Deep Architecture Research Articles

Optimized Deep Learning Model Architecture for the Feature Extraction to Predict Trend in Stock Market

Predicting stock trends is a complex task influenced by various factors such as market sentiment, economic indicators, and company performance. Analysts often employ technical analysis, studying historical price patterns and trading volumes, as well as fundamental analysis, assessing financial statements and industry trends. Deep Learning models have also gained popularity for predicting stock trends, using algorithms to identify patterns and relationships in large datasets. Deep learning algorithms, particularly neural networks, excel at recognizing intricate patterns and relationships within complex datasets, making them well-suited for predicting stock prices, identifying trends, and managing risk. Hence, this paper proposed a Bird Swarm Optimization ARIMA LSTM (BSO-ARIMA-DL) model for stock trend prediction. The proposed BSO-ARIMA-DL model performance is applied in the company datasets Apple, Amazon, and Infosys for stock trend prediction. With the proposed BSO-ARIMA-DL model features are optimized for the identification of features in the dataset for the evaluation of optimal features. Upon the estimation of features, the ARIMA model with the LSTM architecture is implemented for the stock trend analysis. The proposed BSO-ARIMA-DL model deep learning model is implemented for the stock trend prediction in the companies. The results demonstrated that the proposed BSO-ARIMA-DL model exhibits a minimal error of ~10% minimal to the conventional ARIMA model.

Photonic deep residual time-delay reservoir computing

Time-delay reservoir computing (TDRC) represents a simplified variant of recurrent neural networks, employing a nonlinear node with a feedback mechanism to construct virtual nodes. The capabilities of TDRC can be enhanced by transitioning to a deep architecture. In this work, we propose a novel photonic deep residual TDRC (DR-TDRC) with augmented capabilities. The additional time delay added to the residual structure enables DR-TDRC superior to traditional deep structures across various benchmark tasks, especially in memory capability and almost an order of magnitude improvement in nonlinear channel equalization. Additionally, a specifically designed clipping algorithm is utilized to counteract the damage of redundant layers in deep structures, enabling the extension of the deep TDRC to dozens rather than just a few layers, with higher performance. We experimentally demonstrate the proof-of-concept with a 4-layer DR-TDRC containing 960 interrelated neurons (240 neurons per layer), based on four injection-locked distributed feedback lasers. We confirm the potential for scalable deep RC with elevated performance. Our results provide a feasible approach for expanding deep photonic computing to satisfy the boosting demand for artificial intelligence.

Real-Time Deepfake Video Detection Using Eye Movement Analysis with a Hybrid Deep Learning Approach

Deepfake technology uses artificial intelligence to create realistic but false audio, images, and videos. Deepfake technology poses a significant threat to the authenticity of visual content, particularly in live-stream scenarios where the immediacy of detection is crucial. Existing Deepfake detection approaches have limitations and challenges, prompting the need for more robust and accurate solutions. This research proposes an innovative approach: combining eye movement analysis with a hybrid deep learning model to address the need for real-time Deepfake detection. The proposed hybrid deep learning model integrates two deep neural network architectures, MesoNet4 and ResNet101, to leverage their respective architectures’ strengths for effective Deepfake classification. MesoNet4 is a lightweight CNN model designed explicitly to detect subtle manipulations in facial images. At the same time, ResNet101 handles complex visual data and robust feature extraction. Combining the localized feature learning of MesoNet4 with the deeper, more comprehensive feature representations of ResNet101, our robust hybrid model achieves enhanced performance in distinguishing between manipulated and authentic videos, which cannot be conducted with the naked eye or traditional methods. The model is evaluated on diverse datasets, including FaceForensics++, CelebV1, and CelebV2, demonstrating compelling accuracy results, with the hybrid model attaining an accuracy of 0.9873 on FaceForensics++, 0.9689 on CelebV1, and 0.9790 on CelebV2, showcasing its robustness and potential for real-world deployment in content integrity verification and video forensics applications.

