Search results for: learning loss

Authors: Meruyert Serik, Nassipzhan Duisegaliyeva, Danara Tleumagambetova, Madina Ermaganbetova

Abstract:

This research paper explores the process of creating a proctoring system by evaluating the implementation of practical face recognition algorithms. Students of educational programs reviewed the research work "6B01511-Computer Science", "7M01511-Computer Science", "7M01525- STEM Education," and "8D01511-Computer Science" of Eurasian National University named after L.N. Gumilyov. As an outcome, a proctoring system will be created, enabling the conduction of tests and ensuring academic integrity checks within the system. Due to the correct operation of the system, test works are carried out. The result of the creation of the proctoring system will be the basis for the automation of the informational, educational portal developed by machine learning.

Keywords: artificial intelligence, education portal, face recognition, machine learning, proctoring

Procedia PDF Downloads 97

Authors: Mahdi Bazarganigilani

Abstract:

Computer-aided diagnosis (CAD) systems can play crucial roles in diagnosing crucial diseases such as breast cancer at the earliest. In this paper, a CAD system for the diagnosis of breast cancer was introduced and evaluated. This CAD system was developed by using spatio-temporal analysis of data on a set of consecutive thermographic images by employing wavelet transformation. By using this analysis, a very accurate machine learning model using random forest was obtained. The final results showed a promising accuracy of 91% in terms of the F1 measure indicator among 200 patients' sample data. The CAD system was further extended to obtain a detailed analysis of the effect of smaller sub-areas of each breast on the occurrence of cancer.

Keywords: computer-aided diagnosis systems, thermographic analysis, spatio-temporal analysis, image processing, machine learning

Procedia PDF Downloads 198

Authors: Assel Jaxylykova, Alexnder Pak

Abstract:

This study utilized a Self-Organizing Map (SOM) for unsupervised learning on the CONLL-2003 dataset for Named Entity Recognition (NER). The process involved encoding words into 300-dimensional vectors using FastText. These vectors were input into a SOM grid, where training adjusted node weights to minimize distances. The SOM provided a topological representation for identifying and clustering named entities, demonstrating its efficacy without labeled examples. Results showed an F1-measure of 0.86, highlighting SOM's viability. Although some methods achieve higher F1 measures, SOM eliminates the need for labeled data, offering a scalable and efficient alternative. The SOM's ability to uncover hidden patterns provides insights that could enhance existing supervised methods. Further investigation into potential limitations and optimization strategies is suggested to maximize benefits.

Keywords: named entity recognition natural, language processing, self-organizing map, CONLL-2003, semantics

Procedia PDF Downloads 7

Authors: Varun Agarwal

Abstract:

Introduction: With the advent of personalized medicine, histopathological review of whole slide images (WSIs) for cancer diagnosis presents an exceedingly time-consuming, complex task. Specifically, detecting metastatic regions in WSIs of sentinel lymph node biopsies necessitates a full-scanned, holistic evaluation of the image. Thus, digital pathology, low-level image manipulation algorithms, and machine learning provide significant advancements in improving the efficiency and accuracy of WSI analysis. Using Camelyon16 data, this paper proposes a deep learning pipeline to automate and ameliorate breast cancer metastasis localization and WSI classification. Methodology: The model broadly follows five stages -region of interest detection, WSI partitioning into image tiles, convolutional neural network (CNN) image-segment classifications, probabilistic mapping of tumor localizations, and further processing for whole WSI classification. Transfer learning is applied to the task, with the implementation of Inception-ResNetV2 - an effective CNN classifier that uses residual connections to enhance feature representation, adding convolved outputs in the inception unit to the proceeding input data. Moreover, in order to augment the performance of the transfer learning CNN, a stack of convolutional denoising autoencoders (CDAE) is applied to produce embeddings that enrich image representation. Through a saliency-detection algorithm, visual training segments are generated, which are then processed through a denoising autoencoder -primarily consisting of convolutional, leaky rectified linear unit, and batch normalization layers- and subsequently a contrast-normalization function. A spatial pyramid pooling algorithm extracts the key features from the processed image, creating a viable feature map for the CNN that minimizes spatial resolution and noise. Results and Conclusion: The simplified and effective architecture of the fine-tuned transfer learning Inception-ResNetV2 network enhanced with the CDAE stack yields state of the art performance in WSI classification and tumor localization, achieving AUC scores of 0.947 and 0.753, respectively. The convolutional feature retention and compilation with the residual connections to inception units synergized with the input denoising algorithm enable the pipeline to serve as an effective, efficient tool in the histopathological review of WSIs.

Keywords: breast cancer, convolutional neural networks, metastasis mapping, whole slide images

Procedia PDF Downloads 115

Authors: Deeman Yousif Mahmood, Mohammed Abdullah Hussein

Abstract:

The goal of a network-based intrusion detection system is to classify activities of network traffics into two major categories: normal and attack (intrusive) activities. Nowadays, data mining and machine learning plays an important role in many sciences; including intrusion detection system (IDS) using both supervised and unsupervised techniques. However, one of the essential steps of data mining is feature selection that helps in improving the efficiency, performance and prediction rate of proposed approach. This paper applies unsupervised K-means clustering algorithm with information gain (IG) for feature selection and reduction to build a network intrusion detection system. For our experimental analysis, we have used the new NSL-KDD dataset, which is a modified dataset for KDDCup 1999 intrusion detection benchmark dataset. With a split of 60.0% for the training set and the remainder for the testing set, a 2 class classifications have been implemented (Normal, Attack). Weka framework which is a java based open source software consists of a collection of machine learning algorithms for data mining tasks has been used in the testing process. The experimental results show that the proposed approach is very accurate with low false positive rate and high true positive rate and it takes less learning time in comparison with using the full features of the dataset with the same algorithm.

