Leveraging Deep Learning Research Articles

AstroSer: Leveraging Deep Learning for Efficient Content-based Retrieval in Massive Solar-observation Images

Rapid and proficient data retrieval is an essential component of modern astronomical research. In this paper, we address the challenge of retrieving astronomical image content by leveraging state-of-the-art deep learning techniques. We have designed a retrieval model, HybridVR, that integrates the capabilities of the deep learning models ResNet50 and VGG16 and have used it to extract key features of solar activity and solar environmental characteristics from observed images. This model enables efficient image matching and allows for content-based image retrieval (CBIR). Experimental results demonstrate that the model can achieve up to 98% similarity during CBIR while exhibiting adaptability and scalability. Our work has implications for astronomical research, data management, and education, and it can contribute to optimizing the utilization of astronomical image data. It also serves as a useful example of the application of deep learning technology in the field of astronomy.

Cancer Diagnosis through Contour Visualization of Gene Expression Leveraging Deep Learning Techniques.

Prompt diagnostics and appropriate cancer therapy necessitate the use of gene expression databases. The integration of analytical methods can enhance detection precision by capturing intricate patterns and subtle connections in the data. This study proposes a diagnostic-integrated approach combining Empirical Bayes Harmonization (EBS), Jensen-Shannon Divergence (JSD), deep learning, and contour mathematics for cancer detection using gene expression data. EBS preprocesses the gene expression data, while JSD measures the distributional differences between cancerous and non-cancerous samples, providing invaluable insights into gene expression patterns. Deep learning (DL) models are employed for automatic deep feature extraction and to discern complex patterns from the data. Contour mathematics is applied to visualize decision boundaries and regions in the high-dimensional feature space. JSD imparts significant information to the deep learning model, directing it to concentrate on pertinent features associated with cancerous samples. Contour visualization elucidates the model's decision-making process, bolstering interpretability. The amalgamation of JSD, deep learning, and contour mathematics in gene expression dataset analysis diagnostics presents a promising pathway for precise cancer detection. This method taps into the prowess of deep learning for feature extraction while employing JSD to pinpoint distributional differences and contour mathematics for visual elucidation. The outcomes underscore its potential as a formidable instrument for cancer detection, furnishing crucial insights for timely diagnostics and tailor-made treatment strategies.

Abstract 14595: Leveraging Deep Learning to Elucidate Subclinical Aortic Stenosis Risk

Background and Aims: Aortic stenosis (AS) is associated with a high global burden of morbidity. Current guidelines do not fully define normal aortic valve (AV) function or mild AS. We aimed to study normal variation in AV function and address current gaps in knowledge. Methods: We developed a deep learning model to study velocity-encoded magnetic resonance imaging (MRI) in 47,223 UK Biobank participants. We calculated aortic valve area (AVA), peak velocity (PV), and mean gradient (MG). Results: In healthy individuals, we found an annual decrement of 0.03 cm2 in AVA. Normal variation in AV function was predictive of future AV surgery (HR 2.4 per SD increase in PV, P2.6E-110), heart failure (HR 1.23 per SD, P=2.4E-07), and coronary disease (HR 1.23 per SD, P=2.0E-14). Risk was still elevated when excluding participants with moderate-to-severe AS. We determined thresholds for abnormal AV function as being >95% for all age groups in 31,909 healthy people. This established the boundary between normal variation and mild AS as any of: PV >1.65m/s, MG >4.9mmHg, and AVA <2.1cm2 (men) or <1.7cms2 (women). A total of 3,288 participants (7.0%) met the definition of mild AS which was associated with high risk for AV surgery (HR 57.2, P=4.4E-29). Over a mean 3.1±2.0 years of follow-up, the incidence of AV surgery was 0.02% in those with normal AV function and nearly 20% among participants with MG ≥10mmHg. Greater levels of ApoB, triglycerides, Lp(a), and inflammatory markers were associated with higher AV gradients. Conclusion: In this study, we report the largest analysis of AV function and cardiovascular disease in the general population with MRI using deep learning. We propose thresholds to define the lower bounds of mild AS and demonstrate a step change in cardiovascular risk between normal AV function and mild AS. Atherogenic and inflammatory biomarkers were associated with increased AV gradients. These findings may inform earlier surveillance and efforts to prevent complications of AS.

