to 75 cases compared with real-only training. These results highlight the potential of synthetic augmentation to improve clinically relevant outcomes, particularly in reducing missed melanoma diagnoses. Full article
Show Figures

Graphical abstract

16 pages, 5654 KB  
Article
Target Recognition for Ultra-Wideband Radio Fuzes Using 1D-CGAN-Augmented 1D-CNN
by Kaiwei Wu, Shijun Hao, Yanbin Liang, Bing Yang and Zhonghua Huang
Entropy 2025, 27(9), 980; https://doi.org/10.3390/e27090980 - 19 Sep 2025
Viewed by 286
Abstract
In ultra-wideband (UWB) radio fuzes, the signal processing unit’s capability to rapidly and accurately extract target characteristics under battlefield conditions directly determines detonation precision and reliability. Escalating electronic warfare creates complex electromagnetic environments that compromise UWB fuze reliability through false alarms and missed [...] Read more.
In ultra-wideband (UWB) radio fuzes, the signal processing unit’s capability to rapidly and accurately extract target characteristics under battlefield conditions directly determines detonation precision and reliability. Escalating electronic warfare creates complex electromagnetic environments that compromise UWB fuze reliability through false alarms and missed detections. This study pioneers a novel signal processing architecture. The framework integrates: (1) fixed-parameter Least Mean Squares (LMS) front-end filtering for interference suppression; (2) One-Dimensional Convnlutional Neural Network (1D-CNN) recognition trained on One-Dimensional Conditional Generative Adversarial Network (1D-CGAN)-augmented datasets. Validated on test samples, the system achieves 0% false alarm/miss detection rates and 97.66% segment recognition accuracy—representing a 5.32% improvement over the baseline 1D-CNN model trained solely on original data. This breakthrough resolves energy-threshold detection’s critical vulnerability to deliberate jamming while establishing a new technical framework for UWB fuze operation in contested spectra. Full article
(This article belongs to the Section Multidisciplinary Applications)
Show Figures

Figure 1

17 pages, 86811 KB  
Article
The Role of Feature Vector Scale in the Adversarial Vulnerability of Convolutional Neural Networks
by Hyun-Cheol Park and Sang-Woong Lee
Mathematics 2025, 13(18), 3026; https://doi.org/10.3390/math13183026 - 19 Sep 2025
Viewed by 206
Abstract
In image classification, convolutional neural networks (CNNs) remain vulnerable to visually imperceptible perturbations, often called adversarial examples. Although various hypotheses have been proposed to explain this vulnerability, a clear cause has not been established. We hypothesize an unfair learning effect: samples are learned [...] Read more.
In image classification, convolutional neural networks (CNNs) remain vulnerable to visually imperceptible perturbations, often called adversarial examples. Although various hypotheses have been proposed to explain this vulnerability, a clear cause has not been established. We hypothesize an unfair learning effect: samples are learned unevenly depending on the scale (norm) of their feature vectors in feature space. As a result, feature vectors with different scales exhibit different levels of robustness against noise. To test this hypothesis, we conduct vulnerability tests on CIFAR-10 using a standard convolutional classifier, analyzing cosine similarity between original and perturbed feature vectors, as well as error rates across scale intervals. Our experiments show that small-scale feature vectors are highly vulnerable. This is reflected in low cosine similarity and high error rates, whereas large-scale feature vectors consistently exhibit greater robustness with high cosine similarity and low error rates. These findings highlight the critical role of feature vector scale in adversarial vulnerability. Full article
(This article belongs to the Special Issue The Application of Deep Neural Networks in Image Processing)
Show Figures