Automatic removal of large blood vasculature for objective assessment of brain tumors using quantitative dynamic contrast-enhanced magnetic resonance imaging.

The presence of a normal large blood vessel (LBV) in a tumor region can impact the evaluation of quantitative dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) parameters and tumor classification. Hence, there is a need for automatic removal of LBVs from brain tissues including intratumoral regions for achieving an objective assessment of tumors. This retrospective study included 103 histopathologically confirmed brain tumor patients who underwent MRI, including DCE-MRI data acquisition. Quantitative DCE-MRI analysis was performed for computing various parameters such as wash-out slope (Slope-2), relative cerebral blood volume (rCBV), relative cerebral blood flow (rCBF), blood plasma volume fraction (Vp), and volume transfer constant (Ktrans). An approach based on data-clustering algorithm, morphological operations, and quantitative DCE-MRI maps was proposed for the segmentation of normal LBVs in brain tissues, including the tumor region. Here, three widely used data-clustering algorithms were evaluated on two types of quantitative maps: (a) Slope-2, and (b) a new proposed combination of rCBV and Slope-2 maps. Fluid-attenuated inversion recovery-MRI hyperintense lesionswere also automatically segmented using deep learning-based architecture. The accuracy of LBV segmentation was qualitatively assessed blindly by two experienced observers, and Likert scoring was also obtained from each individual and compared using Cohen's Kappa test, and multiple statistical features from quantitative DCE-MRI parameters were obtained in the segmented tumor. t-test and receiver operating characteristic (ROC) curve analysis were performed for comparing the effect of removal of LBVs on parameters as well as on tumor grading. k-means clustering exhibited better accuracy and computational efficiency. Tumors, in particular high-grade gliomas (HGGs), showed a high contrast compared with normal tissues (relative % difference = 18.5%) on quantitative maps after the removal of LBVs. Statistical features (95th percentile values) of all parameters in the tumor region showed a statistically significant difference (p< 0.05) between with and without LBV maps. Similar results were obtained for the ROC curve analysis for differentiation between low-grade gliomas and HGGs. Moreover, after the removal of LBVs, the rCBV, rCBF, and Vp maps show better visualization of tumor regions.

Using Machine Learning to Identify Diseases and Perform Sorting in Apple Fruit

Fruit diseases play a major role in global agriculture, leading to substantial crop losses and influencing food production and economic stability. In this age of Industry 4.0 the fruit sorting is an important part in the food processing wherein this work plays a vital role. In this study, a solution for the detection and classification of apple fruit diseases is proposed and experimentally validated. Deep learning models offer promise for automating disease identification using fruit images, but encounter obstacles such as therequirement for extensive training data, computational complexity, and the risk of overfitting. This study introduces an innovative convolutional neural network (CNN) architecture aimed at addressing these challenges by incorporating a reduced number of layers, thus alleviating computational burdens while maintaining performance. Additionally, augmentation techniques such as shift, shear, scaling, zoom, and flipping are employed to diversify the training set without additional image acquisition. Our CNN model is specifically trained to identify common apple crop diseases like Scab, Rot, and Blotch. Rigorous experimental evaluation demonstrates the effectiveness ofour model, achieving a remarkable classification accuracy of 95.37%. Significantly, our model demonstrates reduced storage requirements and faster execution times compared to existing deep CNN architectures, enabling deployment on handheld devices and resource-limited environments. While other CNN models may offer similar accuracy levels, our approach emphasizes efficiency and resource optimization, rendering it practical for real-world applications in agriculture. Furthermore, our CNN model exhibits resilience to environmental variations and imaging parameters, enhancing its applicability across diverse agricultural settings. By leveraging advanced machine learning techniques, the approach developed in this experimental work contributes to modernizing fruits and vegetables sorting operations in food processing, crop management practices thus promoting agricultural sustainability. The scalability and portability of our model make it suitable for deployment in both small-scale farms and large-scale agricultural operations.

