Search results for: optimum learning outcomes

Authors: Stepan Grebeniuk, Ersi Hodo, Henri Ruotsalainen, Paul Tavolato

Abstract:

Substation security plays an important role in the power delivery system. During the past years, there has been an increase in number of attacks on automation networks of the substations. In spite of that, there hasn’t been enough focus dedicated to the protection of such networks. Aiming to design a specialized anomaly detection system based on machine learning, in this paper we will discuss the IEC 60870-5-104 protocol that is used for communication between substation and control station and focus on the simulation of the substation traffic. Firstly, we will simulate the communication between substation slave and server. Secondly, we will compare the system's normal behavior and its behavior under the attack, in order to extract the right features which will be needed for building an anomaly detection system. Lastly, based on the features we will suggest the anomaly detection system for the asynchronous protocol IEC 60870-5-104.

Keywords: Anomaly detection, IEC-60870-5-104, Machine learning, Man-in-the-Middle attacks, Substation security

Procedia PDF Downloads 368

Authors: Bhalke D. G., Bormane D. S., Kharate G. K.

Abstract:

This paper presents classification of musical instrument using machine learning techniques. The classification has been carried out using temporal, spectral, cepstral and wavelet features. Detail feature analysis is carried out using separate and combined features. Further, instrument model has been developed using K-Nearest Neighbor and Support Vector Machine (SVM). Benchmarked McGill university database has been used to test the performance of the system. Experimental result shows that SVM performs better as compared to KNN classifier.

Keywords: feature extraction, SVM, KNN, musical instruments

Procedia PDF Downloads 480

Authors: Iris Vogeler, Rogerio Cichota, Armin Werner

Abstract:

Dairy farms are under pressure to increase productivity while reducing environmental impacts. Effective fertiliser management practices are critical to achieve this. Determination of optimum nitrogen (N) fertilisation rates which maximise pasture growth and minimise N losses is challenging due to variability in plant requirements and likely near-future supply of N by the soil. Remote sensing can be used for mapping N nutrition status of plants and to rapidly assess the spatial variability within a field. An algorithm is, however, lacking which relates the N status of the plants to the expected yield response to additions of N. The aim of this simulation study was to develop a multi-variate model for determining N fertilisation rate for a target percentage of the maximum achievable yield based on the pasture N concentration (ii) use of an algorithm for guiding fertilisation rates, and (iii) evaluation of the model regarding pasture yield and N losses, including N leaching, denitrification and volatilisation. A simulation study was carried out using the Agricultural Production Systems Simulator (APSIM). The simulations were done for an irrigated ryegrass pasture in the Canterbury region of New Zealand. A multi-variate model was developed and used to determine monthly required N fertilisation rates based on pasture N content prior to fertilisation and targets of 50, 75, 90 and 100% of the potential monthly yield. These monthly optimised fertilisation rules were evaluated by running APSIM for a ten-year period to provide yield and N loss estimates from both nonurine and urine affected areas. Comparison with typical fertilisation rates of 150 and 400 kg N/ha/year was also done. Assessment of pasture yield and leaching from fertiliser and urine patches indicated a large reduction in N losses when N fertilisation rates were controlled by the multi-variate model. However, the reduction in leaching losses was much smaller when taking into account the effects of urine patches. The proposed approach based on biophysical modelling to develop a multi-variate model for determining optimum N fertilisation rates dependent on pasture N content is very promising. Further analysis, under different environmental conditions and validation is required before the approach can be used to help adjust fertiliser management practices to temporal and spatial N demand based on the nitrogen status of the pasture.

Keywords: APSIM modelling, optimum N fertilization rate, pasture N content, ryegrass pasture, three dimensional surface response function.

Procedia PDF Downloads 130

Authors: Muhamad Sani Buang, Shahrul Azam Abdullah, Juri Saedon

Abstract:

There is not much effective guideline on development of design parameters selection on springback for advanced high strength steel sheet metal in U-channel process during cold forming process. This paper presents the development of predictive model for springback in U-channel process on advanced high strength steel sheet employing Response Surface Methodology (RSM). The experimental was performed on dual phase steel sheet, DP590 in U-channel forming process while design of experiment (DoE) approach was used to investigates the effects of four factors namely blank holder force (BHF), clearance (C) and punch travel (Tp) and rolling direction (R) were used as input parameters using two level values by applying Full Factorial design (24). From a statistical analysis of variant (ANOVA), result showed that blank holder force (BHF), clearance (C) and punch travel (Tp) displayed significant effect on springback of flange angle (β2) and wall opening angle (β1), while rolling direction (R) factor is insignificant. The significant parameters are optimized in order to reduce the springback behavior using Central Composite Design (CCD) in RSM and the optimum parameters were determined. A regression model for springback was developed. The effect of individual parameters and their response was also evaluated. The results obtained from optimum model are in agreement with the experimental values

Keywords: advance high strength steel, u-channel process, springback, design of experiment, optimization, response surface methodology (rsm)