Keywords: information gain (IG), intrusion detection system (IDS), k-means clustering, Weka

Procedia PDF Downloads 281

Authors: Usama Shahid

Abstract:

Be it written documents, news columns, or students' essays, verifying the content can be a time-consuming task. Apart from the spelling and grammar mistakes, the proofreader is also supposed to verify whether the content included in the essay or document is relevant or not. The irrelevant content in any document or essay is referred to as off-topic text and in this paper, we will address the problem of off-topic text detection from a document using machine learning techniques. Our study aims to identify the off-topic content from a document using Echo state network model and we will also compare data with other models. The previous study uses Convolutional Neural Networks and TFIDF to detect off-topic text. We will rearrange the existing datasets and take new classifiers along with new word embeddings and implement them on existing and new datasets in order to compare the results with the previously existing CNN model.

Keywords: off topic, text detection, eco state network, machine learning

Procedia PDF Downloads 66

Authors: Tatyana Gavrilova, Vadim Onufriev

Abstract:

Knowledge-based systems’ design requires the developer’s owning the advanced analytical skills. The efficient development of that skills within university courses needs a deep understanding of main pitfalls and drawbacks, which students usually make during their analytical work in form of visual modeling. Thus, it was necessary to hold an analysis of 5-th year students’ learning exercises within courses of 'Intelligent systems' and 'Knowledge engineering' in Saint-Petersburg Polytechnic University. The analysis shows that both lack of system thinking skills and methodological mistakes in course design cause the errors that are discussed in the paper. The conclusion contains an exploration of the issues and topics necessary and sufficient for the implementation of the improved practices in educational design for future curricula of teaching programs.

Keywords: knowledge based systems, knowledge engineering, students’ errors, visual modeling

Procedia PDF Downloads 297

Authors: Dr. Rebecca Kimelman

Abstract:

Teaching students to self-advocate has become a central topic in special education literature and practice. However, many special education programs do not address this important skill area. To this end, I created and implemented the Self Advocacy Made Easy (S.A.M.E.) program of instruction, intended to enhance the self-advocacy skills of young adults with mild to moderate disabilities. The effectiveness of S.A.M.E., the degree to which self-advocacy skills were acquired and demonstrated by the students, the level of parental support, and the impact of culture on the process, and teachers’ beliefs and attitudes about the role of self-advocacy skills for their students were measured using action research that employed mixed methodology. Conducted at an overseas American International School, this action research study sought answers to these questions by providing an in-depth portrayal of the S.A.M.E. program, as well as the attitudes and perceptions of the stakeholders involved in the study (thirteen students, their parents, teachers and counsellors). The findings of this study were very positive. The S.A.M.E. program was found to be a valid and valuable instructional tool for teaching self-advocacy skills to students with learning disabilities and ADHD. The study showed participation in the S.A.M.E. program led to an increased understanding of the important elements of self-advocacy, an increase in students’ skills and abilities to self-advocate, and a positive increase in students’ feelings about themselves. Inclusion in the Student-Led IEP meetings, an authentic student assessment within the S.A.M.E. program, also yielded encouraging results, including a higher level of ownership of one’s profile and learning needs, a higher level of student engagement and participation in the IEP meeting, and a growing student awareness of the relevance of the document and the IEP process to their lives. Without exception, every parent believed that participating in the Student-Led IEP led to a growth in confidence in their children, including that it taught them how to ‘own’ their disability and an improvement in their communication skills. Teachers and counsellors that participated in the study felt the program was worthwhile, and led to an increase in the students’ ability to acknowledge their learning profile and to identify and request the accommodations (such as extended time or use of a calculator) they need to overcome or work around their disability. The implications for further research are many, and include an examination of the degree to which participation in S.A.M.E. fosters student achievement, the long-term effects of participation in the program, and the degree to which student participation in the Student-Led IEP meeting increases parents’ level of understanding and involvement.

Keywords: self-advocacy, learning disabilities, ADHD, student-led IEP process

Procedia PDF Downloads 41

Authors: Nisha Thapa, Ram Prasad Mainali, Prakriti Chand

Abstract:

Broadleaf mustard is a commercialized leafy vegetable of Nepal. However, its utilization is hindered in terms of production and productivity due to the high intensity of insects, pests, and diseases causing great loss. The plant protection part of the crop’s disease and damage intensity has not been studied much from research perspectives in Nepal. The research aimed to evaluate broadleaf mustard genotypes for resistance against different diseases. A total of 35 native genotypes of broadleaf mustard were screened at weekly intervals by scoring the plants for ten weeks. Five different diseases, such as Rhizoctonia root rot, Alternaria blight, black rot, turnip mosaic virus disease, and white rust, were reported from the broad leaf mustard genotypes. Out of 35 genotypes, 23 genotypes were found with very high Rhizoctonia Root Rot severity, whereas 8 genotypes showed very high Alternaria blight severity. Likewise, 3 genotypes were found with high Black rot severity, and 1 genotype was found with very high Turnip mosaic virus disease incidence. Similarly, 2 genotypes were found to have very high White rust severity. Among the disease of national importance, Rhizoctonia root rot was found to be the most severe disease with the greatest loss. Broadleaf mustard genotypes like Rato Rayo, CO 1002, and CO 11007 showed average to the high level of field resistance; therefore, these genotypes should be used, conserved, and stored in a mustard improvement program as the disease resistance quality or susceptibility of these genotypes can be helpful for seed producing farmers, companies and other stakeholders through varietal improvement and developmental works that further aids in sustainable disease management of the vegetable.