Figure 1

27 pages, 5866 KB  
Article
DCGAN Feature-Enhancement-Based YOLOv8n Model in Small-Sample Target Detection
by Peng Zheng, Yun Cheng, Wei Zhu, Bo Liu, Chenhao Ye, Shijie Wang, Shuhong Liu and Jinyin Bai
Computers 2025, 14(9), 389; https://doi.org/10.3390/computers14090389 - 15 Sep 2025
Viewed by 354
Abstract
This paper proposes DCGAN-YOLOv8n, an integrated framework that significantly advances small-sample target detection by synergizing generative adversarial feature enhancement with multi-scale representation learning. The model’s core contribution lies in its novel adversarial feature enhancement module (AFEM), which leverages conditional generative adversarial networks to [...] Read more.
This paper proposes DCGAN-YOLOv8n, an integrated framework that significantly advances small-sample target detection by synergizing generative adversarial feature enhancement with multi-scale representation learning. The model’s core contribution lies in its novel adversarial feature enhancement module (AFEM), which leverages conditional generative adversarial networks to reconstruct discriminative multi-scale features while effectively mitigating mode collapse. Furthermore, the architecture incorporates a deformable multi-scale feature pyramid that dynamically fuses generated high-resolution features with hierarchical semantic representations through an attention mechanism. The proposed triple marginal constraint optimization jointly enhances intra-class compactness and inter-class separation, thereby structuring a highly discriminative feature space. Extensive experiments on the NWPU VHR-10 dataset demonstrate state-of-the-art performance, with the model achieving an mAP50 of 90.46% and an mAP50-95 of 57.06%, representing significant improvements of 4.52% and 4.08% over the baseline YOLOv8n, respectively. These results validate the framework’s effectiveness in addressing critical challenges of feature representation scarcity and cross-scale adaptation in data-limited scenarios. Full article
(This article belongs to the Special Issue Machine Learning Applications in Pattern Recognition)
Show Figures

Figure 1

27 pages, 6873 KB  
Review
Deep Generative Modeling of Protein Conformations: A Comprehensive Review
by Tuan Minh Dao and Taseef Rahman
BioChem 2025, 5(3), 32; https://doi.org/10.3390/biochem5030032 - 15 Sep 2025
Viewed by 476
Abstract
Proteins are dynamic macromolecules whose functions are intricately linked to their structural flexibility. Recent breakthroughs in deep learning have enabled accurate prediction of static protein structures. However, understanding protein function is more complex. It often requires access to a diverse ensemble of conformations. [...] Read more.
Proteins are dynamic macromolecules whose functions are intricately linked to their structural flexibility. Recent breakthroughs in deep learning have enabled accurate prediction of static protein structures. However, understanding protein function is more complex. It often requires access to a diverse ensemble of conformations. Traditional sampling techniques exist to help with this. These include molecular dynamics and Monte Carlo simulations. These techniques can explore conformational landscapes. However, they have limitations as they are often limited by high computational cost and suffer from slow convergence. In response, deep generative models (DGMs) have emerged as a powerful alternative for efficient and scalable protein conformation sampling. Leveraging architectures such as variational autoencoders, normalizing flows, generative adversarial networks, and diffusion models, DGMs can learn complex, high-dimensional distributions over protein conformations directly from data. This survey on generative models for protein conformation sampling provides a comprehensive overview of recent advances in this emerging field. We categorize existing models based on generative architecture, structural representation, and target tasks. We also discuss key datasets, evaluation metrics, limitations, and opportunities for integrating physics-based knowledge with data-driven models. By bridging machine learning and structural biology, DGMs are poised to transform our ability to model, design, and understand dynamic protein behavior. Full article
Show Figures

Figure 1

17 pages, 394 KB  
Article
Boosting Clean-Label Backdoor Attacks on Graph Classification
by Yadong Wang, Zhiwei Zhang, Ye Yuan and Guoren Wang
Electronics 2025, 14(18), 3632; https://doi.org/10.3390/electronics14183632 - 13 Sep 2025
Viewed by 282
Abstract
Graph Neural Networks (GNNs) have become a cornerstone for graph classification, yet their vulnerability to backdoor attacks remains a significant security concern. While clean-label attacks provide a stealthier approach by preserving original labels, they tend to be less effective in graph settings compared [...] Read more.
Graph Neural Networks (GNNs) have become a cornerstone for graph classification, yet their vulnerability to backdoor attacks remains a significant security concern. While clean-label attacks provide a stealthier approach by preserving original labels, they tend to be less effective in graph settings compared to traditional dirty-label methods. This performance gap arises from the inherent dominance of rich, benign structural patterns in target-class graphs, which overshadow the injected backdoor trigger during the GNNs’ learning process. We demonstrate that prior strategies, such as adversarial perturbations used in other domains to suppress benign features, fail in graph settings due to the amplification effects of the GNNs’ message-passing mechanism. To address this issue, we propose two strategies aimed at enabling the model to better learn backdoor features. First, we introduce a long-distance trigger injection method, placing trigger nodes at topologically distant locations. This enhances the global propagation of the backdoor signal while interfering with the aggregation of native substructures. Second, we propose a vulnerability-aware sample selection method, which identifies graphs that contribute more to the success of the backdoor attack based on low model confidence or frequent forgetting events. We conduct extensive experiments on benchmark datasets such as NCI1, NCI109, Mutagenicity, and ENZYMES, demonstrating that our approach significantly improves attack success rates (ASRs) while maintaining a low clean accuracy drop (CAD) compared to existing methods. This work offers valuable insights into manipulating the competition between benign and backdoor features in graph-structured data. Full article
(This article belongs to the Special Issue Security and Privacy for AI)
Show Figures