Procedia PDF Downloads 541

Authors: Julian Wise

Abstract:

Newcrest Mining is one of the world’s top five gold and rare earth mining organizations by production, reserves and market capitalization in the world. This paper elaborates on the data acquisition processes employed by Newcrest in collaboration with Fortune 500 listed organization, Insight Enterprises, to standardize machine learning solutions which process data from over a hundred thousand distributed Internet of Things (IoT) devices located at mine sites globally. Through the utilization of software architecture cloud technologies and edge computing, the technological developments enable for standardized processes of machine learning applications to influence the strategic optimization of mineral processing. Target objectives of the machine learning optimizations include time savings on mineral processing, production efficiencies, risk identification, and increased production throughput. The data acquired and utilized for predictive modelling is processed through edge computing by resources collectively stored within a data lake. Being involved in the digital transformation has necessitated the standardization software architecture to manage the machine learning models submitted by vendors, to ensure effective automation and continuous improvements to the mineral process models. Operating at scale, the system processes hundreds of gigabytes of data per day from distributed mine sites across the globe, for the purposes of increased improved worker safety, and production efficiency through big data applications.

Keywords: mineral technology, big data, machine learning operations, data lake

Procedia PDF Downloads 112

Authors: Chalermpol Tapsai

Abstract:

The objectives of this research were to develop and find the efficiency of integrated online lessons by investigating the usage of online lessons, the relationship between learners’ background knowledge, and the achievement after learning with online lessons. The sample group in this study consisted of 97 students randomly selected from 121 students registering in 1/2012 at Trimitwittayaram Learning Center. The sample technique employed stratified sample technique of 4 groups according to their proficiency, i.e. high, moderate, low, and non-knowledge. The research instrument included online lessons in integration model on the topic of Java Programming, test after each lesson, the achievement test at the end of the course, and the questionnaires to find learners’ satisfaction. The results showed that the efficiency of online lessons was 90.20/89.18 with the achievement of after learning with the lessons higher than that before the lessons at the statistically significant level of 0.05. Moreover, the background knowledge of the learners on the programming showed the positive relationship with the achievement learning at the statistically significant level at 0.05. Learners with high background knowledge employed less exercises and samples than those with lower background knowledge. While learners with different background in the group of moderate and low did not show the significant difference in employing samples and exercises.

Keywords: integration model, online lessons, learners’ background knowledge, efficiency

Procedia PDF Downloads 359

Authors: Begona del Pino, Beatriz Prieto, Alberto Prieto

Abstract:

Two opposite approaches to teaching can be considered: in-class learning (teacher-oriented) versus virtual learning (student-oriented). The most known example of the latter is Massive Online Open Courses (MOOCs). Both methodologies have pros and cons. Nowadays there is an increasing trend towards combining both of them. Blending learning is considered a valuable tool for improving learning since it combines student-centred interactive e-learning and face to face instruction. The aim of this contribution is to exchange and share the experience and research results of a blended-learning project that took place in the University of Granada (Spain). The research objective was to prove how combining didactic resources of a MOOC with in-class teaching, interacting directly with students, can substantially improve academic results, as well as student acceptance. The proposed methodology is based on the use of flipped learning technics applied to the subject ‘Fundamentals of Computer Science’ of the first course of two degrees: Telecommunications Engineering, and Industrial Electronics. In this proposal, students acquire the theoretical knowledges at home through a MOOC platform, where they watch video-lectures, do self-evaluation tests, and use other academic multimedia online resources. Afterwards, they have to attend to in-class teaching where they do other activities in order to interact with teachers and the rest of students (discussing of the videos, solving of doubts and practical exercises, etc.), trying to overcome the disadvantages of self-regulated learning. The results are obtained through the grades of the students and their assessment of the blended experience, based on an opinion survey conducted at the end of the course. The major findings of the study are the following: The percentage of students passing the subject has grown from 53% (average from 2011 to 2014 using traditional learning methodology) to 76% (average from 2015 to 2018 using blended methodology). The average grade has improved from 5.20±1.99 to 6.38±1.66. The results of the opinion survey indicate that most students preferred blended methodology to traditional approaches, and positively valued both courses. In fact, 69% of students felt ‘quite’ or ‘very’ satisfied with the classroom activities; 65% of students preferred the flipped classroom methodology to traditional in-class lectures, and finally, 79% said they were ‘quite’ or ‘very’ satisfied with the course in general. The main conclusions of the experience are the improvement in academic results, as well as the highly satisfactory assessments obtained in the opinion surveys. The results confirm the huge potential of combining MOOCs in formal undergraduate studies with on-campus learning activities. Nevertheless, the results in terms of students’ participation and follow-up have a wide margin for improvement. The method is highly demanding for both students and teachers. As a recommendation, students must perform the assigned tasks with perseverance, every week, in order to take advantage of the face-to-face classes. This perseverance is precisely what needs to be promoted among students because it clearly brings about an improvement in learning.