Keywords: genotype, disease resistance, Rhizoctonia root rot severity, varietal improvement

Procedia PDF Downloads 64

Authors: Amah Nnachi Lofty, Oyefeso Olufemi, Ibiam Udu Ama

Abstract:

New and more powerful communications technologies continue to emerge at a rapid pace and their uses in education are widespread and the impact remarkable in the developing societies. This study investigates Mobile Collaboration Learning Technique (MCLT) on learners’ outcome among students in tertiary institutions of developing nations (a case of Nigeria students). It examines the significance of retention achievement scores of students taught using mobile collaboration and conventional method. The sample consisted of 120 students using Stratified random sampling method. Three research questions and hypotheses were formulated, and tested at a 0.05 level of significance. A student achievement test (SAT) was made of 40 items of multiple-choice objective type, developed and validated for data collection by professionals. The SAT was administered to students as pre-test and post-test. The data were analyzed using t-test statistic to test the hypotheses. The result indicated that students taught using MCLT performed significantly better than their counterparts using the conventional method of instruction. Also, there was no significant difference in the post-test performance scores of male and female students taught using MCLT. Based on the findings, the following recommendations was made that: Mobile collaboration system be encouraged in the institutions to boost knowledge sharing among learners, workshop and trainings should be organized to train teachers on the use of this technique and that schools and government should formulate policies and procedures towards responsible use of MCLT.

Keywords: education, communication, learning, mobile collaboration, technology

Procedia PDF Downloads 206

Authors: Bhaskar Basu

Abstract:

Universities and higher education institutes are finding it increasingly difficult to engage students fruitfully through traditional pedagogic tools. Web 2.0 technologies comprising social networking sites (SNSs) offer a platform for students to collaborate and share information, thereby enhancing their learning experience. Despite the potential and reach of SNSs, its use has been limited in academic settings promoting higher education. The purpose of this paper is to assess the perception of social networking sites among business school students in India and analyze its role in enhancing quality of student experiences in a business school leading to the proposal of an agenda for future research. In this study, more than 300 students of a reputed business school were involved in a survey of their preferences of different social networking sites and their perceptions and attitudes towards these sites. A questionnaire with three major sections was designed, validated and distributed among  a sample of students, the research method being descriptive in nature. Crucial questions were addressed to the students concerning time commitment, reasons for usage, nature of interaction on these sites, and the propensity to share information leading to direct and indirect modes of learning. It was further supplemented with focus group discussion to analyze the findings. The paper notes the resistance in the adoption of new technology by a section of business school faculty, who are staunch supporters of the classical “face-to-face” instruction. In conclusion, social networking sites like Facebook and LinkedIn provide new avenues for students to express themselves and to interact with one another. Universities could take advantage of the new ways  in which students are communicating with one another. Although interactive educational options such as Moodle exist, social networking sites are rarely used for academic purposes. Using this medium opens new ways of academically-oriented interactions where faculty could discover more about students' interests, and students, in turn, might express and develop more intellectual facets of their lives. hitherto unknown intellectual facets.  This study also throws up the enormous potential of mobile phones as a tool for “blended learning” in business schools going forward.

Keywords: business school, India, learning, social media, social networking, university

Procedia PDF Downloads 246

Authors: Krisztina Bohacs, Klaudia Markus

Abstract:

To author’s best knowledge we are in absence of studies on cognitive program evaluation and we are certainly short of programs that prove to have high effect sizes with strong retention results. The purpose of our study was to investigate the effectiveness of a comprehensive cognitive training program, namely BrainRx. This cognitive rehabilitation program target and remediate seven core cognitive skills and related systems of sub-skills through repeated engagement in game-like mental procedures delivered one-on-one by a clinician, supplemented by digital training. A larger sample of children with learning disability were given pretest and post-test cognitive assessments. The experimental group completed a twenty-week cognitive training program in a BrainRx center. A matched control group received another twenty-week intervention with Feuerstein’s Instrumental Enrichment programs. A second matched control group did not receive training. As for pre- and post-test, we used a general intelligence test to assess IQ and a computer-based test battery for assessing cognition across the lifespan. Multiple regression analyses indicated that the experimental BrainRx treatment group had statistically significant higher outcomes in attention, working memory, processing speed, logic and reasoning, auditory processing, visual processing and long-term memory compared to the non-treatment control group with very large effect sizes. With the exception of logic and reasoning, the BrainRx treatment group realized significantly greater gains in six of the above given seven cognitive measures compared to the Feuerstein control group. Our one-year retention measures showed that all the cognitive training gains were above ninety percent with the greatest retention skills in visual processing, auditory processing, logic, and reasoning. The BrainRx program may be an effective tool to establish long-term cognitive changes in case of students with learning disabilities. Recommendations are made for treatment centers and special education institutions on the cognitive training of students with special needs. The importance of our study is that targeted, systematic, progressively loaded and intensive brain training approach may significantly change learning disabilities.

