Search results for: deep learning models
Commenced in January 2007
Frequency: Monthly
Edition: International
Paper Count: 14005

Search results for: deep learning models

13555 Artificial Intelligence-Based Detection of Individuals Suffering from Vestibular Disorder

Authors: Dua Hişam, Serhat İkizoğlu

Abstract:

Identifying the problem behind balance disorder is one of the most interesting topics in the medical literature. This study has considerably enhanced the development of artificial intelligence (AI) algorithms applying multiple machine learning (ML) models to sensory data on gait collected from humans to classify between normal people and those suffering from Vestibular System (VS) problems. Although AI is widely utilized as a diagnostic tool in medicine, AI models have not been used to perform feature extraction and identify VS disorders through training on raw data. In this study, three machine learning (ML) models, the Random Forest Classifier (RF), Extreme Gradient Boosting (XGB), and K-Nearest Neighbor (KNN), have been trained to detect VS disorder, and the performance comparison of the algorithms has been made using accuracy, recall, precision, and f1-score. With an accuracy of 95.28 %, Random Forest Classifier (RF) was the most accurate model.

Keywords: vestibular disorder, machine learning, random forest classifier, k-nearest neighbor, extreme gradient boosting

Procedia PDF Downloads 69
13554 Investigating the performance of machine learning models on PM2.5 forecasts: A case study in the city of Thessaloniki

Authors: Alexandros Pournaras, Anastasia Papadopoulou, Serafim Kontos, Anastasios Karakostas

Abstract:

The air quality of modern cities is an important concern, as poor air quality contributes to human health and environmental issues. Reliable air quality forecasting has, thus, gained scientific and governmental attention as an essential tool that enables authorities to take proactive measures for public safety. In this study, the potential of Machine Learning (ML) models to forecast PM2.5 at local scale is investigated in the city of Thessaloniki, the second largest city in Greece, which has been struggling with the persistent issue of air pollution. ML models, with proven ability to address timeseries forecasting, are employed to predict the PM2.5 concentrations and the respective Air Quality Index 5-days ahead by learning from daily historical air quality and meteorological data from 2014 to 2016 and gathered from two stations with different land use characteristics in the urban fabric of Thessaloniki. The performance of the ML models on PM2.5 concentrations is evaluated with common statistical methods, such as R squared (r²) and Root Mean Squared Error (RMSE), utilizing a portion of the stations’ measurements as test set. A multi-categorical evaluation is utilized for the assessment of their performance on respective AQIs. Several conclusions were made from the experiments conducted. Experimenting on MLs’ configuration revealed a moderate effect of various parameters and training schemas on the model’s predictions. Their performance of all these models were found to produce satisfactory results on PM2.5 concentrations. In addition, their application on untrained stations showed that these models can perform well, indicating a generalized behavior. Moreover, their performance on AQI was even better, showing that the MLs can be used as predictors for AQI, which is the direct information provided to the general public.

Keywords: Air Quality, AQ Forecasting, AQI, Machine Learning, PM2.5

Procedia PDF Downloads 77
13553 Creation and Management of Knowledge for Organization Sustainability and Learning

Authors: Deepa Kapoor, Rajshree Singh

Abstract:

This paper appreciates the emergence and growing importance as a new production factor makes the development of technologies, methodologies and strategies for measurement, creation, and diffusion into one of the main priorities of the organizations in the knowledge society. There are many models for creation and management of knowledge and diverse and varied perspectives for study, analysis, and understanding. In this article, we will conduct a theoretical approach to the type of models for the creation and management of knowledge; we will discuss some of them and see some of the difficulties and the key factors that determine the success of the processes for the creation and management of knowledge.

Keywords: knowledge creation, knowledge management, organizational development, organization learning

Procedia PDF Downloads 345
13552 Improving the Performance of Deep Learning in Facial Emotion Recognition with Image Sharpening

Authors: Ksheeraj Sai Vepuri, Nada Attar

Abstract:

We as humans use words with accompanying visual and facial cues to communicate effectively. Classifying facial emotion using computer vision methodologies has been an active research area in the computer vision field. In this paper, we propose a simple method for facial expression recognition that enhances accuracy. We tested our method on the FER-2013 dataset that contains static images. Instead of using Histogram equalization to preprocess the dataset, we used Unsharp Mask to emphasize texture and details and sharpened the edges. We also used ImageDataGenerator from Keras library for data augmentation. Then we used Convolutional Neural Networks (CNN) model to classify the images into 7 different facial expressions, yielding an accuracy of 69.46% on the test set. Our results show that using image preprocessing such as the sharpening technique for a CNN model can improve the performance, even when the CNN model is relatively simple.