Figure 1

32 pages, 4502 KB  
Article
An Integrated and Robust Vision System for Internal and External Thread Defect Detection with Adversarial Defense
by Liu Fu, Leqi Li, Gengpei Zhang and Zhihao Jiang
Sensors 2025, 25(18), 5664; https://doi.org/10.3390/s25185664 - 11 Sep 2025
Viewed by 372
Abstract
In industrial automation, detecting defects in threaded components is challenging due to their complex geometry and the concealment of micro-flaws. This paper presents an integrated vision system capable of inspecting both internal and external threads with high robustness. A unified imaging platform ensures [...] Read more.
In industrial automation, detecting defects in threaded components is challenging due to their complex geometry and the concealment of micro-flaws. This paper presents an integrated vision system capable of inspecting both internal and external threads with high robustness. A unified imaging platform ensures synchronized capture of thread surfaces, while advanced image enhancement techniques improve clarity under motion blur and low-light conditions. To overcome limited defect samples, we introduce a generative data augmentation strategy that diversifies training data. For detection, a lightweight and optimized deep learning model achieves higher precision and efficiency compared with existing YOLO variants. Moreover, we design a dual-defense mechanism that effectively mitigates stealthy adversarial perturbations, such as alpha channel attacks, preserving system reliability. Experimental results demonstrate that the proposed framework delivers accurate, secure, and efficient thread defect detection, offering a practical pathway toward reliable industrial vision systems. Full article
(This article belongs to the Section Intelligent Sensors)
Show Figures

Figure 1

38 pages, 3071 KB  
Article
A Hybrid Framework for the Sensitivity Analysis of Software-Defined Networking Performance Metrics Using Design of Experiments and Machine Learning Techniques
by Chekwube Ezechi, Mobayode O. Akinsolu, Wilson Sakpere, Abimbola O. Sangodoyin, Uyoata E. Uyoata, Isaac Owusu-Nyarko and Folahanmi T. Akinsolu
Information 2025, 16(9), 783; https://doi.org/10.3390/info16090783 - 9 Sep 2025
Viewed by 416
Abstract
Software-defined networking (SDN) is a transformative approach for managing modern network architectures, particularly in Internet-of-Things (IoT) applications. However, ensuring the optimal SDN performance and security often needs a robust sensitivity analysis (SA). To complement existing SA methods, this study proposes a new SA [...] Read more.
Software-defined networking (SDN) is a transformative approach for managing modern network architectures, particularly in Internet-of-Things (IoT) applications. However, ensuring the optimal SDN performance and security often needs a robust sensitivity analysis (SA). To complement existing SA methods, this study proposes a new SA framework that integrates design of experiments (DOE) and machine-learning (ML) techniques. Although existing SA methods have been shown to be effective and scalable, most of these methods have yet to hybridize anomaly detection and classification (ADC) and data augmentation into a single, unified framework. To fill this gap, a targeted application of well-established existing techniques is proposed. This is achieved by hybridizing these existing techniques to undertake a more robust SA of a typified SDN-reliant IoT network. The proposed hybrid framework combines Latin hypercube sampling (LHS)-based DOE and generative adversarial network (GAN)-driven data augmentation to improve SA and support ADC in SDN-reliant IoT networks. Hence, it is called DOE-GAN-SA. In DOE-GAN-SA, LHS is used to ensure uniform parameter sampling, while GAN is used to generate synthetic data to augment data derived from typified real-world SDN-reliant IoT network scenarios. DOE-GAN-SA also employs a classification and regression tree (CART) to validate the GAN-generated synthetic dataset. Through the proposed framework, ADC is implemented, and an artificial neural network (ANN)-driven SA on an SDN-reliant IoT network is carried out. The performance of the SDN-reliant IoT network is analyzed under two conditions: namely, a normal operating scenario and a distributed-denial-of-service (DDoS) flooding attack scenario, using throughput, jitter, and response time as performance metrics. To statistically validate the experimental findings, hypothesis tests are conducted to confirm the significance of all the inferences. The results demonstrate that integrating LHS and GAN significantly enhances SA, enabling the identification of critical SDN parameters affecting the modeled SDN-reliant IoT network performance. Additionally, ADC is also better supported, achieving higher DDoS flooding attack detection accuracy through the incorporation of synthetic network observations that emulate real-time traffic. Overall, this work highlights the potential of hybridizing LHS-based DOE, GAN-driven data augmentation, and ANN-assisted SA for robust network behavioral analysis and characterization in a new hybrid framework. Full article
(This article belongs to the Special Issue Data Privacy Protection in the Internet of Things)
Show Figures