Integrating multi-task and cost-sensitive learning for predicting mortality risk of chronic diseases in the elderly using real-world data

Background and ObjectiveReal-world data encompass population diversity, enabling insights into chronic disease mortality risk among the elderly. Deep learning excels on large datasets, offering promise for real-world data. However, current models focus on single diseases, neglecting comorbidities prevalent in patients. Moreover, mortality is infrequent compared to illness, causing extreme class imbalance that impedes reliable prediction. We aim to develop a deep learning framework that accurately forecasts mortality risk from real-world data by addressing comorbidities and class imbalance. MethodsWe integrated multi-task and cost-sensitive learning, developing an enhanced deep neural network architecture that extends multi-task learning to predict mortality risk across multiple chronic diseases. Each patient cohort with a chronic disease was assigned to a separate task, with shared lower-level parameters capturing inter-disease complexities through distinct top-level networks. Cost-sensitive functions were incorporated to ensure learning of positive class characteristics for each task and achieve accurate prediction of the risk of death from multiple chronic diseases. ResultsOur study covers 15 prevalent chronic diseases and is experimented with real-world data from 482,145 patients (including 9,516 deaths) in Shenzhen, China. The proposed model is compared with six models including three machine learning models: logistic regression, XGBoost, and CatBoost, and three state-of-the-art deep learning models: 1D-CNN, TabNet, and Saint. The experimental results show that, compared with the other compared algorithms, MTL-CSDNN has better prediction results on the test set (ACC=0.99, REC=0.99, PRAUC=0.97, MCC=0.98, G-means = 0.98). ConclusionsOur method provides valuable insights into leveraging real-world data for precise multi-disease mortality risk prediction, offering potential applications in optimizing chronic disease management, enhancing well-being, and reducing healthcare costs for the elderly population.

Multimodal explainability via latent shift applied to COVID-19 stratification

We are witnessing a widespread adoption of artificial intelligence in healthcare. However, most of the advancements in deep learning in this area consider only unimodal data, neglecting other modalities. Their multimodal interpretation necessary for supporting diagnosis, prognosis and treatment decisions. In this work we present a deep architecture, which jointly learns modality reconstructions and sample classifications using tabular and imaging data. The explanation of the decision taken is computed by applying a latent shift that, simulates a counterfactual prediction revealing the features of each modality that contribute the most to the decision and a quantitative score indicating the modality importance. We validate our approach in the context of COVID-19 pandemic using the AIforCOVID dataset, which contains multimodal data for the early identification of patients at risk of severe outcome. The results show that the proposed method provides meaningful explanations without degrading the classification performance.

A new multi-objective hyperparameter optimization algorithm for COVID-19 detection from x-ray images

The coronavirus occurred in Wuhan (China) first and it was declared a global pandemic. To detect coronavirus X-ray images can be used. Convolutional neural networks (CNNs) are used commonly to detect illness from images. There can be lots of different alternative deep CNN models or architectures. To find the best architecture, hyper-parameter optimization can be used. In this study, the problem is modeled as a multi-objective optimization (MOO) problem. Objective functions are multi-class cross entropy, error ratio, and complexity of the CNN network. For the best solutions to the objective functions, multi-objective hyper-parameter optimization is made by NSGA-III, NSGA-II, R-NSGA-II, SMS-EMOA, MOEA/D, and proposed Swarm Genetic Algorithms (SGA). SGA is a swarm-based algorithm with a cross-over process. All six algorithms are run and give Pareto optimal solution sets. When the figures obtained from the algorithms are analyzed and algorithm hypervolume values are compared, SGA outperforms the NSGA-III, NSGA-II, R-NSGA-II, SMS-EMOA, and MOEA/D algorithms. It can be concluded that SGA is better than others for multi-objective hyper-parameter optimization algorithms for COVID-19 detection from X-ray images. Also, a sensitivity analysis has been made to understand the effect of the number of the parameters of CNN on model success.

An intelligent predictive and optimized wastewater treatment plant.

