Industrial Image Anomaly Detection via Synthetic-Anomaly Contrastive Distillation by Junxian Li, Mingxing Li, Shucheng Huang, Gang Wang, Xinjing Zhao

“…During training, <i>SACD</i> adopts a dual-branch strategy, where one branch encodes multi-scale features from defect-free images, while a Siamese anomaly branch processes synthetically corrupted counterparts. <i>FeaCali</i> modules are trained to strip out a student’s anomalous patterns in anomaly branches, enhancing the student network’s exclusive modeling of normal patterns. …”
Get full text

Deep Learning Approach in Seismology: Enhancing Earthquake Forecasting using K-Means Clustering and LSTM Networks by Tyanita Puti Marindah Wardhani, Zulkifli Tahir, Elly Warni, Anugrayani Bustamin, Muhammad Alief Fahdal Imran Oemar, Muhammad Alwi Kayyum

“…This study utilises historical seismic and proposes innovative data pre-processing methods using K-means clustering to build a Long Short-Term Memory (LSTM) model for earthquake forecasting to overcome high-disparity locations. …”
Get full text

Artificial Intelligence-Based Skin Lesion Analysis and Skin Cancer Detection by Momina Qureshi, Muhammad Athar Javed Sethi, Sayed Shahid Hussain

“…This study demonstrates a distinct training strategy by employing a two-phase approach: first, training only the newly added dense layers while maintaining the layers of the base model frozen, and then, fine-tuning the entire model. …”
Get full text

Explainable light-weight deep learning pipeline for improved drought stress identification by Aswini Kumar Patra, Aswini Kumar Patra, Lingaraj Sahoo

“…This visualization approach sheds light on the internal workings of the deep learning model, often regarded as a ”black box”. By revealing the model’s focus areas within the images, it enhances interpretability and fosters trust in the model’s decision-making process.Results and discussionOur proposed framework achieves superior performance, particularly with the DenseNet121 pre-trained network, reaching a precision of 97% to identify the stressed class with an overall accuracy of 91%. …”
Get full text

Privacy-Diffusion: Privacy-Preserving Stable Diffusion Without FHE and Differential Privacy by Po-Chu Hsu, Ziying Yu, Shuhei Mise, Hideaki Miyaji

“…By utilizing the irreversible property of neural network layers and the property that the predicted noise in the diffusion process is a normalized Gaussian distribution. …”
Get full text

A data driven framework for optimizing droplet microfluidics with residual block and Fourier enhanced networks by Alireza Samari, Kamal Jannati, Azadeh Jafari

“…However, traditional design approaches rely heavily on complex physical models or labor-intensive trial-and-error processes, making them time-consuming and expensive. …”
Get full text

Advanced device for measuring linear and angular components of small displacements of monitor object surfaces by Igor P. Miroshnichenko

“…Areas of application are the following: experimental research; assessment and diagnostics of the structural materials samples; investigations of fast wave processes in the layered structures of complex shape made of the anisotropic composite materials. …”
Get full text

Explainable Machine Learning in Critical Decision Systems: Ensuring Safe Application and Correctness by Julius Wiggerthale, Christoph Reich

“…Machine learning (ML) is increasingly used to support or automate decision processes in critical decision systems such as self driving cars or systems for medical diagnosis. …”
Get full text

Geometric Structures in Textile Design Made with 3D Printing by Cihat Okan Arikan, Samet Doğan, Deja Muck

“…In the production phase, the fused deposition modelling (FDM) process was chosen, and ABS and TPU materials were used. …”
Get full text

Enhancing Neural Network Interpretability Through Deep Prior-Guided Expected Gradients by Su-Ying Guo, Xiu-Jun Gong

“…In these high-stakes applications, understanding the decision-making processes of models is essential for ensuring trust and safety. …”
Get full text

Data-Efficient Bone Segmentation Using Feature Pyramid- Based SegFormer by Naohiro Masuda, Keiko Ono, Daisuke Tawara, Yusuke Matsuura, Kentaro Sakabe