Keywords: facial expression recognittion, image preprocessing, deep learning, CNN

Procedia PDF Downloads 143
13551 Analyzing the Performance of Machine Learning Models to Predict Alzheimer's Disease and its Stages Addressing Missing Value Problem

Authors: Carlos Theran, Yohn Parra Bautista, Victor Adankai, Richard Alo, Jimwi Liu, Clement G. Yedjou

Abstract:

Alzheimer's disease (AD) is a neurodegenerative disorder primarily characterized by deteriorating cognitive functions. AD has gained relevant attention in the last decade. An estimated 24 million people worldwide suffered from this disease by 2011. In 2016 an estimated 40 million were diagnosed with AD, and for 2050 is expected to reach 131 million people affected by AD. Therefore, detecting and confirming AD at its different stages is a priority for medical practices to provide adequate and accurate treatments. Recently, Machine Learning (ML) models have been used to study AD's stages handling missing values in multiclass, focusing on the delineation of Early Mild Cognitive Impairment (EMCI), Late Mild Cognitive Impairment (LMCI), and normal cognitive (CN). But, to our best knowledge, robust performance information of these models and the missing data analysis has not been presented in the literature. In this paper, we propose studying the performance of five different machine learning models for AD's stages multiclass prediction in terms of accuracy, precision, and F1-score. Also, the analysis of three imputation methods to handle the missing value problem is presented. A framework that integrates ML model for AD's stages multiclass prediction is proposed, performing an average accuracy of 84%.

Keywords: alzheimer's disease, missing value, machine learning, performance evaluation

Procedia PDF Downloads 250
13550 Convolutional Neural Networks versus Radiomic Analysis for Classification of Breast Mammogram

Authors: Mehwish Asghar

Abstract:

Breast Cancer (BC) is a common type of cancer among women. Its screening is usually performed using different imaging modalities such as magnetic resonance imaging, mammogram, X-ray, CT, etc. Among these modalities’ mammogram is considered a powerful tool for diagnosis and screening of breast cancer. Sophisticated machine learning approaches have shown promising results in complementing human diagnosis. Generally, machine learning methods can be divided into two major classes: one is Radiomics analysis (RA), where image features are extracted manually; and the other one is the concept of convolutional neural networks (CNN), in which the computer learns to recognize image features on its own. This research aims to improve the incidence of early detection, thus reducing the mortality rate caused by breast cancer through the latest advancements in computer science, in general, and machine learning, in particular. It has also been aimed to ease the burden of doctors by improving and automating the process of breast cancer detection. This research is related to a relative analysis of different techniques for the implementation of different models for detecting and classifying breast cancer. The main goal of this research is to provide a detailed view of results and performances between different techniques. The purpose of this paper is to explore the potential of a convolutional neural network (CNN) w.r.t feature extractor and as a classifier. Also, in this research, it has been aimed to add the module of Radiomics for comparison of its results with deep learning techniques.

Keywords: breast cancer (BC), machine learning (ML), convolutional neural network (CNN), radionics, magnetic resonance imaging, artificial intelligence

Procedia PDF Downloads 224
13549 Integrating Knowledge Distillation of Multiple Strategies

Authors: Min Jindong, Wang Mingxia

Abstract:

With the widespread use of artificial intelligence in life, computer vision, especially deep convolutional neural network models, has developed rapidly. With the increase of the complexity of the real visual target detection task and the improvement of the recognition accuracy, the target detection network model is also very large. The huge deep neural network model is not conducive to deployment on edge devices with limited resources, and the timeliness of network model inference is poor. In this paper, knowledge distillation is used to compress the huge and complex deep neural network model, and the knowledge contained in the complex network model is comprehensively transferred to another lightweight network model. Different from traditional knowledge distillation methods, we propose a novel knowledge distillation that incorporates multi-faceted features, called M-KD. In this paper, when training and optimizing the deep neural network model for target detection, the knowledge of the soft target output of the teacher network in knowledge distillation, the relationship between the layers of the teacher network and the feature attention map of the hidden layer of the teacher network are transferred to the student network as all knowledge. in the model. At the same time, we also introduce an intermediate transition layer, that is, an intermediate guidance layer, between the teacher network and the student network to make up for the huge difference between the teacher network and the student network. Finally, this paper adds an exploration module to the traditional knowledge distillation teacher-student network model. The student network model not only inherits the knowledge of the teacher network but also explores some new knowledge and characteristics. Comprehensive experiments in this paper using different distillation parameter configurations across multiple datasets and convolutional neural network models demonstrate that our proposed new network model achieves substantial improvements in speed and accuracy performance.

Keywords: object detection, knowledge distillation, convolutional network, model compression

Procedia PDF Downloads 278
13548 Multimedia Design in Tactical Play Learning and Acquisition for Elite Gaelic Football Practitioners

Authors: Michael McMahon

Abstract:

The use of media (video/animation/graphics) has long been used by athletes, coaches, and sports scientists to analyse and improve performance in technical skills and team tactics. Sports educators are increasingly open to the use of technology to support coach and learner development. However, an overreliance is a concern., This paper is part of a larger Ph.D. study looking into these new challenges for Sports Educators. Most notably, how to exploit the deep-learning potential of Digital Media among expert learners, how to instruct sports educators to create effective media content that fosters deep learning, and finally, how to make the process manageable and cost-effective. Central to the study is Richard Mayers Cognitive Theory of Multimedia Learning. Mayers Multimedia Learning Theory proposes twelve principles that shape the design and organization of multimedia presentations to improve learning and reduce cognitive load. For example, the Prior Knowledge principle suggests and highlights different learning outcomes for Novice and Non-Novice learners, respectively. Little research, however, is available to support this principle in modified domains (e.g., sports tactics and strategy). As a foundation for further research, this paper compares and contrasts a range of contemporary multimedia sports coaching content and assesses how they perform as learning tools for Strategic and Tactical Play Acquisition among elite sports practitioners. The stress tests applied are guided by Mayers's twelve Multimedia Learning Principles. The focus is on the elite athletes and whether current coaching digital media content does foster improved sports learning among this cohort. The sport of Gaelic Football was selected as it has high strategic and tactical play content, a wide range of Practitioner skill levels (Novice to Elite), and also a significant volume of Multimedia Coaching Content available for analysis. It is hoped the resulting data will help identify and inform the future instructional content design and delivery for Sports Practitioners and help promote best design practices optimal for different levels of expertise.

Keywords: multimedia learning, e-learning, design for learning, ICT

Procedia PDF Downloads 103
13547 Using Deep Learning for the Detection of Faulty RJ45 Connectors on a Radio Base Station

Authors: Djamel Fawzi Hadj Sadok, Marrone Silvério Melo Dantas Pedro Henrique Dreyer, Gabriel Fonseca Reis de Souza, Daniel Bezerra, Ricardo Souza, Silvia Lins, Judith Kelner

Abstract:

A radio base station (RBS), part of the radio access network, is a particular type of equipment that supports the connection between a wide range of cellular user devices and an operator network access infrastructure. Nowadays, most of the RBS maintenance is carried out manually, resulting in a time consuming and costly task. A suitable candidate for RBS maintenance automation is repairing faulty links between devices caused by missing or unplugged connectors. A suitable candidate for RBS maintenance automation is repairing faulty links between devices caused by missing or unplugged connectors. This paper proposes and compares two deep learning solutions to identify attached RJ45 connectors on network ports. We named connector detection, the solution based on object detection, and connector classification, the one based on object classification. With the connector detection, we get an accuracy of 0:934, mean average precision 0:903. Connector classification, get a maximum accuracy of 0:981 and an AUC of 0:989. Although connector detection was outperformed in this study, this should not be viewed as an overall result as connector detection is more flexible for scenarios where there is no precise information about the environment and the possible devices. At the same time, the connector classification requires that information to be well-defined.

Keywords: radio base station, maintenance, classification, detection, deep learning, automation

Procedia PDF Downloads 200
13546 Experiments on Weakly-Supervised Learning on Imperfect Data

Authors: Yan Cheng, Yijun Shao, James Rudolph, Charlene R. Weir, Beth Sahlmann, Qing Zeng-Treitler

Abstract:

Supervised predictive models require labeled data for training purposes. Complete and accurate labeled data, i.e., a ‘gold standard’, is not always available, and imperfectly labeled data may need to serve as an alternative. An important question is if the accuracy of the labeled data creates a performance ceiling for the trained model. In this study, we trained several models to recognize the presence of delirium in clinical documents using data with annotations that are not completely accurate (i.e., weakly-supervised learning). In the external evaluation, the support vector machine model with a linear kernel performed best, achieving an area under the curve of 89.3% and accuracy of 88%, surpassing the 80% accuracy of the training sample. We then generated a set of simulated data and carried out a series of experiments which demonstrated that models trained on imperfect data can (but do not always) outperform the accuracy of the training data, e.g., the area under the curve for some models is higher than 80% when trained on the data with an error rate of 40%. Our experiments also showed that the error resistance of linear modeling is associated with larger sample size, error type, and linearity of the data (all p-values < 0.001). In conclusion, this study sheds light on the usefulness of imperfect data in clinical research via weakly-supervised learning.

Keywords: weakly-supervised learning, support vector machine, prediction, delirium, simulation

Procedia PDF Downloads 198
13545 A Research Agenda for Learner Models for Adaptive Educational Digital Learning Environments

Authors: Felix Böck

Abstract:

Nowadays, data about learners and their digital activities are collected, which could help educational institutions to better understand learning processes, improve them and be able to provide better learning assistance. In this research project, custom knowledge- and data-driven recommendation algorithms will be used to offer students in higher education integrated learning assistance. The pre-requisite for this is a learner model that is as comprehensive as possible, which should first be created and then kept up-to-date largely automatically for being able to individualize and personalize the learning experience. In order to create such a learner model, a roadmap is presented that describes the individual phases up to the creation and evaluation of the finished model. The methodological process for the research project is disclosed, and the research question of how learners can be supported in their learning with personalized, customized learning recommendations is explored.

Keywords: research agenda, user model, learner model, higher education, adaptive educational digital learning environments, personalized learning paths, recommendation system, adaptation, personalization

Procedia PDF Downloads 16
13544 A Survey of Response Generation of Dialogue Systems

Authors: Yifan Fan, Xudong Luo, Pingping Lin

Abstract:

An essential task in the field of artificial intelligence is to allow computers to interact with people through natural language. Therefore, researches such as virtual assistants and dialogue systems have received widespread attention from industry and academia. The response generation plays a crucial role in dialogue systems, so to push forward the research on this topic, this paper surveys various methods for response generation. We sort out these methods into three categories. First one includes finite state machine methods, framework methods, and instance methods. The second contains full-text indexing methods, ontology methods, vast knowledge base method, and some other methods. The third covers retrieval methods and generative methods. We also discuss some hybrid methods based knowledge and deep learning. We compare their disadvantages and advantages and point out in which ways these studies can be improved further. Our discussion covers some studies published in leading conferences such as IJCAI and AAAI in recent years.