The intelligent predictive and optimized wastewater treatment plant method represents a ground-breaking shift in how we manage wastewater. By capitalizing on data-driven predictive modeling, automation, and optimization strategies, it introduces a comprehensive framework designed to enhance the efficiency and sustainability of wastewater treatment operations. This methodology encompasses various essential phases, including data gathering and training, the integration of innovative computational models such as Chimp-based GoogLeNet (CbG), data processing, and performance prediction, all while fine-tuning operational parameters. The designed model is a hybrid of the Chimp optimization algorithm and GoogLeNet. The GoogLeNet is a type of deep convolutional architecture, and the Chimp optimization is one of the bio-inspired optimization models based on chimpanzee behavior. It optimizes the operational parameters, such as pH, dosage rate, effluent quality, and energy consumption, of the wastewater treatment plant, by fixing the optimal settings in the GoogLeNet. The designed model includes the process such as pre-processing and feature analysis for the effective prediction of the operation parameters and its optimization. Notably, this innovative approach provides several key advantages, including cost reduction in operations, improved environmental outcomes, and more effective resource management. Through continuous adaptation and refinement, this methodology not only optimizes wastewater treatment plant performance but also effectively tackles evolving environmental challenges while conserving resources. It represents a significant step forward in the quest for efficient and sustainable wastewater treatment practices. The RMSE, MAE, MAPE, and R2 scores for the suggested technique are 1.103, 0.233, 0.012, and 0.002. Also, the model has shown that power usage decreased to about 1.4%, while greenhouse gas emissions have significantly decreased to 0.12% than the existing techniques.

Enhanced deep learning model architecture for plant disease detection in Chilli plants

Enhanced deep learning model architecture for plant disease detection in Chilli plants

Semantic attention guided low-light image enhancement with multi-scale perception

Low-light environments often lead to complex degradation of captured images. However, most deep learning-based image enhancement methods for low-light conditions only learn a single-channel mapping relationship between the input image in low-light conditions and the desired image in normal light without considering semantic priors. This may cause the network to deviate from the original color of the region. In addition, deep network architectures are not suitable for low-light image recovery due to low pixel values. To address these issues, we propose a novel network called SAGNet. It consists of two branches:the main branch extracts global enhancement features at the level of the original image, and the other branch introduces semantic information through region-based feature learning and learns local enhancement features for semantic regions with multi-level perception to maintain color consistency. The extracted features are merged with the global enhancement features for semantic consistency and visualization. We also propose an unsupervised loss function to improve the network’s adaptability to general scenes and reduce the effect of sparse datasets. Extensive experiments and ablation studies show that SAGNet maintains color accuracy better in all cases and keeps natural luminance consistency across the semantic range.

Computer Vision Model Compression Techniques for Embedded Systems:A Survey

Deep neural networks have consistently represented the state of the art in most computer vision problems. In these scenarios, larger and more complex models have demonstrated superior performance to smaller architectures, especially when trained with plenty of representative data. With the recent adoption of Vision Transformer (ViT) based architectures and advanced Convolutional Neural Networks (CNNs), the total number of parameters of leading backbone architectures increased from 62M parameters in 2012 with AlexNet to 7B parameters in 2024 with AIM-7B. Consequently, deploying such deep architectures faces challenges in environments with processing and runtime constraints, particularly in embedded systems. This paper covers the main model compression techniques applied for computer vision tasks, enabling modern models to be used in embedded systems. We present the characteristics of compression subareas, compare different approaches, and discuss how to choose the best technique and expected variations when analyzing it on various embedded devices. We also share codes to assist researchers and new practitioners in overcoming initial implementation challenges for each subarea and present trends for Model Compression.

