Use Of Deep Learning Models Research Articles

Bridge inspection component registration for damage evolution

There have been great advances in bridge inspection damage detection involving the use of deep learning models. However, automated detection models currently fall short of giving an inspector an understanding of how the damage has progressed from one inspection to the next. The rate-of-change of the damage is a critical piece of information used by engineers to determine appropriate maintenance and rehabilitation actions to prevent structural failures. We propose a simple methodology for registering two bridge inspection videos or still images, collected at different stages of deterioration, so that trained model predictions may be directly measured and damage progression compared. The changes may be documented and presented to the inspector so that they may quickly evaluate key interest regions in the inspection video or image. Three approaches referred to as rigid, deformable, and hybrid image registration methods were experimentally tested and evaluated based on their ability to preserve the geometric characteristics of the referenced image. It was found in all experiments that the rigid, homography-based transformations performed the best for this application over a state-of-the-art deformable registration method, RANSAC-Flow.

A Deep Learning-Based Action Recommendation Model for Cryptocurrency Profit Maximization

Research on the prediction of cryptocurrency prices has been actively conducted, as cryptocurrencies have attracted considerable attention. Recently, researchers have aimed to improve the performance of price prediction methods by applying deep learning-based models. However, most studies have focused on predicting cryptocurrency prices for the following day. Therefore, clients are inconvenienced by the necessity of rapidly making complex decisions on actions that support maximizing their profit, such as “Sell”, “Buy”, and “Wait”. Furthermore, very few studies have explored the use of deep learning models to make recommendations for these actions, and the performance of such models remains low. Therefore, to solve these problems, we propose a deep learning model and three input features: sellProfit, buyProfit, and maxProfit. Through these concepts, clients are provided with criteria on which action would be most beneficial at a given current time. These criteria can be used as decision-making indices to facilitate profit maximization. To verify the effectiveness of the proposed method, daily price data of six representative cryptocurrencies were used to conduct an experiment. The results confirm that the proposed model showed approximately 13% to 21% improvement over existing methods and is statistically significant.

Validation of Soft Labels in Developing Deep Learning Algorithms for Detecting Lesions of Myopic Maculopathy From Optical Coherence Tomographic Images.

It is common for physicians to be uncertain when examining some images. Models trained with human uncertainty could be a help for physicians in diagnosing pathologic myopia. This is a hospital-based study that included 9176 images from 1327 patients that were collected between October 2015 and March 2019. All collected images were graded by 21 myopia specialists according to the presence of myopic neovascularization (MNV), myopic traction maculopathy (MTM), and dome-shaped macula (DSM). Hard labels were made by the rule of major wins, while soft labels were possibilities calculated by whole grading results from the different graders. The area under the curve (AUC) of the receiver operating characteristics curve, the area under precision-recall (AUPR) curve, F-score, and least square errors were used to evaluate the performance of the models. The AUC values of models trained by soft labels in MNV, MTM, and DSM models were 0.985, 0.946, and 0.978; and the AUPR values were 0.908, 0.876, and 0.653 respectively. However, 0.56% of MNV "negative" cases were answered as "positive" with high certainty by the hard label model, whereas no case was graded with extreme errors by the soft label model. The same results were found for the MTM (0.95% vs none) and DSM (0.43% vs 0.09%) models. The predicted possibilities from the models trained by soft labels were close to the results made by myopia specialists. These findings could inspire the novel use of deep learning models in the medical field.

An Empirical Study of the Effectiveness of an Ensemble of Stand-alone Sentiment Detection Tools for Software Engineering Datasets