Keywords: blended learning, educational paradigm, flipped classroom, flipped learning technologies, lessons learned, massive online open course, MOOC, teacher roles through technology

Procedia PDF Downloads 180

Authors: Fei Gao, Rodolfo C. Raga Jr.

Abstract:

This research proposal will ascertain the major risk factors for diabetes and to design a predictive model for risk assessment. The project aims to improve diabetes early detection and management by utilizing data science techniques, which may improve patient outcomes and healthcare efficiency. The phase relation values of each attribute were used to analyze and choose the attributes that might influence the examiner's survival probability using Diabetes Health Indicators Dataset from Kaggle’s data as the research data. We compare and evaluate eight machine learning algorithms. Our investigation begins with comprehensive data preprocessing, including feature engineering and dimensionality reduction, aimed at enhancing data quality. The dataset, comprising health indicators and medical data, serves as a foundation for training and testing these algorithms. A rigorous cross-validation process is applied, and we assess their performance using five key metrics like accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC-ROC). After analyzing the data characteristics, investigate their impact on the likelihood of diabetes and develop corresponding risk indicators.

Keywords: diabetes, risk factors, predictive model, risk assessment, data science techniques, early detection, data analysis, Kaggle

Procedia PDF Downloads 75

Authors: M. Rodionov, Z. Dedovets

Abstract:

The article analyses the composition and structure of the motivationally oriented methodological system of teaching mathematics (purpose, content, methods, forms, and means of teaching), viewed through the prism of the student as the subject of the learning process. Particular attention is paid to the problem of methods of teaching mathematics, which are represented in the form of an ordered triad of attributes corresponding to the selected characteristics. A systematic analysis of possible options and their methodological interpretation enriched existing ideas about known methods and technologies of training, and significantly expanded their nomenclature by including previously unstudied combinations of characteristics. In addition, examples outlined in this article illustrate the possibilities of enhancing the motivational capacity of a particular method or technology in the real learning practice of teaching mathematics through more free goal-setting and varying the conditions of the problem situations. The authors recommend the implementation of different strategies according to their characteristics in teaching and learning mathematics in secondary schools.

Keywords: education, methodological system, the teaching of mathematics, students motivation

Procedia PDF Downloads 354

Authors: Esther Mafunda

Abstract:

The present study interrogated the lecturer’s skills in supporting visually impaired students during the Covid-19 era at the University of Zimbabwe. It particularly assesses how the Covid-19 pandemic affected the learning experience of visually impaired students and which skills the lecturers possessed in order to assist the visually impaired students during online learning. Data was collected from lecturers and visually impaired students at the University of Zimbabwe Disability Resource Centre. Data was collected through the use of interviews and questionnaires. Using content analysis, it was established that visually impaired students faced challenges of lack of familiarity with the Moodle learning platform, marginalization, lack of professional training, and lack of training for parents and guardians. Lecturers faced challenges of lack of training, the curriculum, access, and technical know-how deficit. It was established that lecturers had to resort to social media platforms in order to assist visually impaired students. Visually impaired students also received assistance from their friends and family members. On the basis of the results of the research, it can be concluded that lecturers needed in-service training to be provided with the necessary skills and knowledge to teach students with visual impairments and provide quality education to students with visual impairments.

Keywords: visual impairment, disability, covid-19, inclusive learning

Procedia PDF Downloads 92

Authors: Roshanak Khodabakhsh Jolfaei, Javad Akbari Torkestani

Abstract:

As a cooperative environment for problem-solving, it is necessary that grids develop efficient job scheduling patterns with regard to their goals, domains and structure. Since the Grid environments facilitate distributed calculations, job scheduling appears in the form of a critical problem for the management of Grid sources that influences severely on the efficiency for the whole Grid environment. Due to the existence of some specifications such as sources dynamicity and conditions of the network in Grid, some algorithm should be presented to be adjustable and scalable with increasing the network growth. For this purpose, in this paper a job scheduling algorithm has been presented on the basis of learning automata in computational Grid which the performance of its results were compared with FPSO algorithm (Fuzzy Particle Swarm Optimization algorithm) and GJS algorithm (Grid Job Scheduling algorithm). The obtained numerical results indicated the superiority of suggested algorithm in comparison with FPSO and GJS. In addition, the obtained results classified FPSO and GJS in the second and third position respectively after the mentioned algorithm.

Keywords: computational grid, job scheduling, learning automata, dynamic scheduling

Procedia PDF Downloads 343

Authors: Adam Johnbull

Abstract:

The paper discusses ICT in Education as a quest for quality learning in the 21st century. Education is the key that unlock the door to development, without adequate education of the citizenry, the development of a nation becomes a sham. Information Communication Technologies (ICTs) has revolutionized the way people work today and are now transforming education systems. As a result, if schools train children in yesterday’s skills and technologies they may not be effective and fit in tomorrow’s world. This is a sufficient reason for ICT’s to win global recognition and attention and thus ensure desire quality in our school system. Thus, the purpose of the paper is to discuss amongst others, what is ICT. The roles of ICT’s in education, limitation and key challenges of integrating ICT to education in the enhancement of student learning and experiences in other to encourage policy makers, school administrators and teachers pay the required attention to integrate this technology in the education system. The paper concludes that regardless of all the limitation characterizing it. ICT benefit education system to provide quality education in the 21st century.