Multiscale segmentation net for segregating heterogeneous brain tumors: Gliomas on multimodal MR images

In this research, the 3D volumetric segmentation of heterogeneous brain tumors such as Gliomas- anaplastic astrocytoma, and Glioblastoma Multiforme (GBM) is performed to extract enhancing tumor (ET), whole tumor (WT), and tumor core (TC) regions using T1, T2, and FLAIR images. Therefore, a deep learning-based encoder-decoder architecture named “MS-SegNet” using 3D multi-scale convolutional layers is proposed. The proposed architecture employs multi-scale feature extraction (MS-FE) block the filter size 3 × 3 × 3 to extract confined information like tumor boundary and edges of necrotic part. The filter of size 5 × 5 × 5 focuses on varied features like shape, size, and location of tumor region with edema. The local and global features from different MR modalities are extracted for segmenting thin and meshed boundaries of heterogeneous tumors between anatomical sub-regions like peritumoral edema, enhancing tumor, and necrotic tumor core. The learning parameters on introducing the MS-FE block are reduced to 10 million, which is much less than other architectures like 3D-Unet which takes into consideration 27 million features leading to the consumption of less computational power. A customized loss function is also prophesied based on a combination of dice loss and focal loss along with metrics such as accuracy and Intersection over Union (IoU) i.e. the overlapping of ground truth mask and predicted value for addressing the class imbalance problem. For evaluating the efficacy of the proposed method, four evaluation metrics such as Dice Coefficient (DSC), Sensitivity, Specificity, and Hausdorff95 distance (H95) are employed for analyzing the model's overall performance. It is observed that the proposed MS-SegNet architecture achieved the DSC of 0.81, 0.91, and 0.83 on BraTS 2020; 0.86, 0.92, and 0.84 on BraTS 2021 for ET, WT, and TC respectively. The developed model is also tested on a real-time dataset collected from the Post Graduate Institute of Medical Education & Research (PGIMER), Chandigarh. The DSC of 0.79, 0.76, and 0.68 for ET, WT, and TC respectively on the real-time dataset. These findings show that deep learning models with enhanced feature extraction capabilities can be readily trained to attain high accuracy in segmenting heterogeneous brain tumors and hold promising results. In the future, other tumor datasets will be explored for the detection and treatment planning of brain tumors to check the effectiveness of the model in real-world healthcare environments.

Deep Learning Models for Abdominal CT Organ Segmentation in Children: Development and Validation in Internal and Heterogeneous Public Datasets.

BACKGROUND. Deep learning abdominal organ segmentation algorithms have shown excellent results in adults; validation in children is sparse. OBJECTIVE. The purpose of this article is to develop and validate deep learning models for liver, spleen, and pancreas segmentation on pediatric CT examinations. METHODS. This retrospective study developed and validated deep learning models for liver, spleen, and pancreas segmentation using 1731 CT examinations (1504 training, 221 testing), derived from three internal institutional pediatric (age ≤ 18 years) datasets (n = 483) and three public datasets comprising pediatric and adult examinations with various pathologies (n = 1248). Three deep learning model architectures (SegResNet, DynUNet, and SwinUNETR) from the Medical Open Network for Artificial Intelligence (MONAI) framework underwent training using native training (NT), relying solely on institutional datasets, and transfer learning (TL), incorporating pretraining on public datasets. For comparison, TotalSegmentator, a publicly available segmentation model, was applied to test data without further training. Segmentation performance was evaluated using mean Dice similarity coefficient (DSC), with manual segmentations as reference. RESULTS. For internal pediatric data, the DSC for TotalSegmentator, NT models, and TL models for normal liver was 0.953, 0.964-0.965, and 0.965-0.966, respectively; for normal spleen, 0.914, 0.942-0.945, and 0.937-0.945; for normal pancreas, 0.733, 0.774-0.785, and 0.775-0.786; and for pancreas with pancreatitis, 0.703, 0.590-0.640, and 0.667-0.711. For public pediatric data, the DSC for TotalSegmentator, NT models, and TL models for liver was 0.952, 0.871-0.908, and 0.941-0.946, respectively; for spleen, 0.905, 0.771-0.827, and 0.897-0.926; and for pancreas, 0.700, 0.577-0.648, and 0.693-0.736. For public primarily adult data, the DSC for TotalSegmentator, NT models, and TL models for liver was 0.991, 0.633-0.750, and 0.926-0.952, respectively; for spleen, 0.983, 0.569-0.604, and 0.923-0.947; and for pancreas, 0.909, 0.148-0.241, and 0.699-0.775. The DynUNet TL model was selected as the best-performing NT or TL model considering DSC values across organs and test datasets and was made available as an open-source MONAI bundle (https://github.com/cchmc-dll/pediatric_abdominal_segmentation_bundle.git). CONCLUSION. TL models trained on heterogeneous public datasets and fine-tuned using institutional pediatric data outperformed internal NT models and Total-Segmentator across internal and external pediatric test data. Segmentation performance was better in liver and spleen than in pancreas. CLINICAL IMPACT. The selected model may be used for various volumetry applications in pediatric imaging.