Keywords: deep learning, generative, knowledge, response generation, retrieval

Procedia PDF Downloads 134
13543 Effect of Hybrid Learning in Higher Education

Authors: A. Meydanlioglu, F. Arikan

Abstract:

In recent years, thanks to the development of information and communication technologies, the computer and internet have been used widely in higher education. Internet-based education is impacting traditional higher education as online components increasingly become integrated into face-to-face (FTF) courses. The goal of combined internet-based and traditional education is to take full advantage of the benefits of each platform in order to provide an educational opportunity that can promote student learning better than can either platform alone. Research results show that the use of hybrid learning is more effective than online or FTF models in higher education. Due to the potential benefits, an increasing number of institutions are interested in developing hybrid courses, programs, and degrees. Future research should evaluate the effectiveness of hybrid learning. This paper is designed to determine the impact of hybrid learning on higher education.

Keywords: e-learning, higher education, hybrid learning, online education

Procedia PDF Downloads 909
13542 Hyperspectral Data Classification Algorithm Based on the Deep Belief and Self-Organizing Neural Network

Authors: Li Qingjian, Li Ke, He Chun, Huang Yong

Abstract:

In this paper, the method of combining the Pohl Seidman's deep belief network with the self-organizing neural network is proposed to classify the target. This method is mainly aimed at the high nonlinearity of the hyperspectral image, the high sample dimension and the difficulty in designing the classifier. The main feature of original data is extracted by deep belief network. In the process of extracting features, adding known labels samples to fine tune the network, enriching the main characteristics. Then, the extracted feature vectors are classified into the self-organizing neural network. This method can effectively reduce the dimensions of data in the spectrum dimension in the preservation of large amounts of raw data information, to solve the traditional clustering and the long training time when labeled samples less deep learning algorithm for training problems, improve the classification accuracy and robustness. Through the data simulation, the results show that the proposed network structure can get a higher classification precision in the case of a small number of known label samples.

Keywords: DBN, SOM, pattern classification, hyperspectral, data compression

Procedia PDF Downloads 341
13541 Graphical User Interface Testing by Using Deep Learning

Authors: Akshat Mathur, Sunil Kumar Khatri

Abstract:

This paper presents brief about how the use of Artificial intelligence in respect to GUI testing can reduce workload by using DL-fueled method. This paper also discusses about how graphical user interface and event driven software testing can derive benefits from the use of AI techniques. The use of AI techniques not only reduces the task and work load but also helps in getting better output than manual testing. Although results are same, but the use of Artifical intelligence techniques for GUI testing has proven to provide ideal results. DL-fueled framework helped us to find imperfections of the entire webpage and provides test failure result in a score format between 0 and 1which signifies that are test meets it quality criteria or not. This paper proposes DL-fueled method which helps us to find the genuine GUI bugs and defects and also helped us to scale the existing labour-intensive and skill-intensive methodologies.

Keywords: graphical user interface, GUI, artificial intelligence, deep learning, ML technology

Procedia PDF Downloads 177
13540 Machine Learning Analysis of Student Success in Introductory Calculus Based Physics I Course

Authors: Chandra Prayaga, Aaron Wade, Lakshmi Prayaga, Gopi Shankar Mallu

Abstract:

This paper presents the use of machine learning algorithms to predict the success of students in an introductory physics course. Data having 140 rows pertaining to the performance of two batches of students was used. The lack of sufficient data to train robust machine learning models was compensated for by generating synthetic data similar to the real data. CTGAN and CTGAN with Gaussian Copula (Gaussian) were used to generate synthetic data, with the real data as input. To check the similarity between the real data and each synthetic dataset, pair plots were made. The synthetic data was used to train machine learning models using the PyCaret package. For the CTGAN data, the Ada Boost Classifier (ADA) was found to be the ML model with the best fit, whereas the CTGAN with Gaussian Copula yielded Logistic Regression (LR) as the best model. Both models were then tested for accuracy with the real data. ROC-AUC analysis was performed for all the ten classes of the target variable (Grades A, A-, B+, B, B-, C+, C, C-, D, F). The ADA model with CTGAN data showed a mean AUC score of 0.4377, but the LR model with the Gaussian data showed a mean AUC score of 0.6149. ROC-AUC plots were obtained for each Grade value separately. The LR model with Gaussian data showed consistently better AUC scores compared to the ADA model with CTGAN data, except in two cases of the Grade value, C- and A-.