Graphical abstract

25 pages, 3787 KB  
Article
Early Detection of Tomato Gray Mold Based on Multispectral Imaging and Machine Learning
by Xiaohao Zhong, Huicheng Li, Yixin Cai, Ying Deng, Haobin Xu, Jun Tian, Shuang Liu, Maomao Hou, Haiyong Weng, Lijing Wang, Miaohong Ruan, Fenglin Zhong, Chunhui Zhu and Lu Xu
Horticulturae 2025, 11(9), 1073; https://doi.org/10.3390/horticulturae11091073 - 5 Sep 2025
Viewed by 460
Abstract
Gray mold is one of the major diseases affecting tomato production. Its early symptoms are often inconspicuous, yet the disease spreads rapidly, leading to severe economic losses. Therefore, the development of efficient and non-destructive early detection technologies is of critical importance. At present, [...] Read more.
Gray mold is one of the major diseases affecting tomato production. Its early symptoms are often inconspicuous, yet the disease spreads rapidly, leading to severe economic losses. Therefore, the development of efficient and non-destructive early detection technologies is of critical importance. At present, multispectral imaging-based detection methods are constrained by two major bottlenecks: limited sample size and single modality, which hinder precise recognition at the early stage of infection. To address these challenges, this study explores a detection approach integrating multispectral fluorescence and reflectance imaging, combined with machine learning algorithms, to enhance early recognition of tomato gray mold. Particular emphasis is placed on evaluating the effectiveness of multimodal information fusion in extracting early disease features, and on elucidating the quantitative relationships between disease progression and key physiological indicators such as chlorophyll content, water content, malondialdehyde levels, and antioxidant enzyme activities. Furthermore, an improved WGAN-GP (Wasserstein Generative Adversarial Network with Gradient Penalty) is employed to alleviate data scarcity under small-sample conditions. The results demonstrate that multimodal data fusion significantly improves model sensitivity to early-stage disease detection, while WGAN-GP-based data augmentation effectively enhances learning performance with limited samples. The Random Forest model achieved an early recognition precision of 97.21% on augmented datasets, and transfer learning models attained an overall precision of 97.56% in classifying different disease stages. This study provides an effective approach for the early prediction of tomato gray mold, with potential application value in optimizing disease management strategies and reducing environmental impact. Full article
(This article belongs to the Section Plant Pathology and Disease Management (PPDM))
Show Figures

Figure 1

30 pages, 6483 KB  
Article
The Generative Adversarial Approach: A Cautionary Tale of Finite Samples
by Marcos Escobar-Anel and Yiyao Jiao
Algorithms 2025, 18(9), 564; https://doi.org/10.3390/a18090564 - 5 Sep 2025
Viewed by 360
Abstract
Given the relevance and wide use of the Generative Adversarial (GA) methodology, this paper focuses on finite samples to better understand its benefits and pitfalls. We focus on its finite-sample properties from both statistical and numerical perspectives. We set up a simple and [...] Read more.
Given the relevance and wide use of the Generative Adversarial (GA) methodology, this paper focuses on finite samples to better understand its benefits and pitfalls. We focus on its finite-sample properties from both statistical and numerical perspectives. We set up a simple and ideal “controlled experiment” where the input data are an i.i.d. Gaussian series where the mean is to be learned, and the discriminant and generator are in the same distributional family, not a neural network (NN), as in the popular GAN. We show that, even with the ideal discriminant, the classical GA methodology delivers a biased estimator while producing multiple local optima, confusing numerical methods. The situation worsens when the discriminator is in the correct parametric family but is not the oracle, leading to the absence of a saddle point. To improve the quality of the estimators within the GA method, we propose an alternative loss function, the alternative GA method, that leads to a unique saddle point with better statistical properties. Our findings are intended to start a conversation on the potential pitfalls of GA and GAN methods. In this spirit, the ideas presented here should be explored in other distributional cases and will be extended to the actual use of an NN for discriminators and generators. Full article
(This article belongs to the Section Evolutionary Algorithms and Machine Learning)
Show Figures

