Search results for: classifiers accuracy

Authors: Tamanna Goyal, Divya Bansal, Sanjeev Sofat

Abstract:

In world-threatening terrorist attacks, where early detection, distinction, and prediction are effective diagnosis techniques and for functionally accurate and precise analysis of terrorism data, there are so many data mining & statistical approaches to assure accuracy. The computational extraction of derived patterns is a non-trivial task which comprises specific domain discovery by means of sophisticated algorithm design and analysis. This paper proposes an approach for similarity extraction by obtaining the useful attributes from the available datasets of terrorist attacks and then applying feature selection technique based on the statistical impurity measures followed by clustering techniques on the basis of similarity measures. On the basis of degree of participation of attributes in the rules, the associative dependencies between the attacks are analyzed. Consequently, to compute the similarity among the discovered rules, we applied a weighted similarity measure. Finally, the rules are grouped by applying using hierarchical clustering. We have applied it to an open source dataset to determine the usability and efficiency of our technique, and a literature search is also accomplished to support the efficiency and accuracy of our results.

Keywords: association rules, clustering, similarity measure, statistical approaches

Procedia PDF Downloads 322

Authors: Khalida Douibi, Rodrigo Balp, Solène Le Bars

Abstract:

In the medical field, applications related to human experiments are frequently linked to reduced samples size, which makes the training of machine learning models quite sensitive and therefore not very robust nor generalizable. This is notably the case in Brain-Computer Interface (BCI) studies, where the sample size rarely exceeds 20 subjects or a few number of trials. To address this problem, several resampling approaches are often used during the data preparation phase, which is an overly critical step in a data science analysis process. One of the naive approaches that is usually applied by data scientists consists in the transformation of the entire database before the resampling phase. However, this can cause model’ s performance to be incorrectly estimated when making predictions on unseen data. In this paper, we explored the effect of data leakage observed during our BCI experiments for device control through the real-time classification of SSVEPs (Steady State Visually Evoked Potentials). We also studied potential ways to ensure optimal validation of the classifiers during the calibration phase to avoid overfitting. The results show that the scaling step is crucial for some algorithms, and it should be applied after the resampling phase to avoid data leackage and improve results.

Keywords: data leackage, data science, machine learning, SSVEP, BCI, overfitting

Procedia PDF Downloads 153

Authors: Laxman H. Surwase, Lalit V. Sonawane, Bhagwat N. Poul

Abstract:

A simple stability indicating high performance liquid chromatographic method has been developed for the simultaneous determination of Levosalbutamol Sulphate and Ipratropium Bromide in bulk and pharmaceutical dosage form using reverse phase Zorbax Eclipse Plus C8 column (250mm×4.6mm), with mobile phase phosphate buffer (0.05M KH2PO4): acetonitrile (55:45v/v) pH 3.5 adjusted with ortho-phosphoric acid, the flow rate was 1.0 mL/min and the detection was carried at 212 nm. The retention times of Levosalbutamol Sulphate and Ipratropium Bromide were 2.2007 and 2.6611 min respectively. The correlation coefficient of Levosalbutamol Sulphate and Ipratropium Bromide was found to be 0.997 and 0.998.Calibration plots were linear over the concentration ranges 10-100µg/mL for both Levosalbutamol Sulphate and Ipratropium Bromide. The LOD and LOQ of Levosalbutamol Sulphate were 2.520µg/mL and 7.638µg/mL while for Ipratropium Bromide was 1.201µg/mL and 3.640 µg/mL. The accuracy of the proposed method was determined by recovery studies and found to be 100.15% for Levosalbutamol Sulphate and 100.19% for Ipratropium Bromide respectively. The method was validated for accuracy, linearity, sensitivity, precision, robustness, system suitability. The proposed method could be utilized for routine analysis of Levosalbutamol Sulphate and Ipratropium Bromide in bulk and pharmaceutical capsule dosage form.