Keywords: cognitive rehabilitation training, cognitive skills, learning disability, permanent structural cognitive changes

Procedia PDF Downloads 193

Authors: George Zhou, Yunchan Chen, Candace Chien

Abstract:

Kidney replacement therapy is the current standard of care for end-stage renal diseases. In-center or home hemodialysis remains an integral component of the therapeutic regimen. Arteriovenous fistulas (AVF) make up the vascular circuit through which blood is filtered and returned. Naturally, AVF patency determines whether adequate clearance and filtration can be achieved and directly influences clinical outcomes. Our aim was to build a deep learning model for automated AVF stenosis screening based on the sound of blood flow through the AVF. A total of 311 patients with AVF were enrolled in this study. Blood flow sounds were collected using a digital stethoscope. For each patient, blood flow sounds were collected at 6 different locations along the patient’s AVF. The 6 locations are artery, anastomosis, distal vein, middle vein, proximal vein, and venous arch. A total of 1866 sounds were collected. The blood flow sounds are labeled as “patent” (normal) or “stenotic” (abnormal). The labels are validated from concurrent ultrasound. Our dataset included 1527 “patent” and 339 “stenotic” sounds. We show that blood flow sounds vary significantly along the AVF. For example, the blood flow sound is loudest at the anastomosis site and softest at the cephalic arch. Contextualizing the sound with location metadata significantly improves classification performance. How to encode and incorporate categorical metadata is an active area of research1. Herein, we study ordinal (i.e., integer) encoding schemes. The numerical representation is concatenated to the flattened feature vector. We train a vision transformer (ViT) on spectrogram image representations of the sound and demonstrate that using scalar multiples of our integer encodings improves classification performance. Models are evaluated using a 10-fold cross-validation procedure. The baseline performance of our ViT without any location metadata achieves an AuROC and AuPRC of 0.68 ± 0.05 and 0.28 ± 0.09, respectively. Using the following encodings of Artery:0; Arch: 1; Proximal: 2; Middle: 3; Distal 4: Anastomosis: 5, the ViT achieves an AuROC and AuPRC of 0.69 ± 0.06 and 0.30 ± 0.10, respectively. Using the following encodings of Artery:0; Arch: 10; Proximal: 20; Middle: 30; Distal 40: Anastomosis: 50, the ViT achieves an AuROC and AuPRC of 0.74 ± 0.06 and 0.38 ± 0.10, respectively. Using the following encodings of Artery:0; Arch: 100; Proximal: 200; Middle: 300; Distal 400: Anastomosis: 500, the ViT achieves an AuROC and AuPRC of 0.78 ± 0.06 and 0.43 ± 0.11. respectively. Interestingly, we see that using increasing scalar multiples of our integer encoding scheme (i.e., encoding “venous arch” as 1,10,100) results in progressively improved performance. In theory, the integer values do not matter since we are optimizing the same loss function; the model can learn to increase or decrease the weights associated with location encodings and converge on the same solution. However, in the setting of limited data and computation resources, increasing the importance at initialization either leads to faster convergence or helps the model escape a local minimum.

Keywords: arteriovenous fistula, blood flow sounds, metadata encoding, deep learning

Procedia PDF Downloads 68

Authors: Najlaa Yousef Qusti, Steven J. Brookes, Paul A. Brunton

Abstract:

Background: Tooth discoloration can be an aesthetic problem, and tooth whitening using carbamide peroxide bleaching agents are a popular treatment option. However, there are concerns about possible adverse effects such as demineralisation of the bleached enamel; however, the cause of this demineralisation is unclear. Introduction: Teeth can become stained or discoloured over time. Tooth whitening is an aesthetic solution for tooth discoloration. Bleaching solutions of 10% carbamide peroxide (CP) have become the standard agent used in dentist-prescribed and home-applied ’vital bleaching techniques’. These materials release hydrogen peroxide (H₂O₂), the active whitening agent. However, there is controversy in the literature regarding the effect of bleaching agents on enamel integrity and enamel mineral content. The purpose of this study was to establish if carbamide peroxide bleaching agents affect the acid solubility of enamel (i.e., make teeth more prone to demineralisation). Materials and Methods: Twelve human premolar teeth were sectioned longitudinally along the midline and varnished to leave the natural enamel surface exposed. The baseline behavior of each tooth half in relation to its demineralisation in acid was established by sequential exposure to 4 vials containing 1ml of 10mM acetic acid (1 minute/vial). This was followed by exposure to 10% CP for 8 hours. After washing in distilled water, the tooth half was sequentially exposed to 4 further vials containing acid to test if the acid susceptibility of the enamel had been affected. The corresponding tooth half acted as a control and was exposed to distilled water instead of CP. The mineral loss was determined by measuring [Ca²⁺] and [PO₄³⁻] released in each vial using a calcium ion-selective electrode and the phosphomolybdenum blue method, respectively. The effect of bleaching on the tooth surfaces was also examined using SEM. Results: Exposure to carbamide peroxide did not significantly alter the susceptibility of enamel to acid attack, and SEM of the enamel surface revealed a slight alteration in surface appearance. SEM images of the control enamel surface showed a flat enamel surface with some shallow pits, whereas the bleached enamel appeared with an increase in surface porosity and some areas of mild erosion. Conclusions: Exposure to H₂O₂ equivalent to 10% CP does not significantly increase subsequent acid susceptibility of enamel as determined by Ca²⁺ release from the enamel surface. The effects of bleaching on mineral loss were indistinguishable from distilled water in the experimental system used. However, some surface differences were observed by SEM. The phosphomolybdenum blue method for phosphate is compromised by peroxide bleaching agents due to their oxidising properties. However, the Ca²⁺ electrode is unaffected by oxidising agents and can be used to determine the mineral loss in the presence of peroxides.