Sentiment analysis in software engineering (SE) has shown promise to analyze and support diverse development activities. Recently, several tools are proposed to detect sentiments in software artifacts. While the tools improve accuracy over off-the-shelf tools, recent research shows that their performance could still be unsatisfactory. A more accurate sentiment detector for SE can help reduce noise in analysis of software scenarios where sentiment analysis is required. Recently, combinations, i.e., hybrids of stand-alone classifiers are found to offer better performance than the stand-alone classifiers for fault detection. However, we are aware of no such approach for sentiment detection for software artifacts. We report the results of an empirical study that we conducted to determine the feasibility of developing an ensemble engine by combining the polarity labels of stand-alone SE-specific sentiment detectors. Our study has two phases. In the first phase, we pick five SE-specific sentiment detection tools from two recently published papers by Lin et al. [ 29 , 30 ], who first reported negative results with stand alone sentiment detectors and then proposed an improved SE-specific sentiment detector, POME [ 29 ]. We report the study results on 17,581 units (sentences/documents) coming from six currently available sentiment benchmarks for software engineering. We find that the existing tools can be complementary to each other in 85-95% of the cases, i.e., one is wrong but another is right. However, a majority voting-based ensemble of those tools fails to improve the accuracy of sentiment detection. We develop Sentisead, a supervised tool by combining the polarity labels and bag of words as features. Sentisead improves the performance (F1-score) of the individual tools by 4% (over Senti4SD [ 5 ]) – 100% (over POME [ 29 ]). The initial development of Sentisead occurred before we observed the use of deep learning models for SE-specific sentiment detection. In particular, recent papers show the superiority of advanced language-based pre-trained transformer models (PTM) over rule-based and shallow learning models. Consequently, in a second phase, we compare and improve Sentisead infrastructure using the PTMs. We find that a Sentisead infrastructure with RoBERTa as the ensemble of the five stand-alone rule-based and shallow learning SE-specific tools from Lin et al. [ 29 , 30 ] offers the best F1-score of 0.805 across the six datasets, while a stand-alone RoBERTa shows an F1-score of 0.801.

Deep Spatial-Temporal 2-D CNN-BLSTM Model for Ultrashort-Term LiDAR-Assisted Wind Turbine's Power and Fatigue Load Forecasting

Optimizing wind turbine performance is still a challenge due to the dynamic interactions between the spatially temporally stochastic wind fields and the wind turbine as a complex mechanical system. Recent cost reduction of remote sensing wind measurement technologies, such as light detection and ranging (LiDAR), has opened a new research area on the use of deep learning models for predicting wind turbine's responses. In this article, a LiDAR-aided deep learning model is presented to learn the powerful spatial-temporal characteristics from the input wind fields. In the proposed method, the combination of 2-D convolutional neural networks (CNNs) and bidirectional long short-term memory (BLSTM) units is used to capture high levels of abstractions in wind fields concurrently, and thus, forecasting wind output power and fatigue load as two representatives of wind turbine responses. The LiDAR wind preview information is used as the 2-D-images of wind fields for the CNN. Moreover, the BLSTM is incorporated with the proposed CNN to improve the forecasting accuracy further and learn deep temporal features. The aero-elastic 5-MW reference wind turbine of National Renewable Energy Laboratory (NREL) is used to evaluate the performance of proposed model compared to the state-of-the-art deep-learning-based architectures in the recent literature.

Synthetic Data Generation Using GAN for RUL Prediction of Supercapacitors

The remaining useful life (RUL) prediction of supercapacitors is an important part of supercapacitors management system. To accurately predict the RUL of supercapacitor, a large amount of capacity data is required which can be difficult to acquire due to privacy restrictions and limited access. Previous works have employed the use of deep learning models to synthetically generate data. However, a prerequisite ensuring the success of these models depends on their ability to preserve the temporal dynamics of the data. This paper presents a generative adversarial network (GAN) for synthetic data generation and a long short-term memory (LSTM) network for accurate RUL prediction. Firstly, the GAN model is employed for synthetic data generation and LSTM for RUL prediction. We show that the GAN model is capable of preserving the temporal dynamics of the original data and also prove that the generated data can be used to accurately carry out RUL prediction. Our proposed GAN model was able to achieve an accuracy of 85% after 500 epochs. The performance of the generated data set with the LSTM model achieved an RMSE of 0.29. The overall results show that synthetic data can be used to achieve excellent performance for RUL prediction.

Deploy Cotton Plant Disease Prediction Application using CNN and Flask

Abstract: The use of deep learning models to identify lesions on cotton leaves on the premise of images of the crop within the field is proposed in this article. Its cultivation in tropical regions has made it the target of a large spectrum of agricultural pests and diseases, and efficient economical solutions are needed. Moreover, the symptoms of the main pests and diseases cannot be differentiated within the initial stages, and also the correct identification of a lesion can be troublesome for the producer. To help resolve the problem, the present research provides a solution based on deep learning in the screening of cotton leaves that builds it attainable to watch the health of the cotton crop and make higher choices for its management. For this approach, Automatic classifier CNN will be used for classification based on learningwith some training samples of that twocategories. Finally the simulated result shows that used network classifier provides minimum error during training and better accuracy in classification. Keywords: Plant disease, deep learning, CNN, Classification.