From Prediction to Prevention: Leveraging Deep Learning in Traffic Accident Prediction Systems

We propose a novel system leveraging deep learning-based methods to predict urban traffic accidents and estimate their severity. The major challenge is the data imbalance problem in traffic accident prediction. The problem is caused by numerous zero values in the dataset due to the rarity of traffic accidents. To address the issue, we propose a grid-clustered feature map with the ideas of grids and cells. To predict the occurrence of accidents in the grid, we introduce an accident detector that combines the power of a Convolutional Neural Network (CNN) with a Deep Neural Network (DNN). Then, hierarchical DNNs are supposed to be an accident risk classifier to estimate the risk of each cell in the accident-occurrence grid. The proposed system can effectively reduce instances with no traffic accidents. Furthermore, we introduce the concept of the Accident Risk Index (ARI) to better represent the severity of risk at each cell. Also, we consider all the explanatory variables, such as dangerous driving behaviors, traffic mobility, and safety facility information, that can be related to traffic accidents. To improve the prediction accuracy, we further take into consideration all the explanatory variables, such as dangerous driving behaviors, traffic mobility, and safety facility information, that can be related to traffic accidents. In the experiment, we highlight the benefits of our method for urban traffic accident management by significantly improving model performance compared to the baselines. The feasibility and applicability of the proposed system are validated in the data of Daejeon City, Republic of Korea. The proposed prediction system can dynamically advise and recommend commuters, traffic management systems, and city planners on alternatives, optimizations, and interventions.

A Mathematical Model for Customer Segmentation Leveraging Deep Learning, Explainable AI, and RFM Analysis in Targeted Marketing

In the evolving landscape of targeted marketing, integrating deep learning (DL) and explainable AI (XAI) offers a promising avenue for enhanced customer segmentation. This paper introduces a groundbreaking approach, DeepLimeSeg, which synergizes DL methodologies with Lime-based Explainability to segment customers effectively. The approach employs a comprehensive mathematical model to harness demographic data, behavioral patterns, and purchase histories, categorizing customers into distinct clusters aligned with their preferences and needs. A pivotal component of this research is the mathematical underpinning of the DeepLimeSeg approach. The Lime-based Explainability module ensures that the segmentation results are accurate and interpretable. The mathematical rigor facilitates businesses tailoring their marketing strategies with precision, optimizing sales outcomes. To validate the efficacy of DeepLimeSeg, we employed two real-world datasets: Mall-Customer Segmentation Data and an E-Commerce dataset. A comparative analysis between DeepLimeSeg and the traditional Recency, Frequency, and Monetary (RFM) analysis is presented. The RFM analysis, grounded in its mathematical modeling, segments customers based on purchase recency, frequency, and monetary value. Our preprocessing involved computing RFM scores for each customer, followed by K-means clustering to delineate customer segments. Empirical results underscored the superiority of DeepLimeSeg over other models in terms of MSE, MAE, and R2 metrics. Specifically, the model registered an MSE of 0.9412, indicative of its robust predictive accuracy concerning the spending score. The MAE value stood at 0.9874, signifying minimal deviation from actual values. This paper accentuates the importance of mathematical modeling in enhancing customer segmentation. The DeepLimeSeg approach, with its mathematical foundation and explainable AI integration, paves the way for businesses to make informed, data-driven marketing decisions.

AutoKeras for Fake News Identification in Arabic: Leveraging Deep Learning with an Extensive Dataset

Social media and the World Wide Web have led to a worrying rise in spreading false information, which presents a significant worldwide issue. Identifying and preventing false information is crucial in promoting an informed and knowledgeable society. The identification of false information, specifically in the Arabic dialect, presents inherent difficulties due to its diverse characteristics and linguistic intricacies. This study implements AutoKeras, a deep learning-based machine learning framework. Using advanced optimization techniques, the neural network architecture search, hyperparameter adjustments, and model selection can all be automated in AutoKeras. Therefore, it is suitable for our fake news detection task. The methodology employs proficient deep learning algorithms and natural language processing methods to acquire distinct characteristics that enable accurate differentiation between genuine and fake news. The present study uses various sources, including news websites, social media platforms, and blogs, to construct the dataset. The AutoKeras-based approach is superior to multiple state-of-the-art approaches to detecting fabricated news in Arabic, as evidenced by the experimental results. The suggested method outperforms 93.2% accuracy in identifying fake news, demonstrating its superior efficacy. This demonstrates the great promise of the deep learning-based Auto model for detecting false information.

