Catalogue Search | MBRL
Are you sure you want to remove the book from the shelf?
{{itemTitle}}
3,392
result(s) for
"Jung, Eun Sung"
Sort by:
Comparative Study on Distributed Lightweight Deep Learning Models for Road Pothole Detection
This paper delves into image detection based on distributed deep-learning techniques for intelligent traffic systems or self-driving cars. The accuracy and precision of neural networks deployed on edge devices (e.g., CCTV (closed-circuit television) for road surveillance) with small datasets may be compromised, leading to the misjudgment of targets. To address this challenge, TensorFlow and PyTorch were used to initialize various distributed model parallel and data parallel techniques. Despite the success of these techniques, communication constraints were observed along with certain speed issues. As a result, a hybrid pipeline was proposed, combining both dataset and model distribution through an all-reduced algorithm and NVlinks to prevent miscommunication among gradients. The proposed approach was tested on both an edge cluster and Google cluster environment, demonstrating superior performance compared to other test settings, with the quality of the bounding box detection system meeting expectations with increased reliability. Performance metrics, including total training time, images/second, cross-entropy loss, and total loss against the number of the epoch, were evaluated, revealing a robust competition between TensorFlow and PyTorch. The PyTorch environment’s hybrid pipeline outperformed other test settings.
Journal Article
Improving Air Pollution Prediction System through Multimodal Deep Learning Model Optimization
Many forms of air pollution increase as science and technology rapidly advance. In particular, fine dust harms the human body, causing or worsening heart and lung-related diseases. In this study, the level of fine dust in Seoul after 8 h is predicted to prevent health damage in advance. We construct a dataset by combining two modalities (i.e., numerical and image data) for accurate prediction. In addition, we propose a multimodal deep learning model combining a Long Short Term Memory (LSTM) and Convolutional Neural Network (CNN). An LSTM AutoEncoder is chosen as a model for numerical time series data processing and basic CNN. A Visual Geometry Group Neural Network (VGGNet) (VGG16, VGG19) is also chosen as a CNN model for image processing to compare performance differences according to network depth. The VGGNet is a standard deep CNN architecture with multiple layers. Our multimodal deep learning model using two modalities (i.e., numerical and image data) showed better performance than a single deep learning model using only one modality (numerical data). Specifically, the performance improved up to 14.16% when the VGG19 model, which has a deeper network, was used rather than the VGG16 model.
Journal Article
Development of Cybersecurity Technology and Algorithm Based on Quantum Computing
Many hacking incidents are linked to work files because most companies work with them. However, a variety of file encryption and decryption methods have been proposed. Existing file encryption/decryption technologies are under threat as hacking technologies advance, necessitating the development of stronger encryption algorithms. Therefore, in this study, we propose a modified advanced encryption standard (AES) algorithm and use quantum computing to encrypt/decrypt AES image files. Because the shift is regular during the AES Shift Row procedure, the change technique led the shift to become irregular when using quantum random walk. Computing resources and speeds were simulated using IBM Qiskit quantum simulators for performance evaluation, whereas encryption performance was assessed using number of pixels change rate (NPCR) and unified average changing intensity (UACI).
Journal Article
Design and Implementation of an Automated Disaster-Recovery System for a Kubernetes Cluster Using LSTM
With the increasing importance of data in modern business environments, effective data management and protection strategies are gaining increasing research attention. Data protection in a cloud environment is crucial for safeguarding information assets and maintaining sustainable services. This study introduces a system structure that integrates Kubernetes management platforms with backup and restoration tools. This system is designed to immediately detect disasters and automatically recover applications from another Kubernetes cluster. The experimental results show that this system executes the restoration process within 15 s without human intervention, enabling rapid recovery. This, in turn, significantly reduces the potential for delays and errors compared to manual recovery processes, thereby enhancing data management and recovery efficiency in cloud environments. Moreover, our research model predicts the CPU utilization of the cluster using Long Short-Term Memory (LSTM). The necessity of scheduling through this predict is made clearer through comparison with experiments without scheduling, demonstrating its ability to prevent performance degradation. This research highlights the efficiency and necessity of automatic recovery systems in cloud environments, setting a new direction for future research.
Journal Article
The Patients with Hirschsprung’s Disease Who Underwent Pull-Through at Age Less than 1 Year: Longitudinal Bowel Function
Background
Frequent stooling immediately after pull-through (PT), fecal soiling, and constipation are chronic complications of Hirschsprung’s disease (HD). This study aimed to investigate the longitudinal outcomes in terms of bowel function of patients below the age of 1 year undergoing PT.
Methods
We retrospectively evaluated 396 patients who underwent PT for HD between September 1979 and March 2014. Stool frequency was analyzed up to 10 years of age, and soiling and constipation were analyzed up to 15 years of age.
Results
After resection of the aganglionic segment (AS), stool frequency decreased over time. Furthermore, stool frequency among the three groups was similar 4 years after PT. Among the patients with aganglionic bowel resection, those who underwent the Soave procedure (SP) had an increase (0.56/day) in stool frequency than those who underwent the Duhamel procedure (DP). The soiling severity according to the AS was similar after 5 years of age. More severe soiling was better associated with patients who underwent the SP than those who underwent the DP. The constipation severity increased gradually until around 5 years and declined thereafter. More severe constipation was better associated with the DP than with the SP.
Conclusion
The result of the analysis of stool frequency and soiling in patients with HD indicated that shorter ASs resulted in fewer bowel movements and less severe soiling. However, with the increase in patient age, the differences became similar. Compared to the DP, the SP was associated with an increased frequency of bowel movements and soiling severity; however, the constipation severity was lower.