Keywords: machine learning, student success, physics course, grades, synthetic data, CTGAN, gaussian copula CTGAN

Procedia PDF Downloads 44
13539 CyberSteer: Cyber-Human Approach for Safely Shaping Autonomous Robotic Behavior to Comply with Human Intention

Authors: Vinicius G. Goecks, Gregory M. Gremillion, William D. Nothwang

Abstract:

Modern approaches to train intelligent agents rely on prolonged training sessions, high amounts of input data, and multiple interactions with the environment. This restricts the application of these learning algorithms in robotics and real-world applications, in which there is low tolerance to inadequate actions, interactions are expensive, and real-time processing and action are required. This paper addresses this issue introducing CyberSteer, a novel approach to efficiently design intrinsic reward functions based on human intention to guide deep reinforcement learning agents with no environment-dependent rewards. CyberSteer uses non-expert human operators for initial demonstration of a given task or desired behavior. The trajectories collected are used to train a behavior cloning deep neural network that asynchronously runs in the background and suggests actions to the deep reinforcement learning module. An intrinsic reward is computed based on the similarity between actions suggested and taken by the deep reinforcement learning algorithm commanding the agent. This intrinsic reward can also be reshaped through additional human demonstration or critique. This approach removes the need for environment-dependent or hand-engineered rewards while still being able to safely shape the behavior of autonomous robotic agents, in this case, based on human intention. CyberSteer is tested in a high-fidelity unmanned aerial vehicle simulation environment, the Microsoft AirSim. The simulated aerial robot performs collision avoidance through a clustered forest environment using forward-looking depth sensing and roll, pitch, and yaw references angle commands to the flight controller. This approach shows that the behavior of robotic systems can be shaped in a reduced amount of time when guided by a non-expert human, who is only aware of the high-level goals of the task. Decreasing the amount of training time required and increasing safety during training maneuvers will allow for faster deployment of intelligent robotic agents in dynamic real-world applications.

Keywords: human-robot interaction, intelligent robots, robot learning, semisupervised learning, unmanned aerial vehicles

Procedia PDF Downloads 259
13538 A Machine Learning Approach for Intelligent Transportation System Management on Urban Roads

Authors: Ashish Dhamaniya, Vineet Jain, Rajesh Chouhan

Abstract:

Traffic management is one of the gigantic issue in most of the urban roads in al-most all metropolitan cities in India. Speed is one of the critical traffic parameters for effective Intelligent Transportation System (ITS) implementation as it decides the arrival rate of vehicles on an intersection which are majorly the point of con-gestions. The study aimed to leverage Machine Learning (ML) models to produce precise predictions of speed on urban roadway links. The research objective was to assess how categorized traffic volume and road width, serving as variables, in-fluence speed prediction. Four tree-based regression models namely: Decision Tree (DT), Random Forest (RF), Extra Tree (ET), and Extreme Gradient Boost (XGB)are employed for this purpose. The models' performances were validated using test data, and the results demonstrate that Random Forest surpasses other machine learning techniques and a conventional utility theory-based model in speed prediction. The study is useful for managing the urban roadway network performance under mixed traffic conditions and effective implementation of ITS.

Keywords: stream speed, urban roads, machine learning, traffic flow

Procedia PDF Downloads 70
13537 Enhanced Image Representation for Deep Belief Network Classification of Hyperspectral Images

Authors: Khitem Amiri, Mohamed Farah

Abstract:

Image classification is a challenging task and is gaining lots of interest since it helps us to understand the content of images. Recently Deep Learning (DL) based methods gave very interesting results on several benchmarks. For Hyperspectral images (HSI), the application of DL techniques is still challenging due to the scarcity of labeled data and to the curse of dimensionality. Among other approaches, Deep Belief Network (DBN) based approaches gave a fair classification accuracy. In this paper, we address the problem of the curse of dimensionality by reducing the number of bands and replacing the HSI channels by the channels representing radiometric indices. Therefore, instead of using all the HSI bands, we compute the radiometric indices such as NDVI (Normalized Difference Vegetation Index), NDWI (Normalized Difference Water Index), etc, and we use the combination of these indices as input for the Deep Belief Network (DBN) based classification model. Thus, we keep almost all the pertinent spectral information while reducing considerably the size of the image. In order to test our image representation, we applied our method on several HSI datasets including the Indian pines dataset, Jasper Ridge data and it gave comparable results to the state of the art methods while reducing considerably the time of training and testing.

Keywords: hyperspectral images, deep belief network, radiometric indices, image classification

Procedia PDF Downloads 280
13536 Analysis of Biomarkers Intractable Epileptogenic Brain Networks with Independent Component Analysis and Deep Learning Algorithms: A Comprehensive Framework for Scalable Seizure Prediction with Unimodal Neuroimaging Data in Pediatric Patients

Authors: Bliss Singhal

Abstract:

Epilepsy is a prevalent neurological disorder affecting approximately 50 million individuals worldwide and 1.2 million Americans. There exist millions of pediatric patients with intractable epilepsy, a condition in which seizures fail to come under control. The occurrence of seizures can result in physical injury, disorientation, unconsciousness, and additional symptoms that could impede children's ability to participate in everyday tasks. Predicting seizures can help parents and healthcare providers take precautions, prevent risky situations, and mentally prepare children to minimize anxiety and nervousness associated with the uncertainty of a seizure. This research proposes a comprehensive framework to predict seizures in pediatric patients by evaluating machine learning algorithms on unimodal neuroimaging data consisting of electroencephalogram signals. The bandpass filtering and independent component analysis proved to be effective in reducing the noise and artifacts from the dataset. Various machine learning algorithms’ performance is evaluated on important metrics such as accuracy, precision, specificity, sensitivity, F1 score and MCC. The results show that the deep learning algorithms are more successful in predicting seizures than logistic Regression, and k nearest neighbors. The recurrent neural network (RNN) gave the highest precision and F1 Score, long short-term memory (LSTM) outperformed RNN in accuracy and convolutional neural network (CNN) resulted in the highest Specificity. This research has significant implications for healthcare providers in proactively managing seizure occurrence in pediatric patients, potentially transforming clinical practices, and improving pediatric care.

Keywords: intractable epilepsy, seizure, deep learning, prediction, electroencephalogram channels

Procedia PDF Downloads 84
13535 Detecting Memory-Related Gene Modules in sc/snRNA-seq Data by Deep-Learning

Authors: Yong Chen

Abstract:

To understand the detailed molecular mechanisms of memory formation in engram cells is one of the most fundamental questions in neuroscience. Recent single-cell RNA-seq (scRNA-seq) and single-nucleus RNA-seq (snRNA-seq) techniques have allowed us to explore the sparsely activated engram ensembles, enabling access to the molecular mechanisms that underlie experience-dependent memory formation and consolidation. However, the absence of specific and powerful computational methods to detect memory-related genes (modules) and their regulatory relationships in the sc/snRNA-seq datasets has strictly limited the analysis of underlying mechanisms and memory coding principles in mammalian brains. Here, we present a deep-learning method named SCENTBOX, to detect memory-related gene modules and causal regulatory relationships among themfromsc/snRNA-seq datasets. SCENTBOX first constructs codifferential expression gene network (CEGN) from case versus control sc/snRNA-seq datasets. It then detects the highly correlated modules of differential expression genes (DEGs) in CEGN. The deep network embedding and attention-based convolutional neural network strategies are employed to precisely detect regulatory relationships among DEG genes in a module. We applied them on scRNA-seq datasets of TRAP; Ai14 mouse neurons with fear memory and detected not only known memory-related genes, but also the modules and potential causal regulations. Our results provided novel regulations within an interesting module, including Arc, Bdnf, Creb, Dusp1, Rgs4, and Btg2. Overall, our methods provide a general computational tool for processing sc/snRNA-seq data from case versus control studie and a systematic investigation of fear-memory-related gene modules.

Keywords: sc/snRNA-seq, memory formation, deep learning, gene module, causal inference

Procedia PDF Downloads 120
13534 An Exploration of First Year Bachelor of Education Degree Students’ Learning Preferences in Academic Literacy in a Private Higher Education Institution: A Case for the Blended Learning Approach

Authors: K. Kannapathi-Naidoo

Abstract:

The higher education landscape has undergone changes in the past decade, with concepts such as blended learning, online learning, and hybrid models appearing more frequently in research and practice. The year 2020 marked a mass migration from face-to-face learning and more traditional forms of education to online learning in higher education institutions across the globe due to the Covid-19 pandemic. As a result, contact learning students and lecturing staff alike were thrust into the world of online learning at an unprecedented pace. Traditional modes of learning had to be amended, and pedagogical strategies required adjustments. This study was located within a compulsory first-year academic literacy module in a higher education institution. The study aimed to explore students’ learning preferences between online, face-face, and blended learning within the context of academic literacy. Data was collected through online qualitative questionnaires administered to 150 first-year students, which were then analysed thematically. The findings of the study revealed that 48.5% of the participants preferred a blended learning approach to academic literacy. The main themes that emerged in support of their preference were best of both worlds, flexibility, productivity, and lecturer accessibility. As a result, this paper advocates for the blended learning approach for academic literacy skills-based modules.

Keywords: academic literacy, blended learning, online learning, student learning preferences

Procedia PDF Downloads 75
13533 Churn Prediction for Savings Bank Customers: A Machine Learning Approach

Authors: Prashant Verma

Abstract:

Commercial banks are facing immense pressure, including financial disintermediation, interest rate volatility and digital ways of finance. Retaining an existing customer is 5 to 25 less expensive than acquiring a new one. This paper explores customer churn prediction, based on various statistical & machine learning models and uses under-sampling, to improve the predictive power of these models. The results show that out of the various machine learning models, Random Forest which predicts the churn with 78% accuracy, has been found to be the most powerful model for the scenario. Customer vintage, customer’s age, average balance, occupation code, population code, average withdrawal amount, and an average number of transactions were found to be the variables with high predictive power for the churn prediction model. The model can be deployed by the commercial banks in order to avoid the customer churn so that they may retain the funds, which are kept by savings bank (SB) customers. The article suggests a customized campaign to be initiated by commercial banks to avoid SB customer churn. Hence, by giving better customer satisfaction and experience, the commercial banks can limit the customer churn and maintain their deposits.