Utilizing AI and Machine Learning for Human Emotional Analysis through Speech-to-Text Engine Data Conversion

Artificial Intelligence (AI) and Machine Learning (ML) technologies have revolutionized various domains, and one of their fascinating applications is in analyzing human emotions through speech-to-text data conversion. This article delves into the innovative use of AI and ML to process spoken words into text and subsequently evaluate human emotions embedded within the speech. We explore the development of speech-to-text technologies, the essential function of natural language processing (NLP), and the use of deep learning models to understand emotional cues. Our research discoveries illustrate how these technologies are altering fields like mental health, customer service, and market research, granting deeper understanding of human emotional states. By scrutinizing emotions conveyed in spoken language, artificial intelligence (AI) and machine learning (ML).

Cotton Plant Disease Prediction Using Deep Learning

Abstract: The use of deep learning models to identify lessions on cotton leaves on the basis of images of the crop in the field is proposed in this article. Its cultivation in tropical regions has made it the target of a wide spectrum of agricultural pests and diseases, and efficient solutions are required. Moreover, the symptoms of the main pests and diseases cannot be differentiated in the initial stages, and the correct identification of a lesion can be difficult for the producer. To help resolve the problem, the present research provides a solution based on deep learning in the screening of cotton leaves which makes it possible to monitor the health of the cotton crop and make better decisions for its management. For this approach, Automatic classifierCNN will be used for classification based on learningwith some training samples of that twocategories. Finally the simulated result shows that used network classifier provides minimum error during training and better accuracy in classification. Keywords: Plant disease, deep learning, CNN, Classification.

Deep Learning Methods in Short-Term Traffic Prediction: A Survey

Nowadays, traffic congestion has become a serious problem that plagues the development of many cities aroundthe world and the travel and life of urban residents. Compared with the costly and long implementation cyclemeasures such as the promotion of public transportation construction, vehicle restriction, road reconstruction, etc., traffic prediction is the lowest cost and best means to solve traffic congestion. Relevant departmentscan give early warnings on congested road sections based on the results of traffic prediction, rationalize thedistribution of police forces, and solve the traffic congestion problem. At the same time, due to the increasingreal-time requirements of current traffic prediction, short-term traffic prediction has become a subject of widespread concern and research. Currently, the most widely used model for short-term traffic prediction are deeplearning models. This survey studied the relevant literature on the use of deep learning models to solve shortterm traffic prediction problem in the top journals of transportation in recent years, summarized the currentcommonly used traffic datasets, the mainstream deep learning models and their applications in this field. Finally, the challenges and future development trends of deep learning models applied in this field are discussed.

A Neural Approach to Improve the Lee-Carter Mortality Density Forecasts

Several countries worldwide are experiencing a continuous increase in life expectancy, extending the challenges of life actuaries and demographers in forecasting mortality. Although several stochastic mortality models have been proposed in the literature, mortality forecasting research remains a crucial task. Recently, various research works have encouraged the use of deep learning models to extrapolate suitable patterns within mortality data. Such learning models allow achieving accurate point predictions, though uncertainty measures are also necessary to support both model estimate reliability and risk evaluation. As a new advance in mortality forecasting, we formalize the deep neural network integration within the Lee-Carter framework, as a first bridge between the deep learning and the mortality density forecasts. We test our model proposal in a numerical application considering three representative countries worldwide and for both genders, scrutinizing two different fitting periods. Exploiting the meaning of both biological reasonableness and plausibility of forecasts, as well as performance metrics, our findings confirm the suitability of deep learning models to improve the predictive capacity of the Lee-Carter model, providing more reliable mortality boundaries in the long run.

Citation Context Analysis Using Combined Feature Embedding and Deep Convolutional Neural Network Model

Citation creates a link between citing and the cited author, and the frequency of citation has been regarded as the basic element to measure the impact of research and knowledge-based achievements. Citation frequency has been widely used to calculate the impact factor, H index, i10 index, etc., of authors and journals. However, for a fair evaluation, the qualitative aspect should be considered along with the quantitative measures. The sentiments expressed in citation play an important role in evaluating the quality of the research because the citation may be used to indicate appreciation, criticism, or a basis for carrying on research. In-text citation analysis is a challenging task, despite the use of machine learning models and automatic sentiment annotation. Additionally, the use of deep learning models and word embedding is not studied very well. This study performs several experiments with machine learning and deep learning models using fastText, fastText subword, global vectors, and their blending for word representation to perform in-text sentiment analysis. A dimensionality reduction technique called principal component analysis (PCA) is utilized to reduce the feature vectors before passing them to the classifier. Additionally, a customized convolutional neural network (CNN) is presented to obtain higher classification accuracy. Results suggest that the deep learning CNN coupled with fastText word embedding produces the best results in terms of accuracy, precision, recall, and F1 measure.