A deep learning framework for solving forward and inverse problems of power-law fluids

We for the first time leverage deep learning approaches to solve forward and inverse problems of two-dimensional laminar flows for power-law fluids. We propose a deep-learning framework, called Power-Law-Fluid-Net (PL-Net). We develop a surrogate model to solve the forward problems of the power-law fluids, and solve the inverse problems utilizing only a small set of measurement data under the assumption that boundary conditions (BCs) can be partially known. In the design of the methods, we incorporate the hard boundary condition constraints to accelerate the iteration of stochastic gradient descent methods for minimizing loss functions. For the forward problems, by incorporating the constitutive parameters into the input variables of neural networks, the PL-Net serves as a surrogate model for simulating the pressure-driven flows inside pipes having cross sections of varying shapes. We investigate the influences of the BC type, activation function type, and number of collocation points on the accuracy of numerical solutions. For the inverse problems, the PL-Net infers the physical quantities or constitutive parameters from a small number of measurements of flow field variables. The BCs of the inverse problems can even be partially known. We demonstrate the effects of BC type, number of sensors, and noise level on accuracy of inferred quantities. Computational examples indicate the high accuracy of the PL-Net in tackling both the forward and inverse problems of the power-law fluids.

P06-41: Leveraging Deep Learning to Optimize Initial Condition Determination in Next-Generation Systems Biology Models

P06-41: Leveraging Deep Learning to Optimize Initial Condition Determination in Next-Generation Systems Biology Models

NLP for Qualitative Analysis: Leveraging Deep Learning to Capture Contextual Meaning

NLP for Qualitative Analysis: Leveraging Deep Learning to Capture Contextual Meaning

Pre-Trained Embeddings for Entity Resolution: An Experimental Analysis

Many recent works on Entity Resolution (ER) leverage Deep Learning techniques involving language models to improve effectiveness. This is applied to both main steps of ER, i.e., blocking and matching. Several pre-trained embeddings have been tested, with the most popular ones being fastText and variants of the BERT model. However, there is no detailed analysis of their pros and cons. To cover this gap, we perform a thorough experimental analysis of 12 popular language models over 17 established benchmark datasets. First, we assess their vectorization overhead for converting all input entities into dense embeddings vectors. Second, we investigate their blocking performance, performing a detailed scalability analysis, and comparing them with the state-of-the-art deep learning-based blocking method. Third, we conclude with their relative performance for both supervised and unsupervised matching. Our experimental results provide novel insights into the strengths and weaknesses of the main language models, facilitating researchers and practitioners to select the most suitable ones in practice.

Deep-Learning-Based COVID-19 Detection: Challenges and Future Directions

Coronavirus disease 2019 (COVID-19) is an ecumenical pandemic that has affected the whole world drastically by raising a global calamitous situation. Owing to this pernicious disease, millions of people have lost their lives. The scientists are still far from knowing how to tackle the coronavirus due to its multiple mutations found around the globe. The standard testing technique called polymerase chain reaction for the clinical diagnosis of COVID-19 is expensive and time consuming. However, to assist specialists and radiologists in COVID-19 detection and diagnosis, deep learning plays an important role. Many research efforts have been done that leverage deep learning techniques and technologies for the identification or categorization of COVID-19-positive patients, and these techniques are proved to be a powerful tool that can automatically detect or diagnose COVID-19 cases. In this article, we identify significant challenges regarding deep-learning-based systems and techniques that use different medical imaging modalities, including cough and breadth, chest X-ray, and computed tomography, to combat COVID-19 outbreak. We also pinpoint important research questions for each category of challenges.

Deepfake Detection on Social Media: Leveraging Deep Learning and FastText Embeddings for Identifying Machine-Generated Tweets

Deepfake Detection on Social Media: Leveraging Deep Learning and FastText Embeddings for Identifying Machine-Generated Tweets

Unveiling the Veiled: Leveraging Deep Learning and Network Analysis for De-Anonymization in Social Networks

Online anonymity allows individuals to safeguard their personal information. It provides the freedom for anyone to express themselves without being concerned about censorship, discrimination, or retaliation. This pseudonym also provides opportunities for individuals to promote open discourse and diverse viewpoints. However, in today's digital age, this online anonymity has become a growing concern. Although the safeguarding of people’s rights, the advancement of free speech, and the development of a more diverse and democratic online community all depend heavily on the anonymity of online users, there are risks as well, such as cyberbullying, harassment, and the propagation of false information. Because of their anonymity, predators, groomers, and other unscrupulous individuals may be able to take advantage of vulnerable individuals, especially children and adolescents. In order to trick victims into hazardous or abusive situations, adversaries can hide their identities. Deep learning and network analysis can be used to reveal the true identities of anonymous social media users in order to combat this. Deep learning algorithms are capable of analyzing a wide range of social network data, including user behavior, relationships, and content, in order to find patterns and correlations that can lead to the true identities of users that go anonymous. The proposed system includes examining network topology and group dynamics to identify potential anomalies and connections that could lead to de-anonymization. This paper proposes a novel module for de-anonymization integrating deep learning and network analysis in social networks.