Keywords: savings bank, customer churn, customer retention, random forests, machine learning, under-sampling

Procedia PDF Downloads 143
13532 Geovisualisation for Defense Based on a Deep Learning Monocular Depth Reconstruction Approach

Authors: Daniel R. dos Santos, Mateus S. Maldonado, Estevão J. R. Batista

Abstract:

The military commanders increasingly dependent on spatial awareness, as knowing where enemy are, understanding how war battle scenarios change over time, and visualizing these trends in ways that offer insights for decision-making. Thanks to advancements in geospatial technologies and artificial intelligence algorithms, the commanders are now able to modernize military operations on a universal scale. Thus, geovisualisation has become an essential asset in the defense sector. It has become indispensable for better decisionmaking in dynamic/temporal scenarios, operation planning and management for the war field, situational awareness, effective planning, monitoring, and others. For example, a 3D visualization of war field data contributes to intelligence analysis, evaluation of postmission outcomes, and creation of predictive models to enhance decision-making and strategic planning capabilities. However, old-school visualization methods are slow, expensive, and unscalable. Despite modern technologies in generating 3D point clouds, such as LIDAR and stereo sensors, monocular depth values based on deep learning can offer a faster and more detailed view of the environment, transforming single images into visual information for valuable insights. We propose a dedicated monocular depth reconstruction approach via deep learning techniques for 3D geovisualisation of satellite images. It introduces scalability in terrain reconstruction and data visualization. First, a dataset with more than 7,000 satellite images and associated digital elevation model (DEM) is created. It is based on high resolution optical and radar imageries collected from Planet and Copernicus, on which we fuse highresolution topographic data obtained using technologies such as LiDAR and the associated geographic coordinates. Second, we developed an imagery-DEM fusion strategy that combine feature maps from two encoder-decoder networks. One network is trained with radar and optical bands, while the other is trained with DEM features to compute dense 3D depth. Finally, we constructed a benchmark with sparse depth annotations to facilitate future research. To demonstrate the proposed method's versatility, we evaluated its performance on no annotated satellite images and implemented an enclosed environment useful for Geovisualisation applications. The algorithms were developed in Python 3.0, employing open-source computing libraries, i.e., Open3D, TensorFlow, and Pythorch3D. The proposed method provides fast and accurate decision-making with GIS for localization of troops, position of the enemy, terrain and climate conditions. This analysis enhances situational consciousness, enabling commanders to fine-tune the strategies and distribute the resources proficiently.

Keywords: depth, deep learning, geovisualisation, satellite images

Procedia PDF Downloads 8
13531 Numerical Modeling of Various Support Systems to Stabilize Deep Excavations

Authors: M. Abdallah

Abstract:

Urban development requires deep excavations near buildings and other structures. Deep excavation has become more a necessity for better utilization of space as the population of the world has dramatically increased. In Lebanon, some urban areas are very crowded and lack spaces for new buildings and underground projects, which makes the usage of underground space indispensable. In this paper, a numerical modeling is performed using the finite element method to study the deep excavation-diaphragm wall soil-structure interaction in the case of nonlinear soil behavior. The study is focused on a comparison of the results obtained using different support systems. Furthermore, a parametric study is performed according to the remoteness of the structure.

Keywords: deep excavation, ground anchors, interaction soil-structure, struts

Procedia PDF Downloads 414
13530 Leveraging Learning Analytics to Inform Learning Design in Higher Education

Authors: Mingming Jiang

Abstract:

This literature review aims to offer an overview of existing research on learning analytics and learning design, the alignment between the two, and how learning analytics has been leveraged to inform learning design in higher education. Current research suggests a need to create more alignment and integration between learning analytics and learning design in order to not only ground learning analytics on learning sciences but also enable data-driven decisions in learning design to improve learning outcomes. In addition, multiple conceptual frameworks have been proposed to enhance the synergy and alignment between learning analytics and learning design. Future research should explore this synergy further in the unique context of higher education, identifying learning analytics metrics in higher education that can offer insight into learning processes, evaluating the effect of learning analytics outcomes on learning design decision-making in higher education, and designing learning environments in higher education that make the capturing and deployment of learning analytics outcomes more efficient.

Keywords: learning analytics, learning design, big data in higher education, online learning environments

Procedia PDF Downloads 170
13529 A U-Net Based Architecture for Fast and Accurate Diagram Extraction

Authors: Revoti Prasad Bora, Saurabh Yadav, Nikita Katyal

Abstract:

In the context of educational data mining, the use case of extracting information from images containing both text and diagrams is of high importance. Hence, document analysis requires the extraction of diagrams from such images and processes the text and diagrams separately. To the author’s best knowledge, none among plenty of approaches for extracting tables, figures, etc., suffice the need for real-time processing with high accuracy as needed in multiple applications. In the education domain, diagrams can be of varied characteristics viz. line-based i.e. geometric diagrams, chemical bonds, mathematical formulas, etc. There are two broad categories of approaches that try to solve similar problems viz. traditional computer vision based approaches and deep learning approaches. The traditional computer vision based approaches mainly leverage connected components and distance transform based processing and hence perform well in very limited scenarios. The existing deep learning approaches either leverage YOLO or faster-RCNN architectures. These approaches suffer from a performance-accuracy tradeoff. This paper proposes a U-Net based architecture that formulates the diagram extraction as a segmentation problem. The proposed method provides similar accuracy with a much faster extraction time as compared to the mentioned state-of-the-art approaches. Further, the segmentation mask in this approach allows the extraction of diagrams of irregular shapes.