Keywords: ICTs, quest, information, global, sham, century

Procedia PDF Downloads 426

Authors: Taylan Kabbani, Ekrem Duman

Abstract:

The design of adaptive systems that take advantage of financial markets while reducing the risk can bring more stagnant wealth into the global market. However, most efforts made to generate successful deals in trading financial assets rely on Supervised Learning (SL), which suffered from various limitations. Deep Reinforcement Learning (DRL) offers to solve these drawbacks of SL approaches by combining the financial assets price "prediction" step and the "allocation" step of the portfolio in one unified process to produce fully autonomous systems capable of interacting with its environment to make optimal decisions through trial and error. In this paper, a continuous action space approach is adopted to give the trading agent the ability to gradually adjust the portfolio's positions with each time step (dynamically re-allocate investments), resulting in better agent-environment interaction and faster convergence of the learning process. In addition, the approach supports the managing of a portfolio with several assets instead of a single one. This work represents a novel DRL model to generate profitable trades in the stock market, effectively overcoming the limitations of supervised learning approaches. We formulate the trading problem, or what is referred to as The Agent Environment as Partially observed Markov Decision Process (POMDP) model, considering the constraints imposed by the stock market, such as liquidity and transaction costs. More specifically, we design an environment that simulates the real-world trading process by augmenting the state representation with ten different technical indicators and sentiment analysis of news articles for each stock. We then solve the formulated POMDP problem using the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm, which can learn policies in high-dimensional and continuous action spaces like those typically found in the stock market environment. From the point of view of stock market forecasting and the intelligent decision-making mechanism, this paper demonstrates the superiority of deep reinforcement learning in financial markets over other types of machine learning such as supervised learning and proves its credibility and advantages of strategic decision-making.

Keywords: the stock market, deep reinforcement learning, MDP, twin delayed deep deterministic policy gradient, sentiment analysis, technical indicators, autonomous agent

Procedia PDF Downloads 178

Authors: David Ming Liu, Benjamin Hirsch, Bashir Aden

Abstract:

Syndromic surveillance is to detect or predict disease outbreaks through the analysis of medical sources of data. Using social media data like tweets to do syndromic surveillance becomes more and more popular with the aid of open platform to collect data and the advantage of microblogging text and mobile geographic location features. In this paper, a Syndromic Surveillance Framework is presented with machine learning kernel using tweets data analytics. Influenza and the three cities Abu Dhabi, Al Ain and Dubai of United Arabic Emirates are used as the test disease and trial areas. Hospital cases data provided by the Health Authority of Abu Dhabi (HAAD) are used for the correlation purpose. In our model, Latent Dirichlet allocation (LDA) engine is adapted to do supervised learning classification and N-Fold cross validation confusion matrix are given as the simulation results with overall system recall 85.595% performance achieved.

Keywords: Syndromic surveillance, Tweets, Machine Learning, data mining, Latent Dirichlet allocation (LDA), Influenza

Procedia PDF Downloads 116

Authors: Giovanni Cicceri, Roberta Maisano, Nathalie Morey, Salvatore Distefano

Abstract:

The main goal of this paper is to present a solution for a water purification system based on an Environmental Internet of Things (EIoT) platform to monitor and control water quality and machine learning (ML) models to support decision making and speed up the processes of purification of water. A real case study has been implemented by deploying an EIoT platform and a network of devices, called Gramb meters and belonging to the Gramb project, on wastewater purification systems located in Calabria, south of Italy. The data thus collected are used to control the wastewater quality, detect anomalies and predict the behaviour of the purification system. To this extent, three different statistical and machine learning models have been adopted and thus compared: Autoregressive Integrated Moving Average (ARIMA), Long Short Term Memory (LSTM) autoencoder, and Facebook Prophet (FP). The results demonstrated that the ML solution (LSTM) out-perform classical statistical approaches (ARIMA, FP), in terms of both accuracy, efficiency and effectiveness in monitoring and controlling the wastewater purification processes.

Keywords: environmental internet of things, EIoT, machine learning, anomaly detection, environment monitoring

Procedia PDF Downloads 151

Authors: Liza Mariah Hj. Azahari

Abstract:

This paper will look at a case study on English Camp which was an activity coordinated at the Sultan Sharif Ali Islamic University in 2011. English Camp is a fun and motivation filled activity which brings students and teachers together outside of the classroom setting into a more diverse environment. It also enables teacher and students to gain proximate time together for a mutual purpose which is to explore the language in a more dynamic and relaxed way. First of all, the study will look into the background of English Camp, and how it was introduced and implemented from different contexts. Thereafter, it will explain the objectives of the English Camp coordinated at our university, UNISSA, and what types of activities were conducted. It will then evaluate the effectiveness of the camp as to what extent it managed to meet its motto, which was to foster dynamic interactive learning of English Language. To conclude, the paper presents a potential for further research on the topic as well as a guideline for educators who wish to coordinate the activity. Proposal for collaboration in this activity is further highlighted and encouraged within the paper for future implementation and endeavor.