Deep Learning Based Walking Tasks Classification in Older Adults Using fNIRS.

Decline in gait features is common in older adults and an indicator of increased risk of disability, morbidity, and mortality. Under dual task walking (DTW) conditions, further degradation in the performance of both the gait and the secondary cognitive task were found in older adults which were significantly correlated to falls history. Cortical control of gait, specifically in the pre-frontal cortex (PFC) as measured by functional near infrared spectroscopy (fNIRS), during DTW in older adults has recently been studied. However, the automatic classification of differences in cognitive activations under single and dual task gait conditions has not been extensively studied yet. In this paper, by considering single task walking (STW) as a lower attentional walking state and DTW as a higher attentional walking state, we aimed to formulate this as an automatic detection of low and high attentional walking states and leverage deep learning methods to perform their classification. We conduct analysis on the data samples which reveals the characteristics on the difference between HbO2 and Hb values that are subsequently used as additional features. We perform feature engineering to formulate the fNIRS features as a 3-channel image and apply various image processing techniques for data augmentation to enhance the performance of deep learning models. Experimental results show that pre-trained deep learning models that are fine-tuned using the collected fNIRS dataset together with gender and cognitive status information can achieve around 81% classification accuracy which is about 10% higher than the traditional machine learning algorithms. We present additional sensitivity metrics such as confusion matrix, precision and F1 score, as well as accuracy on two-way classification between condition pairings. We further performed an extensive ablation study to evaluate factors such as the voxel locations, channels of input images, zero-paddings and pre-training of deep learning model on their contribution or impact to the classification task. Results showed that using pre-trained model, all the voxel locations, and HbO2 - Hb as the third channel of the input image can achieve the best classification accuracy.

An inception V3 approach for malware classification using machine learning and transfer learning

Malware instances have been extremely used for illegitimate purposes, and new variants of malware are observed every day. Machine learning in network security is one of the prime areas of research today because of its performance and has shown tremendous growth in the last decade. In this paper, we formulate the malware signature as a 2D image representation and leverage deep learning approaches to characterize the signature of malware contained in BIG15 dataset across nine classes. The current research compares the performance of various machine learning and deep learning technologies towards malware classification such as Logistic Regression (LR), Artificial Neural Network (ANN), Convolutional Neural Network (CNN), transfer learning on CNN and Long Short Term Memory (LSTM). The transfer learning approach using InceptionV3 resulted in a good performance with respect to the compared models like LSTM with a classification accuracy of 98.76% on the test dataset and 99.6% on the train dataset.

Safety Implications of Variability in Autonomous Driving Assist Alerting

Advanced Driving Assist Systems (ADAS) are on the rise in new cars, including versions that embed artificial intelligence in computer vision systems that leverage deep learning algorithms. Because these systems, at the present time, cannot operate in all operational driving domains, they employ some type of driver monitoring system for assessing driver attention, so that drivers can effectively take control if and when an ADAS system can no longer control the car. To determine the reliability of a driver alerting system when linked to autonomy that leverages deep learning, a set of increasingly complex tests were conducted on three Tesla Model 3 vehicles. Tests were conducted on a highway and a closed test track to test road departure and construction zone detection capabilities. Results revealed significant between- and within-vehicle variation on a number of metrics related to driver monitoring, alerting, and safe operation of the underlying autonomy. In some cases, cars performed better than expected but all cars exhibited both inconsistent and unsafe behaviors as well as poor driver alerting. These results highlight that a post-deployment regulatory process is ill-equipped to flag significant issues in vehicles with embedded artificial intelligence.