Keywords: computer vision, deep-learning, educational data mining, faster-RCNN, figure extraction, image segmentation, real-time document analysis, text extraction, U-Net, YOLO

Procedia PDF Downloads 137
13528 Nonparametric Sieve Estimation with Dependent Data: Application to Deep Neural Networks

Authors: Chad Brown

Abstract:

This paper establishes general conditions for the convergence rates of nonparametric sieve estimators with dependent data. We present two key results: one for nonstationary data and another for stationary mixing data. Previous theoretical results often lack practical applicability to deep neural networks (DNNs). Using these conditions, we derive convergence rates for DNN sieve estimators in nonparametric regression settings with both nonstationary and stationary mixing data. The DNN architectures considered adhere to current industry standards, featuring fully connected feedforward networks with rectified linear unit activation functions, unbounded weights, and a width and depth that grows with sample size.

Keywords: sieve extremum estimates, nonparametric estimation, deep learning, neural networks, rectified linear unit, nonstationary processes

Procedia PDF Downloads 41
13527 Speech Emotion Recognition: A DNN and LSTM Comparison in Single and Multiple Feature Application

Authors: Thiago Spilborghs Bueno Meyer, Plinio Thomaz Aquino Junior

Abstract:

Through speech, which privileges the functional and interactive nature of the text, it is possible to ascertain the spatiotemporal circumstances, the conditions of production and reception of the discourse, the explicit purposes such as informing, explaining, convincing, etc. These conditions allow bringing the interaction between humans closer to the human-robot interaction, making it natural and sensitive to information. However, it is not enough to understand what is said; it is necessary to recognize emotions for the desired interaction. The validity of the use of neural networks for feature selection and emotion recognition was verified. For this purpose, it is proposed the use of neural networks and comparison of models, such as recurrent neural networks and deep neural networks, in order to carry out the classification of emotions through speech signals to verify the quality of recognition. It is expected to enable the implementation of robots in a domestic environment, such as the HERA robot from the RoboFEI@Home team, which focuses on autonomous service robots for the domestic environment. Tests were performed using only the Mel-Frequency Cepstral Coefficients, as well as tests with several characteristics of Delta-MFCC, spectral contrast, and the Mel spectrogram. To carry out the training, validation and testing of the neural networks, the eNTERFACE’05 database was used, which has 42 speakers from 14 different nationalities speaking the English language. The data from the chosen database are videos that, for use in neural networks, were converted into audios. It was found as a result, a classification of 51,969% of correct answers when using the deep neural network, when the use of the recurrent neural network was verified, with the classification with accuracy equal to 44.09%. The results are more accurate when only the Mel-Frequency Cepstral Coefficients are used for the classification, using the classifier with the deep neural network, and in only one case, it is possible to observe a greater accuracy by the recurrent neural network, which occurs in the use of various features and setting 73 for batch size and 100 training epochs.

Keywords: emotion recognition, speech, deep learning, human-robot interaction, neural networks

Procedia PDF Downloads 170
13526 Multi-Impairment Compensation Based Deep Neural Networks for 16-QAM Coherent Optical Orthogonal Frequency Division Multiplexing System

Authors: Ying Han, Yuanxiang Chen, Yongtao Huang, Jia Fu, Kaile Li, Shangjing Lin, Jianguo Yu

Abstract:

In long-haul and high-speed optical transmission system, the orthogonal frequency division multiplexing (OFDM) signal suffers various linear and non-linear impairments. In recent years, researchers have proposed compensation schemes for specific impairment, and the effects are remarkable. However, different impairment compensation algorithms have caused an increase in transmission delay. With the widespread application of deep neural networks (DNN) in communication, multi-impairment compensation based on DNN will be a promising scheme. In this paper, we propose and apply DNN to compensate multi-impairment of 16-QAM coherent optical OFDM signal, thereby improving the performance of the transmission system. The trained DNN models are applied in the offline digital signal processing (DSP) module of the transmission system. The models can optimize the constellation mapping signals at the transmitter and compensate multi-impairment of the OFDM decoded signal at the receiver. Furthermore, the models reduce the peak to average power ratio (PAPR) of the transmitted OFDM signal and the bit error rate (BER) of the received signal. We verify the effectiveness of the proposed scheme for 16-QAM Coherent Optical OFDM signal and demonstrate and analyze transmission performance in different transmission scenarios. The experimental results show that the PAPR and BER of the transmission system are significantly reduced after using the trained DNN. It shows that the DNN with specific loss function and network structure can optimize the transmitted signal and learn the channel feature and compensate for multi-impairment in fiber transmission effectively.

Keywords: coherent optical OFDM, deep neural network, multi-impairment compensation, optical transmission

Procedia PDF Downloads 143