Interpretable wind power forecasting combining seasonal-trend representations learning with temporal fusion transformers architecture

Wind power forecasting plays a critical role in the efficient integration of wind energy into power systems. Accurate short-term WPF helps maintain grid stability, optimize power dispatch, and reduce operational costs. However, current deep learning-based WPF architectures rely on mapping spatial–temporal features into a single high-dimensional representation, compromising performance and interpretability. In this work, the interpretable seasonal-trend representation algorithm (ISTR) is proposed to learn a couple of disentangled latent representations that describe seasonal-trend of wind power time series. The ISTR is developed in a self-supervised manner with no need for prior knowledge, providing sufficient interpretable information while avoiding the feature entanglement problem. The ISTR is then integrated into the temporal fusion transformer (TFT) model through seasonal time-point embedding to perform accurate multi-horizon WPF with interpretable meanings. In the case study, ISTR-TFT outperforms the current state-of-the-art methods by achieving an 8.9% improvement in mean absolute percentage error (MAPE) compared to DeepAR and a 15.7% improvement in pinball score. The proposed ISTR-TFT also shows enhanced robustness to noise and missing data scenarios. The interpretability results generated by ISTR-TFT reveal the importance of various input features and the learned temporal patterns, offering valuable insights for decision-making in power system operations.

XAI-TRIS: non-linear image benchmarks to quantify false positive post-hoc attribution of feature importance

The field of ‘explainable’ artificial intelligence (XAI) has produced highly acclaimed methods that seek to make the decisions of complex machine learning (ML) methods ‘understandable’ to humans, for example by attributing ‘importance’ scores to input features. Yet, a lack of formal underpinning leaves it unclear as to what conclusions can safely be drawn from the results of a given XAI method and has also so far hindered the theoretical verification and empirical validation of XAI methods. This means that challenging non-linear problems, typically solved by deep neural networks, presently lack appropriate remedies. Here, we craft benchmark datasets for one linear and three different non-linear classification scenarios, in which the important class-conditional features are known by design, serving as ground truth explanations. Using novel quantitative metrics, we benchmark the explanation performance of a wide set of XAI methods across three deep learning model architectures. We show that popular XAI methods are often unable to significantly outperform random performance baselines and edge detection methods, attributing false-positive importance to features with no statistical relationship to the prediction target rather than truly important features. Moreover, we demonstrate that explanations derived from different model architectures can be vastly different; thus, prone to misinterpretation even under controlled conditions.

Extra-abdominal trocar and instrument detection for enhanced surgical workflow understanding.

Video-based intra-abdominal instrument tracking for laparoscopic surgeries is a common research area. However, the tracking can only be done with instruments that are actually visible in the laparoscopic image. By using extra-abdominal cameras to detect trocars and classify their occupancy state, additional information about the instrument location, whether an instrument is still in the abdomen or not, can be obtained. This can enhance laparoscopic workflow understanding and enrich already existing intra-abdominal solutions. A data set of four laparoscopic surgeries recorded with two time-synchronized extra-abdominal 2D cameras was generated. The preprocessed and annotated data were used to train a deep learning-based network architecture consisting of a trocar detection, a centroid tracker and a temporal model to provide the occupancy state of all trocars during the surgery. The trocar detection model achieves an F1 score of . The prediction of the occupancy state yields an F1 score of , providing a first step towards enhanced surgical workflow understanding. The current method shows promising results for the extra-abdominal tracking of trocars and their occupancy state. Future advancements include the enlargement of the data set and incorporation of intra-abdominal imaging to facilitate accurate assignment of instruments to trocars.