Coastal Bathymetry Estimation from Sentinel-2 Satellite Imagery: Comparing Deep Learning and Physics-Based Approaches

The ability to monitor the evolution of the coastal zone over time is an important factor in coastal knowledge, development, planning, risk mitigation, and overall coastal zone management. While traditional bathymetry surveys using echo-sounding techniques are expensive and time consuming, remote sensing tools have recently emerged as reliable and inexpensive data sources that can be used to estimate bathymetry using depth inversion models. Deep learning is a growing field of artificial intelligence that allows for the automatic construction of models from data and has been successfully used for various Earth observation and model inversion applications. In this work, we make use of publicly available Sentinel-2 satellite imagery and multiple bathymetry surveys to train a deep learning-based bathymetry estimation model. We explore for the first time two complementary approaches, based on color information but also wave kinematics, as inputs to the deep learning model. This offers the possibility to derive bathymetry not only in clear waters as previously done with deep learning models but also at common turbid coastal zones. We show competitive results with a state-of-the-art physical inversion method for satellite-derived bathymetry, Satellite to Shores (S2Shores), demonstrating a promising direction for worldwide applicability of deep learning models to inverse bathymetry from satellite imagery and a novel use of deep learning models in Earth observation.

Kidney Tumor Semantic Segmentation Using Deep Learning: A Survey of State-of-the-Art.

Cure rates for kidney cancer vary according to stage and grade; hence, accurate diagnostic procedures for early detection and diagnosis are crucial. Some difficulties with manual segmentation have necessitated the use of deep learning models to assist clinicians in effectively recognizing and segmenting tumors. Deep learning (DL), particularly convolutional neural networks, has produced outstanding success in classifying and segmenting images. Simultaneously, researchers in the field of medical image segmentation employ DL approaches to solve problems such as tumor segmentation, cell segmentation, and organ segmentation. Segmentation of tumors semantically is critical in radiation and therapeutic practice. This article discusses current advances in kidney tumor segmentation systems based on DL. We discuss the various types of medical images and segmentation techniques and the assessment criteria for segmentation outcomes in kidney tumor segmentation, highlighting their building blocks and various strategies.

Benchmark of Deep Learning and a Proposed HSV Colour Space Models for the Detection and Classification of Greenhouse Tomato

The harvesting operation is a recurring task in the production of any crop, thus making it an excellent candidate for automation. In protected horticulture, one of the crops with high added value is tomatoes. However, its robotic harvesting is still far from maturity. That said, the development of an accurate fruit detection system is a crucial step towards achieving fully automated robotic harvesting. Deep Learning (DL) and detection frameworks like Single Shot MultiBox Detector (SSD) or You Only Look Once (YOLO) are more robust and accurate alternatives with better response to highly complex scenarios. The use of DL can be easily used to detect tomatoes, but when their classification is intended, the task becomes harsh, demanding a huge amount of data. Therefore, this paper proposes the use of DL models (SSD MobileNet v2 and YOLOv4) to efficiently detect the tomatoes and compare those systems with a proposed histogram-based HSV colour space model to classify each tomato and determine its ripening stage, through two image datasets acquired. Regarding detection, both models obtained promising results, with the YOLOv4 model standing out with an F1-Score of 85.81%. For classification task the YOLOv4 was again the best model with an Macro F1-Score of 74.16%. The HSV colour space model outperformed the SSD MobileNet v2 model, obtaining results similar to the YOLOv4 model, with a Balanced Accuracy of 68.10%.

Recurrent neural network model for high-speed train vibration prediction from time series

In this article, we want to discuss the use of deep learning model to predict potential vibrations of high-speed trains. In our research, we have tested and developed deep learning model to predict potential vibrations from time series of recorded vibrations during travel. We have tested various training models, different time steps and potential error margins to examine how well we are able to predict situation on the track. Summarizing, in our article we have used the RNN-LSTM neural network model with hyperbolic tangent in hidden layers and rectified linear unit gate at the final layer in order to predict future values from the time series data. Results of our research show the our system is able to predict vibrations with Accuracy of above 99% in series of values forward.

Main requirements of end-to-end deep learning models for biomedical time series classification in healthcare environments

The use of deep learning models with medical data is becoming more widespread. However, although numerous models have shown high accuracy in medical-related tasks, such as medical image recognition (e.g. radiographs), there are still many problems with seeing these models operating in a real healthcare environment. This article presents a series of basic requirements that must be taken into account when developing deep learning models for biomedical time series classification tasks, with the aim of facilitating the subsequent production of the models in healthcare. These requirements range from the correct collection of data, to the existing techniques for a correct explanation of the results obtained by the models. This is due to the fact that one of the main reasons why the use of deep learning models is not more widespread in healthcare settings is their lack of clarity when it comes to explaining decision making.