Keywords: bleaching, carbamide peroxide, demineralisation, teeth whitening

Procedia PDF Downloads 110

Authors: C. B. Le, V. N. Pham

Abstract:

In modern data analysis, multi-source data appears more and more in real applications. Multi-source data clustering has emerged as a important issue in the data mining and machine learning community. Different data sources provide information about different data. Therefore, multi-source data linking is essential to improve clustering performance. However, in practice multi-source data is often heterogeneous, uncertain, and large. This issue is considered a major challenge from multi-source data. Ensemble is a versatile machine learning model in which learning techniques can work in parallel, with big data. Clustering ensemble has been shown to outperform any standard clustering algorithm in terms of accuracy and robustness. However, most of the traditional clustering ensemble approaches are based on single-objective function and single-source data. This paper proposes a new clustering ensemble method for multi-source data analysis. The fuzzy optimized multi-objective clustering ensemble method is called FOMOCE. Firstly, a clustering ensemble mathematical model based on the structure of multi-objective clustering function, multi-source data, and dark knowledge is introduced. Then, rules for extracting dark knowledge from the input data, clustering algorithms, and base clusterings are designed and applied. Finally, a clustering ensemble algorithm is proposed for multi-source data analysis. The experiments were performed on the standard sample data set. The experimental results demonstrate the superior performance of the FOMOCE method compared to the existing clustering ensemble methods and multi-source clustering methods.

Keywords: clustering ensemble, multi-source, multi-objective, fuzzy clustering

Procedia PDF Downloads 166

Authors: Helen Zhang

Abstract:

Accurate segmentation of the airways from CT scans is crucial for early diagnosis of lung cancer. However, the existing airway segmentation algorithms often rely on thin-slice CT scans, which can be inconvenient and costly. This paper presents a set of machine learning-based 3D super-resolution algorithms along heterogeneous dimensions to improve the resolution of thicker CT scans to reduce the reliance on thin-slice scans. To evaluate the efficacy of the super-resolution algorithms, quantitative assessments using PSNR (Peak Signal to Noise Ratio) and SSIM (Structural SIMilarity index) were performed. The impact of super-resolution on airway segmentation accuracy is also studied. The proposed approach has the potential to make airway segmentation more accessible and affordable, thereby facilitating early diagnosis and treatment of lung cancer.

Keywords: 3D super-resolution, airway segmentation, thin-slice CT scans, machine learning

Procedia PDF Downloads 93

Authors: RAM PAL SINGH, VIKASH CHAUDHARY, MONIKA VERMA

Abstract:

In this paper, we proposed the implementation of optimization based Extreme Learning Machine (ELM) for watermarking of B-channel of color image in discrete wavelet transform (DWT) domain. ELM, a regularization algorithm, works based on generalized single-hidden-layer feed-forward neural networks (SLFNs). However, hidden layer parameters, generally called feature mapping in context of ELM need not to be tuned every time. This paper shows the embedding and extraction processes of watermark with the help of ELM and results are compared with already used machine learning models for watermarking.Here, a cover image is divide into suitable numbers of non-overlapping blocks of required size and DWT is applied to each block to be transformed in low frequency sub-band domain. Basically, ELM gives a unified leaning platform with a feature mapping, that is, mapping between hidden layer and output layer of SLFNs, is tried for watermark embedding and extraction purpose in a cover image. Although ELM has widespread application right from binary classification, multiclass classification to regression and function estimation etc. Unlike SVM based algorithm which achieve suboptimal solution with high computational complexity, ELM can provide better generalization performance results with very small complexity. Efficacy of optimization method based ELM algorithm is measured by using quantitative and qualitative parameters on a watermarked image even though image is subjected to different types of geometrical and conventional attacks.

Keywords: BER, DWT, extreme leaning machine (ELM), PSNR

Procedia PDF Downloads 295

Authors: Harish Bisht, Bahadur Singh Kotlia, Kireet Kumar

Abstract:

In order to reconstruct past retreating rates, total area loss, volume change and shift in snout position were measured through multi-temporal satellite data from 1989 to 2016 and kinematic GPS survey from 2015 to 2016. The results obtained from satellite data indicate that in the last 27 years, Chaturangi glacier snout has retreated 1172.57 ± 38.3 m (average 45.07 ± 4.31 m/year) with a total area and volume loss of 0.626 ± 0.001 sq. Km and 0.139 Km³, respectively. The field measurements through differential global positioning system survey revealed that the annual retreating rate was 22.84 ± 0.05 m/year. The large variations in results derived from both the methods are probably because of higher difference in their accuracy. Snout monitoring of the Gangotri glacier during the ablation season (May to September) in the years 2005 and 2015 reveals that the retreating rate has been comparatively more declined than that shown by the earlier studies. The GPS dataset shows that the average recession rate is 10.26 ± 0.05 m/year. In order to determine the possible causes of decreased retreating rate, a relationship between debris thickness and melt rate was also established by using ablation stakes. The present study concludes that remote sensing method is suitable for large area and long term study, while kinematic GPS is more appropriate for the annual monitoring of retreating rate of glacier snout. The present study also emphasizes on mapping of all the tributary glaciers in order to assess the overall changes in the main glacier system and its health.