Keywords: English camp, UNISSA, interactive learning, outside

Procedia PDF Downloads 569

Authors: Nicole V. Shifrin

Abstract:

Scheduling flexibility is becoming an increasingly available option for employees struggling to balance their work and life responsibilities, allowing employees to coordinate work schedules with their additional roles. The goal of such opportunities is to help employees manage the demands they face across domains of life by allowing employees to work from home, design their own work hours, take time off when necessary, along with various other scheduling accommodations. Organizations are also turning to utilizing scheduling flexibility to facilitate employee health and wellbeing through the reduction of stress and maximization of efficiency. The purpose of the present study is to investigate the effects of scheduling flexibility on employee health-related behaviors and outcomes through a synthesis of research. The current meta-analytic review of 19 samples within 16 studies with a total sample size of 20,707 employees examines the relationship between the degree of scheduling flexibility available to employees and the resulting health outcomes and exercise habits. The results demonstrate that reduced scheduling flexibility is associated with poorer health status, suggesting that schedule inflexibility can hinder employees’ ability to maintain and support their health. These findings hold practical implications for developing work schedules to promote employee health and health-related behaviors, such as eating well and exercising. Additionally, there was a positive association between increased scheduling flexibility and engagement in exercise, suggesting that employees with more flexible schedules exercise more frequently than those with less flexible schedules. A potential explanation for the resulting relationship is that flexible schedules leave employees more time due to shorter work days, shorter or eliminated commutes, etc. with which they can use to engage in healthy behaviors. These findings stress the importance of promoting job designs that facilitate employee engagement in healthy behaviors, which directly impact their overall health status. Implications for practice are discussed as well as future directions in examining the link between job design and employee health and well-being.

Keywords: exercise, health, meta-analysis, job design, scheduling flexibility

Procedia PDF Downloads 138

Authors: Amit Kumar, Amanpreet Chander, Ashish Sahani

Abstract:

This paper presents our recent work on real-time human activity detection based on the media pipe pipeline and machine learning algorithms. The proposed system can detect human activities, including running, jumping, squatting, bending to the left or right, and standing still. This is a robust solution for developing a yoga, dance, metaverse, and fitness application that checks for the correction of the pose without having any additional monitor like a personal trainer. MediaPipe solution offers an open-source cross-platform which utilizes a two-step detector-tracker ML pipeline for live detection of key landmarks on our body which can be used for motion data collection. The prediction of real-time poses uses a variety of machine learning techniques and different types of analysis. Without primarily relying on powerful desktop environments for inference, our method achieves real-time performance on the majority of contemporary mobile phones, desktops/laptops, Python, or even the web. Experimental results show that our method outperforms the existing method in terms of accuracy and real-time capability, achieving an accuracy of 99.92% on testing datasets.

Keywords: human activity detection, media pipe, machine learning, metaverse applications

Procedia PDF Downloads 179

Authors: A. O. Adebayo, G. A. Idowu, F. Odegbemi

Abstract:

Investigations on the recovery of nickel from aqueous solution using 2-hydroxy-5-nonyl- acetophenone oxime (LIX-84I) impregnated on activated charcoal was carried out. The LIX-84I was impregnated onto the pores of dried activated charcoal by dry method and optimum conditions for different equilibrium parameters (pH, adsorbent dosage, extractant concentration, agitation time and temperature) were determined using a simulated solution of nickel. The kinetics and adsorption isotherm studies were also evaluated. It was observed that the efficiency of recovery with LIX-84I impregnated on charcoal was dependent on the pH of the aqueous solution as there was little or no recovery at pH below 4. However, as the pH was raised, percentage recovery increases and peaked at pH 5.0. The recovery was found to increase with temperature up to 60ºC. Also it was observed that nickel adsorbed onto the loaded charcoal best at a lower concentration (0.1M) of the extractant when compared with higher concentrations. Similarly, a moderately low dosage (1 g) of the adsorbent showed better recovery than larger dosages. These optimum conditions were used to recover nickel from the leachate of Ni-MH batteries dissolved with sulphuric acid, and a 99.6% recovery was attained. Adsorption isotherm studies showed that the equilibrium data fitted best to Temkin model, with a negative value of constant, b (-1.017 J/mol) and a high correlation coefficient, R² of 0.9913. Kinetic studies showed that the adsorption process followed a pseudo-second order model. Thermodynamic parameter values (∆G⁰, ∆H⁰, and ∆S⁰) showed that the adsorption was endothermic and spontaneous. The impregnated charcoal appreciably recovered nickel using a relatively smaller volume of extractant than what is required in solvent extraction. Desorption studies showed that the loaded charcoal is reusable for three times, and so might be economical for nickel recovery from waste battery.