Journal Article
Analysis of Growth, Nutritional Status and Hospital Visitation Scores Associated with Reflux After Nissen Fundoplication in Neurologically Impaired Children with Gastroesophageal Reflux
Background
Neurologically impaired children (NIC) often experience swallowing difficulties and gastroesophageal reflux disease (GERD). Although these conditions could place children in a state of poor nutritional status and prevent them from thriving, there is insufficient research evaluating growth and nutritional status following fundoplication in these patients.
Method
This is a retrospective study of patients who were neurologically impaired and underwent Nissen fundoplication between April 2001 and March 2015. Seventy-six patients were enrolled, and the follow-up period was 12 months or longer. Growth was measured by the change in body weight and height. Nutritional status was measured by the change in body mass index, serum albumin and protein level.
Results
Median age at operation was 1.85 years old, and median body weight was 10 kg. The respective
Z
scores for weight and height showed significant improvements after 1 year since the operation compared to 1 year within the operation (−2.42 ± 2.19 vs. −1.31 ± 1.96,
P
< 0.001) (−1.6 ± 2.16 vs. −1.05 ± 1.69,
P
= 0.002). The respective
Z
scores for body mass index, albumin and protein also showed improvements after 1 year since the operation compared to 1 year within the operation (−2.07 ± 2.99 vs. −0.89 ± 2.1,
P
< 0.001) (3.55 ± 0.48 vs. 3.86 ± 0.45,
P
< 0.001) (6.22 ± 0.76 vs. 6.65 ± 0.51,
P
< 0.001). Hospital visitation scores associated with reflux were significantly lower after the operation (4.1 ± 3.43 vs. 1.18 ± 1.67,
P
< 0.001).
Conclusions
In summary, after Nissen fundoplication in NIC with GER, growth and nutritional status improved significantly. Also, hospital visitation scores associated with reflux decreased after the operation.
Journal Article
Risk factors of meconium-related ileus in very low birth weight infants: patients-control study
Very low birth weight (VLBW) neonates experience various problems, including meconium-related ileus (MRI). This study investigated the risk factors of MRI and surgical MRI in VLBW infants. VLBW neonates admitted to the Neonatal Intensive Care Unit of Seoul National University Children’s Hospital from October 2002 to September 2016 were included in the study. The diagnostic criteria for MRI were a decreased frequency of defecation with intolerable feeding, vomiting, and increased gastric residue (>50%); meconium-filled bowel dilatation in an imaging study; and no evidence of necrotizing enteritis or spontaneous intestinal perforation. Medical MRIs and surgical MRIs were managed through conventional treatment and surgical intervention. Of 1543 neonates, 69 and 1474 were in the patient and control groups, respectively. The risk factors for MRI include low birth weight (BW), cesarean section delivery, fetal distress, maternal diabetes, maternal hypertension, and maternal steroid use. Low BW and fetal distress were independent risk factors for MRI. Compared to the medical MRI group (n = 44), the risk factors for surgical MRI (n = 25) included males, younger gestational age, low BW, and meconium located at the small bowel. Male gender and low BW were independent risk factors for surgical MRI. Low BW and fetal distress were independent risk factors for MRI and male gender and low BW were independent risk factors for surgical MRI. In VLBW neonates, careful attention to the risk factors for MRI could minimize or avoid surgical interventions.
Journal Article
High-performance IoT streaming data prediction system using Spark: a case study of air pollution
Internet-of-Things (IoT) devices are becoming prevalent, and some of them, such as sensors, generate continuous time-series data, i.e., streaming data. These IoT streaming data are one of Big Data sources, and they require careful consideration for efficient data processing and analysis. Deep learning is emerging as a solution to IoT streaming data analytics. However, there is a persistent problem in deep learning that it takes a long time to learn neural networks. In this paper, we propose a high-performance IoT streaming data prediction system to improve the learning speed and to predict in real time. We showed the efficacy of the system through a case study of air pollution. The experimental results show that the modified LSTM autoencoder model shows the best performance compared to a generic LSTM model. We noticed that achieving the best performance requires optimizing many parameters, including learning rate, epoch, memory cell size, input timestep size, and the number of features/predictors. In that regard, we show that the high-performance data learning/prediction frameworks (e.g., Spark, Dist-Keras, and Hadoop) are essential to rapidly fine-tune a model for training and testing before real deployment of the model as data accumulate.
Journal Article
Predicting Survival of Congenital Diaphragmatic Hernia on the First Day of Life
Background
This study aimed to determine perinatal risk factors for 30-day mortality of congenital diaphragmatic hernia (CDH) patients and develop a prognostic index to predict 30-day mortality of CDH patients. Identifying risk factors that can prognosticate outcome is critical to obtain the best management practices for patients.
Methods
A retrospective study was performed for patients who were diagnosed with CDH from November 2000 to August 2016. A total of 10 prenatal risk factors and 14 postnatal risk factors were analyzed. All postnatal variables were measured within 24 h after birth.
Results
A total of 95 CDH patients were enrolled in this study, including 61 males and 34 females with mean gestational age of 38.86 ± 1.51 weeks. The overall 30-day survival rate was 63.2%. Multivariate analysis revealed that five factors (polyhydramnios, gestational age at diagnosis <25 weeks, observed-to-expected lung-to-head ratio ≤45, best oxygenation index in 24 h >11, and severity of tricuspid regurgitation ≥ mild) were independent predictors of 30-day mortality of CDH. Using these five factors, a perinatal prognostic index for 30-day mortality was developed. Four predictive models (poor, bad, good, and excellent) of the perinatal prognostic index were constructed, and external validation was performed.
Conclusions
Awareness of risk factors is very important for predicting prognosis and managing patients. Five independent perinatal risk factors were identified in this study. A perinatal prognostic index was developed for 30-day mortality for patients with CDH. This index may be used to help manage CDH patients.
Journal Article