Keywords: Chaturangi glacier, Gangotri glacier, glacier snout, kinematic global positioning system, retreat rate

Procedia PDF Downloads 124

Authors: Chhavi Rana, Sanjay K Jain

Abstract:

All sides of every education debate agree that quality learning happens when knowledgeable, caring teachers use sound pedagogy. Many deliberations of pedagogy make the fault of considering it as principally being about teaching. There has been lot of research about how to build a positive climate for learning, improve student curiosity, and enhance classroom association. However, these things can only be facilitated when teachers are equipped with better teaching techniques that use sound and accurate pedagogy. Pedagogy is the science and art of education. Its aims range from the full development of the human being to skills acquisition. In India, a project named Mission 10 x has been started by an esteemed IT Corporation Wipro as a faculty development programme (FDP) that particularly focus on elements that facilitated teachers in developing curriculum and new pedagogies that can lead to improvement in student engagement. This paper presents a study of these FDPs and examines (1) the parameters that help teachers in building new pedagogies (2) the extent to which appropriate usage of pedagogy is improved after the conduct of Mission 10 x FDPs, and (3) whether institutions differ in terms of their ability to convert usage of improved pedagogy into academic performance via these FDPs. The sample consisted of 2,236 students at 6 four-year engineering colleges and universities that completed several FDPs during 2012-2014. Many measures of usage of better pedagogy were linked positively with such FDPs, although some of the relationships were weak in strength. The results suggest that the usage of pedagogy were more benefited after conducting these FDPs and application of novel approaches in conducting classes.

Keywords: student engagement, critical thinking; achievement, student learning, pedagogy

Procedia PDF Downloads 404

Authors: Brandon Foggo, Nanpeng Yu

Abstract:

Power distribution circuits undergo frequent network topology changes that are often left undocumented. As a result, the documentation of a circuit’s connectivity becomes inaccurate with time. The lack of reliable circuit connectivity information is one of the biggest obstacles to model, monitor, and control modern distribution systems. To enhance the reliability and efficiency of electric power distribution systems, the circuit’s connectivity information must be updated periodically. This paper focuses on one critical component of a distribution circuit’s topology - the secondary transformer to phase association. This topology component describes the set of phase lines that feed power to a given secondary transformer (and therefore a given group of power consumers). Finding the documentation of this component is call Phase Identification, and is typically performed with physical measurements. These measurements can take time lengths on the order of several months, but with supervised learning, the time length can be reduced significantly. This paper compares several such methods applied to Phase Identification for a large range of real distribution circuits, describes a method of training data selection, describes preprocessing steps unique to the Phase Identification problem, and ultimately describes a method which obtains high accuracy (> 96% in most cases, > 92% in the worst case) using only 5% of the measurements typically used for Phase Identification.

Keywords: distribution network, machine learning, network topology, phase identification, smart grid

Procedia PDF Downloads 283

Authors: Ali Almadan, Anoop Krishnan, Ajita Rattani

Abstract:

Facial recognition is a widely adopted technology in surveillance, border control, healthcare, banking services, and lately, in mobile user authentication with Apple introducing “Face ID” moniker with iPhone X. A lot of research has been conducted in the area of face recognition on datasets captured by surveillance cameras, DSLR, and mobile devices. Recently, face recognition technology has also been deployed on body-worn cameras to keep officers safe, enabling situational awareness and providing evidence for trial. However, limited academic research has been conducted on this topic so far, without the availability of any publicly available datasets with a sufficient sample size. This paper aims to advance research in the area of face recognition using body-worn cameras. To this aim, the contribution of this work is two-fold: (1) collection of a dataset consisting of a total of 136,939 facial images of 102 subjects captured using body-worn cameras in in-door and daylight conditions and (2) evaluation of various deep-learning architectures for face identification on the collected dataset. Experimental results suggest a maximum True Positive Rate(TPR) of 99.86% at False Positive Rate(FPR) of 0.000 obtained by SphereFace based deep learning architecture in daylight condition. The collected dataset and the baseline algorithms will promote further research and development. A downloadable link of the dataset and the algorithms is available by contacting the authors.

Keywords: face recognition, body-worn cameras, deep learning, person identification

Procedia PDF Downloads 151

Authors: Theresia El Khoury, Georges Badr, Amir Hajjam El Hassani, Stéphane N’Guyen Van Ky

Abstract:

The widespread adoption of machine learning and artificial intelligence across different domains can be attributed to the digitization of data over several decades, resulting in vast amounts of data, types, and structures. Thus, data processing and preparation turn out to be a crucial stage. However, applying these techniques to S1000D standard-based data poses a challenge due to its complexity and the need to preserve logical information. This paper describes SALERNO, an S1000d AppLicability dEpended pRocessiNg mOdel. This python-based model analyzes and converts the XML S1000D-based files into an easier data format that can be used in machine learning techniques while preserving the different logic and relationships in files. The model parses the files in the given folder, filters them, and extracts the required information to be saved in appropriate data frames and Excel sheets. Its main idea is to group the extracted information by applicability. In addition, it extracts the full text by replacing internal and external references while maintaining the relationships between files, as well as the necessary requirements. The resulting files can then be saved in databases and used in different models. Documents in both English and French languages were tested, and special characters were decoded. Updates on the technical manuals were taken into consideration as well. The model was tested on different versions of the S1000D, and the results demonstrated its ability to effectively handle the applicability, requirements, references, and relationships across all files and on different levels.