Keywords: charcoal, impregnated, LIX-84I, nickel, recovery

Procedia PDF Downloads 151

Authors: Haojie Wang, Xun Li, Rui Yin

Abstract:

In order to solve the memorization overfitting in the model-agnostic meta-learning MAML algorithm, a method of generating mutually exclusive tasks based on data augmentation is proposed. This method generates a mutex task by corresponding one feature of the data to multiple labels so that the generated mutex task is inconsistent with the data distribution in the initial dataset. Because generating mutex tasks for all data will produce a large number of invalid data and, in the worst case, lead to an exponential growth of computation, this paper also proposes a key data extraction method that only extract part of the data to generate the mutex task. The experiments show that the method of generating mutually exclusive tasks can effectively solve the memorization overfitting in the meta-learning MAML algorithm.

Keywords: mutex task generation, data augmentation, meta-learning, text classification.

Procedia PDF Downloads 143

Authors: Ali Nadi, Ali Edrissi

Abstract:

Relief demand and transportation links availability is the essential information that is needed for every natural disaster operation. This information is not in hand once a disaster strikes. Relief demand and network condition has been evaluated based on prediction method in related works. Nevertheless, prediction seems to be over or under estimated due to uncertainties and may lead to a failure operation. Therefore, in this paper a stochastic programming model is proposed to evaluate real-time relief demand and network condition at the onset of a natural disaster. To address the time sensitivity of the emergency response, the proposed model uses reinforcement learning for optimization of the total relief assessment time. The proposed model is tested on a real size network problem. The simulation results indicate that the proposed model performs well in the case of collecting real-time information.

Keywords: disaster management, real-time demand, reinforcement learning, relief demand

Procedia PDF Downloads 316

Authors: Parthiban Srinivasan

Abstract:

With recent advances in computational power and access to enough data in biosciences, artificial intelligence methods are increasingly being used in drug discovery research. These methods are essentially a series of advanced statistics based exercises that review the past to indicate the likely future. Our goal is to develop a model that accurately predicts biological activity and toxicity parameters for novel compounds. We have compiled a robust library of over 150,000 chemical compounds with different pharmacological properties from literature and public domain databases. The compounds are stored in simplified molecular-input line-entry system (SMILES), a commonly used text encoding for organic molecules. We utilize an automated process to generate an array of numerical descriptors (features) for each molecule. Redundant and irrelevant descriptors are eliminated iteratively. Our prediction engine is based on a portfolio of machine learning algorithms. We found Random Forest algorithm to be a better choice for this analysis. We captured non-linear relationship in the data and formed a prediction model with reasonable accuracy by averaging across a large number of randomized decision trees. Our next step is to apply deep neural network (DNN) algorithm to predict the biological activity and toxicity properties. We expect the DNN algorithm to give better results and improve the accuracy of the prediction. This presentation will review all these prominent machine learning and deep learning methods, our implementation protocols and discuss these techniques for their usefulness in biomedical and health informatics.

Keywords: deep learning, drug discovery, health informatics, machine learning, toxicity prediction

Procedia PDF Downloads 357

Authors: Faran Chaudhry, Anser Daud, Doris Braunstein, Oleg Safir, Allan Gross, Paul Kuzyk

Abstract:

Background: Acetabular reconstruction in the context of massive acetabular bone loss is challenging. In rare scenarios where the extent of bone loss precludes shell placement (cup-cage), reconstruction at our center consisted of a cage combined with highly porous metal augments. This study evaluates survivorship, complications, and functional outcomes using this technique. Methods: A total of 131 cup-cage implants (129 patients) were included in our retrospective review of revisions of total hip arthroplasty from January 2003 to January 2022. Among these cases, 100/131 (76.3%) were women, the mean age at surgery time was 68.7 years (range, 29.0 to 92.0; SD, 12.4), and the mean follow-up was 7.7 years (range, 0.02 to 20.3; SD, 5.1). Kaplan-Meier survivorship analysis was conducted with failure defined as revision surgery and/or failure of the cup-cage reconstruction. Results: A total of 30 implants (23%) reached the study endpoint involving all-cause revision. Overall survivorship was 74.8% at 10 years and 69.8% at 15 years. Reasons for revision included infection 12/131 (9.1%), dislocation 10/131 (7.6%), aseptic loosening of cup and/or cage 5/131 (3.8%), and aseptic loosening of the femoral stem 2/131 (1.5%). The mean LLD improved from 12.2 ± 15.9 mm to 3.9 ± 11.8 (p<0.05). The horizontal and vertical hip centres on plain film radiographs were significantly improved (p<0.05). Functionally, there was a decrease in the number of patients requiring the use of gait aids, with fewer patients (34, 25.9%) using a cane, walker, or wheelchair post-operatively compared to pre-operatively (58, 44%). There was a significant increase in the number of independent ambulators from 24 to 47 (36%). Conclusion: The cup-cage construct is a reliable treatment option for the treatment of various acetabular defects. There are favourable survivorship, clinical and radiographic outcomes, with a satisfactory complication rate.