Incorporating deep learning into capacitive images for smartphone user authentication

We incorporate deep learning techniques into capacitive images of body parts (ear, four fingers, and thumb) to improve the performance of user authentication in smartphones. Use of a capacitive touchscreen as an image sensor has several advantages, such as it is less sensitive to poor illumination conditions, occlusions, and pose variations. Also, it does not need an additional hardware like iris or fingerprint scanner. Use of capacitive images for user authentication is not new. However, the performance, specially, false reject rates (FRRs) of the state-of-the-art capacitive image-based systems are poor. In this paper, we focus on improving the performance and leverage deep learning. Deep learning techniques demonstrated spectacular performance in previous physical biometrics-based research. However, to our knowledge, effectiveness of deep learning is still unexplored in capacitive touchscreen-based user authentication. In order to bridge this research gap, we devise a multi-modal deep learning model, namely UASNet, and compare its performance with a large set of uni- and multi-modal baselines. Using the UASNet, we achieve an accuracy of 99.77%, an EER of 0.48%, and an FRR of 1.19% at FAR of 0.06%.

Analysis Of PIV Images Of Transonic Buffet Flow By Recurrent Deep Learning Based Optical Flow Prediction

In the past few years, several algorithms have been proposed that leverage deep learning techniques within the data analysis workflow of particle-image velocimetry (PIV) experiments. This emerging body of work has shown that deep learning has the potential to match or outperform state-of-the-art classical algorithms in terms of efficiency, accuracy, and spatial resolution. Due to the significant relevance of PIV experiments in the broader fluid mechanics community, progress in PIV processing approaches based on state-of-the-art machine learning tools has a major impact across a range of problems in applied physics and engineering where velocity components of flow fields need to be determined. In contrast to existing methods, these approaches are general, near-automated and yield per-pixel flow estimates. Primary work on deep learning for PIV is promising, but important questions concerning the application to challenging engineering problems remain open and more scientific analyses are necessary to substantiate the general confidence amongst practitioners in such neural network based tools. This motivates us to employ a novel deep learning based PIV approach called RAFT-PIV to a challenging application of a recent research subject. That is, we use our neural optical flow approach to evaluate Background-Oriented Schlieren (BOS) and PIV images obtained by measurements of the transonic flow around a supercritical DRA-2303 profile and demonstrate that it can be used as a direct one-to-one replacement for standard cross-correlation based counterparts.

Combining deep ensemble learning and explanation for intelligent ticket management

Intelligent Ticket Management Systems, equipped with automated ticket classification tools, are an advanced solution for handling customer-support activities. Some recent approaches to ticket classification leverage Deep Learning (DL) methods, in place of traditional ones using standard Machine Learning and feature engineering techniques. However, two challenging objectives should be addressed when applying DL methods to real-life contexts: (i) curbing the risk of having an overfitting model that hinges on spurious ticket features, and (ii) trying to explain the ticket classifications returned by such black-box models. In this work, we propose a comprehensive ticket classification framework, which relies on training a novel kind of ensemble of deep classifiers, and on providing AI-based interpretation methods to help both the operator in recognizing misclassification errors and the analyst in improving and fine-tuning the model.Tests on real data confirmed the accuracy of the classifications returned by the framework, and the practical value of their associated explanations.

A Hierarchical Approach toward Prediction of Human Biological Age from Masked Facial Image Leveraging Deep Learning Techniques

The lifestyle of humans has changed noticeably since the contagious COVID-19 disease struck globally. People should wear a face mask as a protective measure to curb the spread of the contagious disease. Consequently, real-world applications (i.e., electronic customer relationship management) dealing with human ages extracted from face images must migrate to a robust system proficient to estimate the age of a person wearing a face mask. In this paper, we proposed a hierarchical age estimation model from masked facial images in a group-to-specific manner rather than a single regression model because age progression across different age groups is quite dissimilar. Our intention was to squeeze the feature space among limited age classes so that the model could fairly discern age. We generated a synthetic masked face image dataset over the IMDB-WIKI face image dataset to train and validate our proposed model due to the absence of a benchmark masked face image dataset with real age annotations. We somewhat mitigated the data sparsity problem of the large public IMDB-WIKI dataset using off-the-shelf down-sampling and up-sampling techniques as required. The age estimation task was fully modeled like a deep classification problem, and expected ages were formulated from SoftMax probabilities. We performed a classification task by deploying multiple low-memory and higher-accuracy-based convolutional neural networks (CNNs). Our proposed hierarchical framework demonstrated marginal improvement in terms of mean absolute error (MAE) compared to the one-off model approach for masked face real age estimation. Moreover, this research is perhaps the maiden attempt to estimate the real age of a person from his/her masked face image.