Keywords: aeronautics, big data, data processing, machine learning, S1000D

Procedia PDF Downloads 119

Authors: Jamison Duckworth, Shankarachary Ragi

Abstract:

Detection and segmentation of wildfire smoke plumes from remote sensing imagery are being pursued as a solution for early fire detection and response. Smoke plume detection can be automated and made robust by the application of artificial intelligence methods. Specifically, in this study, the deep learning approach Mask Region-based Convolutional Neural Network (RCNN) is being proposed to learn smoke patterns across different spectral bands. This method is proposed to separate the smoke regions from the background and return masks placed over the smoke plumes. Multispectral data was acquired using NASA’s Earthdata and WorldView and services and satellite imagery. Due to the use of multispectral bands along with the three visual bands, we show that Mask R-CNN can be applied to distinguish smoke plumes from clouds and other landscape features that resemble smoke.

Keywords: deep learning, mask-RCNN, smoke plumes, spectral bands

Procedia PDF Downloads 107

Authors: Alessandro Preziosi, Antonio Prioletti, Luca Castangia

Abstract:

Deformable part models achieve high precision in pedestrian recognition, but all publicly available implementations are too slow for real-time applications. We implemented a deformable part model algorithm fast enough for real-time use by exploiting information about the camera position and orientation. This implementation is both faster and more precise than alternative DPM implementations. These results are obtained by computing convolutions in the frequency domain and using lookup tables to speed up feature computation. This approach is almost an order of magnitude faster than the reference DPM implementation, with no loss in precision. Knowing the position of the camera with respect to horizon it is also possible prune many hypotheses based on their size and location. The range of acceptable sizes and positions is set by looking at the statistical distribution of bounding boxes in labelled images. With this approach it is not needed to compute the entire feature pyramid: for example higher resolution features are only needed near the horizon. This results in an increase in mean average precision of 5% and an increase in speed by a factor of two. Furthermore, to reduce misdetections involving small pedestrians near the horizon, input images are supersampled near the horizon. Supersampling the image at 1.5 times the original scale, results in an increase in precision of about 4%. The implementation was tested against the public KITTI dataset, obtaining an 8% improvement in mean average precision over the best performing DPM-based method. By allowing for a small loss in precision computational time can be easily brought down to our target of 100ms per image, reaching a solution that is faster and still more precise than all publicly available DPM implementations.

Keywords: autonomous vehicles, deformable part model, dpm, pedestrian detection, real time

Procedia PDF Downloads 265

Authors: A. Bouabid, B. Bielenberg, S. Ainane, N. Pasha

Abstract:

In this paper, a team of faculty members of the Petroleum Institute in Abu Dhabi, UAE representing six different courses across General Engineering (ENGR), Communication (COMM), and Design (STPS) worked together to establish a clear developmental progression of learning outcomes and performance indicators for targeted knowledge, areas of competency, and skills for the first three semesters of the Bachelor of Sciences in Engineering curriculum. The sequences of courses studied in this project were ENGR/COMM, COMM/STPS, and ENGR/STPS. For each course’s nine areas of knowledge, competency, and skills, the research team reviewed the existing learning outcomes and related performance indicators with a focus on identifying linkages across disciplines as well as within the courses of a discipline. The team reviewed existing performance indicators for developmental progression from semester to semester for same discipline related courses (vertical alignment) and for different discipline courses within the same semester (horizontal alignment). The results of this work have led to recommendations for modifications of the initial indicators when incoherence was identified, and/or for new indicators based on best practices (identified through literature searches) when gaps were identified. It also led to recommendations for modifications of the level of emphasis within each course to ensure developmental progression. The exercise has led to a revised Sequence Performance Indicator Mapping for the knowledge, skills, and competencies across the six core courses.

Keywords: curriculum alignment, horizontal and vertical progression, performance indicators, skill level

Procedia PDF Downloads 206

Authors: Bander Mohayya Alotaibi

Abstract:

The valuable partnerships between parents and teachers may develop positive and effective interactions between home and school. This will help these stakeholders share information and resources regarding student academics during ongoing interactions. Thus, partnerships will build a solid foundation for both families and schools to help children succeed in school. Parental involvement can be seen as an effective tool that can change homes and communities and not just schools’ systems. Seeking parents and stakeholders’ attitudes toward learning and learners can help schools design a curriculum. Subsequently, this information can be used to find ways to help improve the academic performance of students, especially in low performing schools. There may be some conflicts when designing curriculum. In addition, designing curriculum might bring more educational expectations to all the sides. There is a lack of research that targets the specific attitude of parents toward specific concepts on curriculum contents. More research is needed to study the perspective that parents of children with learning disabilities (LD) have regarding early reading curriculum. Parents and stakeholders’ perspectives on early reading intervention implemented as a curriculum for children with LD was studied through an advanced quantitative research. The purpose of this study seeks to understand stakeholders and parents’ perspectives of key concepts and essential early reading skills that impact the design of curriculum that will serve as an intervention for early struggler readers who have LD. Those concepts or stages include phonics, phonological awareness, and reading fluency as well as strategies used in house by parents. A survey instrument was used to gather the data. Participants were recruited through 29 schools and districts of the metropolitan area of the northern part of Saudi Arabia. Participants were stakeholders including parents of children with learning disability. Data were collected using distribution of paper and pen survey to schools. Psychometric properties of the instrument were evaluated for the validity and reliability of the survey; face validity, content validity, and construct validity including an Exploratory Factor Analysis were used to shape and reevaluate the structure of the instrument. Multivariate analysis of variance (MANOVA) used to find differences between the variables. The study reported the results of the perspectives of stakeholders toward reading strategies, phonics, phonological awareness, and reading fluency. Also, suggestions and limitations are discussed.