Keywords: revision total hip arthroplasty, acetabular defect, pelvic discontinuity, trabecular metal augment, cup-cage

Procedia PDF Downloads 67

Authors: Oleg Vlasovets, Fabian Schaipp, Christian L. Mueller

Abstract:

we conduct a network analysis and predict the sex of 1000 participants based on ”connectome” - pairwise Pearson’s correlation across 436 brain parcels. We solve the non-smooth convex optimization problem, known under the name of Graphical Lasso, where the solution includes a low-rank component. With this solution and machine learning model for a sex prediction, we explain the brain parcels-sex connectivity patterns.

Keywords: network analysis, neuroscience, machine learning, optimization

Procedia PDF Downloads 147

Authors: Boon Yih Mah

Abstract:

Web-based Cognitive Writing Instruction (WeCWI) is a hybrid e-framework that consolidates instructional design and language development towards the development of a web-based instruction (WBI). WeCWI divides instructional design into macro and micro perspectives. In macro perspective, a 21st century educator is encouraged to disseminate knowledge and share ideas with in-class and global learners. By leveraging the virtue of technology, WeCWI aims to transform the educator into an aggregator, curator, publisher, social networker and finally, a web-based instructor. Since the most notable contribution of integrating technology is being a tool of teaching as well as a stimulus for learning, WeCWI focuses on the use of contemporary web tools based on the multiple roles played by the 21st century educator. The micro perspective draws attention to the pedagogical approaches focussing on three main aspects: reading, discussion, and writing. With the effective use of pedagogical approaches, technology adds new dimensions and expands the bounds of learning capacity. Lastly, WeCWI also imparts the fundamental theoretical concepts for web-based instructors’ awareness such as interactionism, e-learning interactional-based model, computer-mediated communication (CMC), cognitive theories, and learning style model.

Keywords: web-based cognitive writing instruction, WeCWI, instructional design, e-framework, web-based instructor

Procedia PDF Downloads 439

Authors: Zenah Gheblawi, Susan Armijo-Olivo, Elisa B. Pelai, Vaishali Sharma, Musa Tashfeen, Angela Fung, Francisca Claveria

Abstract:

Due to physiological differences between men and women, pain is experienced differently between the two sexes. Chronic pain disorders, notably temporomandibular disorders (TMDs), disproportionately affect women in diagnosis, and pain severity in opposition of their male counterparts. TMDs are a type of musculoskeletal disorder that target the masticatory muscles, temporalis muscle, and temporomandibular joints, causing considerable orofacial pain which can usually be referred to the neck and back. Therapeutic methods are scarce, and are not TMD-centered, with the latest research suggesting that subjects with chronic musculoskeletal pain disorders have abnormal alterations in the grey matter of their brains which can be remedied with exercise, and thus, decreasing the pain experienced. The aim of the study is to investigate the effects of exercise therapy in TMD female patients experiencing chronic jaw pain and to assess the consequential effects on brain activity. In a randomized controlled trial, the effectiveness of an exercise program to improve brain alterations and clinical outcomes in women with TMD pain will be tested. Women with chronic TMD pain will be randomized to either an intervention arm or a placebo control group. Women in the intervention arm will receive 8 weeks of progressive exercise of motor control training using visual feedback (MCTF) of the cervical muscles, twice per week. Women in the placebo arm will receive innocuous transcutaneous electrical nerve stimulation during 8 weeks as well. The primary outcomes will be changes in 1) pain, measured with the Visual Analogue Scale, 2) brain structure and networks, measured by fractional anisotropy (brain structure) and the blood-oxygen level dependent signal (brain networks). Outcomes will be measured at baseline, after 8 weeks of treatment, and 4 months after treatment ends and will determine effectiveness of MCTF in managing TMD, through improved clinical outcomes. Results will directly inform and guide clinicians in prescribing more effective interventions for women with TMD. This study is underway, and no results are available at this point. The results of this study will have substantial implications on the advancement in understanding the scope of plasticity the brain has in regards with pain, and how it can be used to improve the treatment and pain of women with TMD, and more generally, other musculoskeletal disorders.

Keywords: exercise therapy, musculoskeletal disorders, physical therapy, rehabilitation, tempomandibular disorders

Procedia PDF Downloads 292

Authors: M. Zezou

Abstract:

This paper focuses on the design of a Massive Open Online Course (MOOC) about Content Language Integrated Learning (CLIL) and more specifically about how teachers can use CLIL as an educational approach incorporating technology in their teaching as well. All the four weeks of the MOOC will be presented and a step-by-step analysis of each lesson will be offered. Additionally, the paper includes detailed lesson plans about CLIL lessons with proposed CLIL activities and games in which technology plays a central part. The MOOC is structured based on certain criteria, in order to ensure success, as well as a positive experience that the learners need to have after completing this MOOC. It addresses to all language teachers who would like to implement CLIL into their teaching. In other words, it presents the methodology that needs to be followed so as to successfully carry out a CLIL lesson and achieve the learning objectives set at the beginning of the course. Firstly, in this paper, it is very important to give the definitions of MOOCs and LMOOCs, as well as to explore the difference between a structure-based MOOC (xMOOC) and a connectivist MOOC (cMOOC) and present the criteria of a successful MOOC. Moreover, the notion of CLIL will be explored, as it is necessary to fully understand this concept before moving on to the design of the MOOC. Onwards, the four weeks of the MOOC will be introduced as well as lesson plans will be presented: The type of the activities, the aims of each activity and the methodology that teachers have to follow. Emphasis will be placed on the role of technology in foreign language learning and on the ways in which we can involve technology in teaching a foreign language. Final remarks will be made and a summary of the main points will be offered at the end.