Figure 1

15 pages, 2578 KB  
Article
Effects of Composite Cross-Entropy Loss on Adversarial Robustness
by Ning Ding and Knut Möller
Electronics 2025, 14(17), 3529; https://doi.org/10.3390/electronics14173529 - 4 Sep 2025
Viewed by 463
Abstract
Convolutional neural networks (CNNs) can efficiently extract image features and perform corresponding classification. Typically, the CNN architecture uses the softmax layer to map the extracted features to classification probabilities, and the cost function used for training is the cross-entropy loss. In this paper, [...] Read more.
Convolutional neural networks (CNNs) can efficiently extract image features and perform corresponding classification. Typically, the CNN architecture uses the softmax layer to map the extracted features to classification probabilities, and the cost function used for training is the cross-entropy loss. In this paper, we evaluate the influence of a number of representative composite cross-entropy loss functions on the learned feature space at the fully connected layer, when a target classification is introduced into a multi-class classification task. In addition, the accuracy and robustness of CNN models trained with different composite cross-entropy loss functions are investigated. Improved robustness is achieved by changing the loss between the input and the target classification. Preliminary experiments were conducted using ResNet-50 on the Cholec80 dataset for surgical tool recognition. Furthermore, the model trained with the proposed composite cross-entropy loss incorporating another target all-one classification demonstrates a 31% peak improvement in adversarial robustness. Adversarial training with target adversarial samples yields 80% robustness against PGD attack. This investigation shows that the careful choice of the loss function can improve the robustness of CNN models. Full article
(This article belongs to the Special Issue Convolutional Neural Networks and Vision Applications, 4th Edition)
Show Figures

Figure 1

25 pages, 4385 KB  
Article
Robust DeepFake Audio Detection via an Improved NeXt-TDNN with Multi-Fused Self-Supervised Learning Features
by Gul Tahaoglu
Appl. Sci. 2025, 15(17), 9685; https://doi.org/10.3390/app15179685 - 3 Sep 2025
Viewed by 1000
Abstract
Deepfake audio refers to speech that has been synthetically generated or altered through advanced neural network techniques, often with a degree of realism sufficient to convincingly imitate genuine human voices. As these manipulations become increasingly indistinguishable from authentic recordings, they present significant threats [...] Read more.
Deepfake audio refers to speech that has been synthetically generated or altered through advanced neural network techniques, often with a degree of realism sufficient to convincingly imitate genuine human voices. As these manipulations become increasingly indistinguishable from authentic recordings, they present significant threats to security, undermine media integrity, and challenge the reliability of digital authentication systems. In this study, a robust detection framework is proposed, which leverages the power of self-supervised learning (SSL) and attention-based modeling to identify deepfake audio samples. Specifically, audio features are extracted from input speech using two powerful pretrained SSL models: HuBERT-Large and WavLM-Large. These distinctive features are then integrated through an Attentional Multi-Feature Fusion (AMFF) mechanism. The fused features are subsequently classified using a NeXt-Time Delay Neural Network (NeXt-TDNN) model enhanced with Efficient Channel Attention (ECA), enabling improved temporal and channel-wise feature discrimination. Experimental results show that the proposed method achieves a 0.42% EER and 0.01 min-tDCF on ASVspoof 2019 LA, a 1.01% EER on ASVspoof 2019 PA, and a pooled 6.56% EER on the cross-channel ASVspoof 2021 LA evaluation, thus highlighting its effectiveness for real-world deepfake detection scenarios. Furthermore, on the ASVspoof 5 dataset, the method achieved a 7.23% EER, outperforming strong baselines and demonstrating strong generalization ability. Moreover, the macro-averaged F1-score of 96.01% and balanced accuracy of 99.06% were obtained on the ASVspoof 2019 LA dataset, while the proposed method achieved a macro-averaged F1-score of 98.70% and balanced accuracy of 98.90% on the ASVspoof 2019 PA dataset. On the highly challenging ASVspoof 5 dataset, which includes crowdsourced, non-studio-quality audio, and novel adversarial attacks, the proposed method achieves macro-averaged metrics exceeding 92%, with a precision of 92.07%, a recall of 92.63%, an F1-measure of 92.35%, and a balanced accuracy of 92.63%. Full article
Show Figures