EMPLOYEE PAYROLL DATA ANALYSIS USING MACHINE LEARNING

Every company is using machine learning to predict their employees' working styles and competence by estimating how long it will take them to complete a task. The effectiveness of various classification algorithms for predicting employee salary classes has been the subject of numerous recent studies. Salary prediction appears to be a challenging endeavor from the perspective of machine learning because of the low sample size, relatively high dimensionality, and presence of noise. Deeper architectures are required to address this and locate more useful features. Also, more information investigation and information handling can be commonsense to make the forecast model go past the relationship and accuracy guidelines by highlight extraction methods.

A comprehensive study of auto-encoders for anomaly detection: Efficiency and trade-offs

Unsupervised anomaly detection (UAD) is a diverse research area explored across various application domains. Over time, numerous anomaly detection techniques, including clustering, generative, and variational inference-based methods, are developed to address specific drawbacks and advance state-of-the-art techniques. Deep learning and generative models recently played a significant role in identifying unique challenges and devising advanced approaches. Auto-encoders (AEs) represent one such powerful technique that combines generative and probabilistic variational modeling with deep architecture. Auto-Encoder aims to learn the underlying data distribution to generate consequential sample data. This concept of data generation and the adoption of generative modeling have emerged in extensive research and variations in Auto-Encoder design, particularly in unsupervised representation learning. This study systematically reviews 11 Auto-Encoder architectures categorized into three groups, aiming to differentiate their reconstruction ability, sample generation, latent space visualization, and accuracy in classifying anomalous data using the Fashion-MNIST (FMNIST) and MNIST datasets. Additionally, we closely observed the reproducibility scope under different training parameters. We conducted reproducibility experiments utilizing similar model setups and hyperparameters and attempted to generate comparative results to address the scope of improvements for each Auto-Encoder. We conclude this study by analyzing the experimental results, which guide us in identifying the efficiency and trade-offs among auto-encoders, providing valuable insights into their performance and applicability in unsupervised anomaly detection techniques.

Element-wise parallel deep learning for structural distributed damage diagnosis by leveraging physical properties of long-gauge static strain transmissibility under moving loads

The Transmissibility Function (TF) has gained considerable interest in structural damage detection because of its relatively high sensitivity to damage and robustness to excitation. This study proposed a distributed damage diagnosis method for beam-like structures based on a long-gauge static strain TF defined as the ratio of the Fourier transform of static strain response under moving loads from a target long-gauge sensor to that from the reference sensor. It has been discovered that each long-gauge static strain TF is independent of moving loads, making it an ideal damage indicator. It exhibits direct proportionality to the ratio of bending stiffness between the two measured zones covered by the target and reference long-gauge strain sensors while being unaffected by the stiffness of any other zones not covered by these two sensors. Moreover, the TF at zero frequency was proved to be equivalent to the ratio of static strain time history areas, relying solely on the stiffness of the covered zones. Leveraged by the physical properties of long-gauge static strain TF, an element-wise parallel deep neural network architecture was developed to decompose damage diagnosis into independent sub-tasks on the element level, utilizing a variational autoencoder (VAE) in the Bayesian inference framework for extracting features from the long-gauge strain TF and a regressor for mapping the features to elemental stiffness reduction. The divide-and-conquer strategy and element-wise parallel learning architecture allow for a significant reduction in the number of labeled training samples generated based on the updated numerical baseline model as it only requires simulated scenarios involving a single damage element. The efficiency and robustness of the proposed method were demonstrated through a numerical simply supported beam and a laboratory experiment on a two-span bridge.