Keywords: CLIL, cMOOC, lesson plan, LMOOC, MOOC criteria, MOOC, technology, xMOOC

Procedia PDF Downloads 194

Authors: Md. Shaiful Islam, Mahani Bt. Stapa

Abstract:

To help learners overcome their struggle in developing proficiency in spoken English, self-regulated learning strategies seem to be promising. Students in the private universities in Bangladesh are expected to communicate with the teachers, peers, and staff members in English, but most of them suffer from their inadequate oral communicative competence in English. To address this problem, the researchers adopted a qualitative research approach to answer the research questions. They employed the learner diary method to collect data from the first-semester undergraduate students of a reputed private university in Bangladesh who were involved in writing weekly diaries about their use of self-regulated learning strategies to improve speaking in an English speaking course. The learners were provided with prompts for writing the diaries. The thematic analysis method was applied to analyze the entries of the diaries for the identification of themes. Seven strategies related to the effectiveness of SRL for the improvement of spoken English were identified from the data, and they include goal-setting, strategic planning, identifying the sources of self-motivation, help-seeking, environmental restructuring, self-monitoring, and self-evaluation. However, the students reported in their diaries that they faced challenges that impeded their SRL strategy use. Five challenges were identified, and they entail the complex nature of SRL, lack of literacy on SRL, teachers’ preference for controlling the class, learners’ past habit of learning, and students’ addiction to gadgets. The implications the study addresses include revising the syllabus and curriculum, facilitating SRL training for students and teachers, and integrating SRL in the lessons.

Keywords: private university in Bangladesh, proficiency, self-regulated learning, spoken English

Procedia PDF Downloads 160

Authors: Ghazi Alotaibi

Abstract:

The current study examined and described the views of secondary school students and their parents on their reasons for seeking private tutors in English. These views were obtained through two group interviews with the students and parents separately. Several causes were brought up during the two interviews. These causes included difficulty of the English language, weak teacher performance, the need to pass exams with high marks, lack of parents’ follow-up of student school performance, social pressure, variability in student comprehension levels at school, weak English foundation in previous school years, repeated student absence from school, large classes, as well as English teachers’ heavy teaching loads. The study started with a description of the EFL educational system in Saudi Arabia and concluded with recommendations for the improvement of the school learning environment.

Keywords: english, learning difficulty, private tutoring, Saudi, teaching practices, learning environment

Procedia PDF Downloads 456

Authors: Migena Mana, Ahmed Khalid Syed, Abdul Malik, Nikhil Cherian

Abstract:

In recent years, advancements in deep learning enabled researchers to tackle the problem of self-driving cars. Car companies use huge datasets to train their deep learning models to make autonomous cars a reality. However, this approach has certain drawbacks in that the state space of possible actions for a car is so huge that there cannot be a dataset for every possible road scenario. To overcome this problem, the concept of reinforcement learning (RL) is being investigated in this research. Since the problem of autonomous driving can be modeled in a simulation, it lends itself naturally to the domain of reinforcement learning. The advantage of this approach is that we can model different and complex road scenarios in a simulation without having to deploy in the real world. The autonomous agent can learn to drive by finding the optimal policy. This learned model can then be easily deployed in a real-world setting. In this project, we focus on three RL algorithms: Q-learning, Deep Deterministic Policy Gradient (DDPG), and Proximal Policy Optimization (PPO). To model the environment, we have used TORCS (The Open Racing Car Simulator), which provides us with a strong foundation to test our model. The inputs to the algorithms are the sensor data provided by the simulator such as velocity, distance from side pavement, etc. The outcome of this research project is a comparative analysis of these algorithms. Based on the comparison, the PPO algorithm gives the best results. When using PPO algorithm, the reward is greater, and the acceleration, steering angle and braking are more stable compared to the other algorithms, which means that the agent learns to drive in a better and more efficient way in this case. Additionally, we have come up with a dataset taken from the training of the agent with DDPG and PPO algorithms. It contains all the steps of the agent during one full training in the form: (all input values, acceleration, steering angle, break, loss, reward). This study can serve as a base for further complex road scenarios. Furthermore, it can be enlarged in the field of computer vision, using the images to find the best policy.

Keywords: autonomous driving, DDPG (deep deterministic policy gradient), PPO (proximal policy optimization), reinforcement learning

Procedia PDF Downloads 148