Figure 1

19 pages, 3770 KB  
Article
Segmentation of 220 kV Cable Insulation Layers Using WGAN-GP-Based Data Augmentation and the TransUNet Model
by Liang Luo, Song Qing, Yingjie Liu, Guoyuan Lu, Ziying Zhang, Yuhang Xia, Yi Ao, Fanbo Wei and Xingang Chen
Energies 2025, 18(17), 4667; https://doi.org/10.3390/en18174667 - 2 Sep 2025
Viewed by 606
Abstract
This study presents a segmentation framework for images of 220 kV cable insulation that addresses sample scarcity and blurred boundaries. The framework integrates data augmentation using the Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) and the TransUNet architecture. Considering the difficulty and [...] Read more.
This study presents a segmentation framework for images of 220 kV cable insulation that addresses sample scarcity and blurred boundaries. The framework integrates data augmentation using the Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP) and the TransUNet architecture. Considering the difficulty and high cost of obtaining real cable images, WGAN-GP generates high-quality synthetic data to expand the dataset and improve the model’s generalization. The TransUNet network, designed to handle the structural complexity and indistinct edge features of insulation layers, combines the local feature extraction capability of convolutional neural networks (CNNs) with the global context modeling strength of Transformers. This combination enables accurate delineation of the insulation regions. The experimental results show that the proposed method achieves mDice, mIoU, MP, and mRecall scores of 0.9835, 0.9677, 0.9840, and 0.9831, respectively, with improvements of approximately 2.03%, 3.05%, 2.08%, and 1.98% over a UNet baseline. Overall, the proposed approach outperforms UNet, Swin-UNet, and Attention-UNet, confirming its effectiveness in delineating 220 kV cable insulation layers under complex structural and data-limited conditions. Full article
(This article belongs to the Special Issue Fault Detection and Diagnosis of Power Distribution System)
Show Figures

Figure 1

32 pages, 2361 KB  
Article
Exploring the Use and Misuse of Large Language Models
by Hezekiah Paul D. Valdez, Faranak Abri, Jade Webb and Thomas H. Austin
Information 2025, 16(9), 758; https://doi.org/10.3390/info16090758 - 1 Sep 2025
Viewed by 596
Abstract
Language modeling has evolved from simple rule-based systems into complex assistants capable of tackling a multitude of tasks. State-of-the-art large language models (LLMs) are capable of scoring highly on proficiency benchmarks, and as a result have been deployed across industries to increase productivity [...] Read more.
Language modeling has evolved from simple rule-based systems into complex assistants capable of tackling a multitude of tasks. State-of-the-art large language models (LLMs) are capable of scoring highly on proficiency benchmarks, and as a result have been deployed across industries to increase productivity and convenience. However, the prolific nature of such tools has provided threat actors with the ability to leverage them for attack development. Our paper describes the current state of LLMs, their availability, and their role in benevolent and malicious applications. In addition, we propose how an LLM can be combined with text-to-speech (TTS) voice cloning to create a framework capable of carrying out social engineering attacks. Our case study analyzes the realism of two different open-source TTS models, Tortoise TTS and Coqui XTTS-v2, by calculating similarity scores between generated and real audio samples from four participants. Our results demonstrate that Tortoise is able to generate realistic voice clone audios for native English speaking males, which indicates that easily accessible resources can be leveraged to create deceptive social engineering attacks. As such tools become more advanced, defenses such as awareness, detection, and red teaming may not be able to keep up with dangerously equipped adversaries. Full article
Show Figures

Figure 1

Back to TopTop