“…Specifically, these include the data-efficient model, which deepens the hierarchical encoder by adding convolution layers to transformer blocks and increases feature map resolution within transformer blocks, and the FPN-based model, which enhances the decoder through a Feature Pyramid Network (FPN) and attention mechanisms. …”
Get full text

Inhibition of restenosis after balloon injury in rabbit vessels by integrin αvβ3-targeted 10058-F4 nanoparticles by Wenling Man, Qiang Fu, Yujia Wang, Chang Dong, Po Yang, Wei Li

“…The small molecule inhibitor of the c-Myc gene, 10058-F4, can play a significant inhibitory role in the process of vascular proliferation, and the locally targeted nano-delivery method allows the drug to be activated within the smooth muscle cell layer to prevent stenosis and does not retard endothelial healing.…”
Get full text

Three-Dimensional CKANs: UUV Noncooperative Target State Estimation Approach Based on 3D Convolutional Kolmogorov–Arnold Networks by Changjian Lin, Dan Yu, Shibo Lin

“…The incorporation of the Kolmogorov–Arnold representation within the convolutional layers enhances the network’s capacity for nonlinear expression and adaptability in processing spatial information. …”
Get full text

Research on entity recognition and alignment of APT attack based on Bert and BiLSTM-CRF by Xiuzhang YANG, Guojun PENG, Zichuan LI, Yangqi LYU, Side LIU, Chenguang LI

“…Also, the proposed model's training process is more stable, and the entire curve converges faster, which can achieve higher accuracy with fewer training batches. …”
Get full text

Research on entity recognition and alignment of APT attack based on Bert and BiLSTM-CRF by Xiuzhang YANG, Guojun PENG, Zichuan LI, Yangqi LYU, Side LIU, Chenguang LI

“…Also, the proposed model's training process is more stable, and the entire curve converges faster, which can achieve higher accuracy with fewer training batches. …”
Get full text

AMST2: aggregated multi-level spatial and temporal context-based transformer for robust aerial tracking by Hasil Park, Injae Lee, Dasol Jeong, Joonki Paik

“…Abstract Recently, many existing visual trackers have made significant progress by incorporating either spatial information from multi-level convolution layers or temporal information for tracking. However, the complementary advantages of both spatial and temporal information cannot be leveraged when these two types of information are used separately. …”
Get full text

Speed up integer-arithmetic-only inference via bit-shifting by Mingjun Song, Yiming Zhou, Mengmeng Song, Sujie Liu, Shungen Xiao, Youshun Zheng

“…Abstract Quantization is a widely adopted technique in model deployment as it offers a favorable trade-off between computational overhead and performance loss. …”
Get full text

An integrated workflow combining machine learning and wavelet transform for automated characterization of heterogeneous groundwater systems by Musaab A. A. Mohammed, Norbert P. Szabó, Abdelrhim Eltijani, Péter Szűcs

“…The delineated lithology aligned closely with the deterministically estimated shale volume and permeability, revealing higher permeability and lower shale volume in the sandy and gravelly layers. This model provides a robust foundation for groundwater flow and contaminant transport modeling and can be extended to other regions for improved aquifer management and development.…”
Get full text

Radar-Based Hand Gesture Recognition With Feature Fusion Using Robust CNN-LSTM and Attention Architecture by Irshad Khan, Young-Woo Kwon

“…We evaluate our method on the UWB-Gestures dataset, which has 12 gestures from eight people that were recorded using three X-Thru X4 UWB impulse radar sensors. Our processing pipeline integrates feature extraction, LSTM-attention blocks, and dense layers for final classification. …”
Get full text

Sky images based photovoltaic power forecasting: A novel approach with optimized VMD and Vision Mamba by Chenhao Cai, Leyao Zhang, Jianguo Zhou, Luming Zhou

“…Finally, by integrating the outputs from all processing steps through normalization and feed-forward layers, the final predictive results are produced. …”
Get full text