Keywords: stakeholders, learning disability, early reading, perspectives, parents, intervention, curriculum

Procedia PDF Downloads 139

Authors: Mais Nijim, Rama Devi Chennuboyina, Waseem Al Aqqad

Abstract:

Advances in spatial and spectral resolution of satellite images have led to tremendous growth in large image databases. The data we acquire through satellites, radars and sensors consists of important geographical information that can be used for remote sensing applications such as region planning, disaster management. Spatial data classification and object recognition are important tasks for many applications. However, classifying objects and identifying them manually from images is a difficult task. Object recognition is often considered as a classification problem, this task can be performed using machine-learning techniques. Despite of many machine-learning algorithms, the classification is done using supervised classifiers such as Support Vector Machines (SVM) as the area of interest is known. We proposed a classification method, which considers neighboring pixels in a region for feature extraction and it evaluates classifications precisely according to neighboring classes for semantic interpretation of region of interest (ROI). A dataset has been created for training and testing purpose; we generated the attributes by considering pixel intensity values and mean values of reflectance. We demonstrated the benefits of using knowledge discovery and data-mining techniques, which can be on image data for accurate information extraction and classification from high spatial resolution remote sensing imagery.

Keywords: remote sensing, object recognition, classification, data mining, waterbody identification, feature extraction

Procedia PDF Downloads 323

Authors: Hana Maresova, Daniel Ecler

Abstract:

The paper focuses on the use of 3D multi-user virtual environments (MUVE) in language teaching and presents the results of four years of research at the Faculty of Education, Palacký University in Olomouc (Czech Republic). In the form of an experiment, mother tongue language teaching in the 3D virtual worlds Second Life and Kitely (experimental group) and parallel traditional teaching on identical topics representing teacher's interpretation using a textbook (control group) were implemented. The didactic test, which was presented to the experimental and control groups in an identical form before and after the instruction, verified the effect of the instruction in the experimental group by comparing the results obtained by both groups. Within the three components of mother-tongue teaching (vocabulary, literature, style and communication education), the students in the literature group achieved partially better results (statistically significant in the case of items devoted to the area of visualization of the learning topic), while in the case of grammar and style education the respondents of the control group achieved better results. On the basis of the results obtained, we can conclude that the most appropriate use of MUVE can be seen in the teaching of those topics that provide the possibility of dramatization, experiential learning and group involvement and cooperation, on the contrary, with regard to the need to divide students attention between the topic taught and the control of avatar and movement in virtual reality as less suitable for teaching in the area of memorization of the topic or concepts.

Keywords: distance learning, 3D virtual environments, online teaching, language teaching

Procedia PDF Downloads 147

Authors: Kyle De Freitas, Margaret Bernard

Abstract:

Enabling data mining analysis among a wider audience of educators is an active area of research within the educational data mining (EDM) community. The paper proposes a framework for developing an environment that caters for educators who have little technical data mining skills as well as for more advanced users with some data mining expertise. This framework architecture was developed through the review of the strengths and weaknesses of existing models in the literature. The proposed framework provides a modular architecture for future researchers to focus on the development of specific areas within the EDM process. Finally, the paper also highlights a strategy of enabling analysis through either the use of predefined questions or a guided data mining process and highlights how the developed questions and analysis conducted can be reused and extended over time.

Keywords: educational data mining, learning management system, learning analytics, EDM framework

Procedia PDF Downloads 309

Authors: Jack R. McKenzie, Peter A. Appleby, Thomas House, Neil Walton

Abstract:

Cold-start is a notoriously difficult problem which can occur in recommendation systems, and arises when there is insufficient information to draw inferences for users or items. To address this challenge, a contextual bandit algorithm – the Fast Approximate Bayesian Contextual Cold Start Learning algorithm (FAB-COST) – is proposed, which is designed to provide improved accuracy compared to the traditionally used Laplace approximation in the logistic contextual bandit, while controlling both algorithmic complexity and computational cost. To this end, FAB-COST uses a combination of two moment projection variational methods: Expectation Propagation (EP), which performs well at the cold start, but becomes slow as the amount of data increases; and Assumed Density Filtering (ADF), which has slower growth of computational cost with data size but requires more data to obtain an acceptable level of accuracy. By switching from EP to ADF when the dataset becomes large, it is able to exploit their complementary strengths. The empirical justification for FAB-COST is presented, and systematically compared to other approaches on simulated data. In a benchmark against the Laplace approximation on real data consisting of over 670, 000 impressions from autotrader.co.uk, FAB-COST demonstrates at one point increase of over 16% in user clicks. On the basis of these results, it is argued that FAB-COST is likely to be an attractive approach to cold-start recommendation systems in a variety of contexts.

Keywords: cold-start learning, expectation propagation, multi-armed bandits, Thompson Sampling, variational inference

Procedia PDF Downloads 97