DESSO-DB: A web database for sequence and shape motif analyses and identification

Cis-regulatory motif (motif for short) identification and analyses are essential steps in detecting gene regulatory mechanisms. Deep learning (DL) models have shown substantial advances in motif prediction. In parallel, intuitive and integrative web databases are needed to make effective use of DL models and ensure easy access to the identified motifs. Here, we present DESSO-DB, a web database developed to allow efficient access to the identified motifs and diverse motif analyses. DESSO-DB provides motif prediction results and visualizations of 690 ENCODE human Chromatin Immunoprecipitation sequencing (ChIP-seq) data (including 161 transcription factors (TFs) in 91 cell lines) and 1,677 human ChIP-seq data (including 547 TFs in 359 cell lines) from Cistrome DB using DESSO, which is an in-house developed DL tool for motif prediction. It also provides online motif finding and scanning functions for new ChIP-seq/ATAC-seq datasets and downloadable motif results of the above 690 DECODE datasets, 126 cancer ChIP-seq, 55 RNA Crosslinking-Immunoprecipitation and high-throughput sequencing (CLIP-seq) data. DESSO-DB is deployed on the Google Cloud Platform, providing stabilized and efficient resources freely to the public. DESSO-DB is free and available at https://desso.osubmi.org/.

Graph-based molecular Pareto optimisation.

Computer-assisted design of small molecules has experienced a resurgence in academic and industrial interest due to the widespread use of data-driven techniques such as deep generative models. While the ability to generate molecules that fulfil required chemical properties is encouraging, the use of deep learning models requires significant, if not prohibitive, amounts of data and computational power. At the same time, open-sourcing of more traditional techniques such as graph-based genetic algorithms for molecular optimisation [Jensen, Chem. Sci., 2019, 12, 3567–3572] has shown that simple and training-free algorithms can be efficient and robust alternatives. Further research alleviated the common genetic algorithm issue of evolutionary stagnation by enforcing molecular diversity during optimisation [Van den Abeele, Chem. Sci., 2020, 42, 11485–11491]. The crucial lesson distilled from the simultaneous development of deep generative models and advanced genetic algorithms has been the importance of chemical space exploration [Aspuru-Guzik, Chem. Sci., 2021, 12, 7079–7090]. For single-objective optimisation problems, chemical space exploration had to be discovered as a useable resource but in multi-objective optimisation problems, an exploration of trade-offs between conflicting objectives is inherently present. In this paper we provide state-of-the-art and open-source implementations of two generations of graph-based non-dominated sorting genetic algorithms (NSGA-II, NSGA-III) for molecular multi-objective optimisation. We provide the results of a series of benchmarks for the inverse design of small molecule drugs for both the NSGA-II and NSGA-III algorithms. In addition, we introduce the dominated hypervolume and extended fingerprint based internal similarity as novel metrics for these benchmarks. By design, NSGA-II, and NSGA-III outperform a single optimisation method baseline in terms of dominated hypervolume, but remarkably our results show they do so without relying on a greater internal chemical diversity.

Fatigue State Detection for Tired Persons in Presence of Driving Periods

Because of the increased number of traffic accidents, there is an urgent need to control and reduce driving mistakes. Driver fatigue or drowsiness is one of these major mistakes. Many algorithms have been developed to address this issue by detecting fatigue and alerting the driver to this potentially dangerous condition. The developed algorithms’ main problem is their detection accuracy, as well as the time required to detect fatigue status and alert the driver. The accuracy and time represent a critical condition that affects the reduction of traffic accidents. Several datasets have been used in the development of fatigue or drowsy detection techniques. These data are gathered from the deriver’s brain Electroencephalogram (EEG) signals or video streaming recordings of the driver’s behavior. This paper proposes two distinct approaches to producing a high-performance fatigue detection system, the first based on the use of machine learning classifiers and the second depending on the use of deep learning models. The machine learning approach is used to process EEG signals, whereas the deep learning approach is used to process video streams. In machine learning classifiers, Support Vector Machine (SVM) provides up to 98% of detection accuracy, which is the highest accuracy among the other five deployed classifiers. In deep learning models, Convolutional Neural Network (CNN) provides up to 99% detection accuracy, which is the highest accuracy among the other two deployed models. The experimental results demonstrate that the two proposed algorithms provide the highest detection accuracy with the shortest Testing Time (TT) when compared to all other recent and efficient fatigue detection algorithms.