Keywords: levosalbutamol sulphate, ipratropium bromide, RP-HPLC, phosphate buffer, acetonitrile

Procedia PDF Downloads 351

Authors: Mohsen Bahrami, Banafsheh Nikmehr, Yueqiang Song, Anuradha Koduru, Ayse K. Vuruskan, Hongkun Lu, Tamer M. Yalcinkaya

Abstract:

In recent years, we have witnessed an explosion of studies aimed at using a combination of artificial intelligence (AI) and time-lapse imaging data on embryos to improve IVF outcomes. However, despite promising results, no study has used a matched cohort of transferred embryos which only differ in pregnancy outcome, i.e., embryos from a single clinic which are similar in parameters, such as: morphokinetic condition, patient age, and overall clinic and lab performance. Here, we used time-lapse data on embryos with known pregnancy outcomes to see if the rich spatiotemporal information embedded in this data would allow the prediction of the pregnancy outcome regardless of such critical parameters. Methodology—We did a retrospective analysis of time-lapse data from our IVF clinic utilizing Embryoscope 100% of the time for embryo culture to blastocyst stage with known clinical outcomes, including live birth vs nonpregnant (embryos with spontaneous abortion outcomes were excluded). We used time-lapse data from 200 elective single transfer embryos randomly selected from January 2019 to June 2021. Our sample included 100 embryos in each group with no significant difference in patient age (P=0.9550) and morphokinetic scores (P=0.4032). Data from all patients were combined to make a 4th order tensor, and feature extraction were subsequently carried out by a tensor decomposition methodology. The features were then used in a machine learning classifier to classify the two groups. Major Findings—The performance of the model was evaluated using 100 random subsampling cross validation (train (80%) - test (20%)). The prediction accuracy, averaged across 100 permutations, exceeded 80%. We also did a random grouping analysis, in which labels (live birth, nonpregnant) were randomly assigned to embryos, which yielded 50% accuracy. Conclusion—The high accuracy in the main analysis and the low accuracy in random grouping analysis suggest a consistent spatiotemporal pattern which is associated with pregnancy outcomes, regardless of patient age and embryo morphokinetic condition, and beyond already known parameters, such as: early cleavage or early blastulation. Despite small samples size, this ongoing analysis is the first to show the potential of AI methods in capturing the complex morphokinetic changes embedded in embryo time-lapse data, which contribute to successful pregnancy outcomes, regardless of already known parameters. The results on a larger sample size with complementary analysis on prediction of other key outcomes, such as: euploidy and aneuploidy of embryos will be presented at the meeting.

Keywords: IVF, embryo, machine learning, time-lapse imaging data

Procedia PDF Downloads 93

Authors: M. S. Mousavi-Daramoroudi, H. Yousefnia, S. Zolghadri, F. Abbasi-Davani

Abstract:

Dosimetry is an indispensable and precious factor in patient treatment planning; to minimize the absorbed dose in vital tissues. In this study, In accordance with the proper characteristics of DOTATOC and ¹⁷⁷Lu, after preparing ¹⁷⁷Lu-DOTATOC at the optimal conditions for the first time in Iran, radionuclidic and radiochemical purity of the solution was investigated using an HPGe spectrometer and ITLC method, respectively. The biodistribution of the compound was assayed for treatment of adenocarcinoma breast cancer in bearing BALB/c mice. The results have demonstrated that ¹⁷⁷Lu-DOTATOC is a profitable selection for therapy of the tumors. Because of the vital role of internal dosimetry before and during therapy, the effort to improve the accuracy and rapidity of dosimetric calculations is necessary. For this reason, a new method was accomplished to calculate the absorbed dose through mixing between compartmental model, animal dosimetry and extrapolated data from animal to human and using MIRD method. Despite utilization of compartmental model based on the experimental data, it seems this approach may increase the accuracy of dosimetric data, confidently.

Keywords: ¹⁷⁷Lu-DOTATOC, biodistribution modeling, compartmental model, internal dosimetry

Procedia PDF Downloads 221

Authors: Yuntao Liu, Lei Wang, Haoran Xia

Abstract:

Machine learning has shown extensive applications in the development of classification models for autism spectrum disorder (ASD) using neural image data. This paper proposes a fusion multi-modal classification network based on a graph neural network. First, the brain is segmented into 116 regions of interest using a medical segmentation template (AAL, Anatomical Automatic Labeling). The image features of sMRI and the signal features of fMRI are extracted, which build the node and edge embedding representations of the brain map. Then, we construct a dynamically updated brain map neural network and propose a method based on a dynamic brain map adjacency matrix update mechanism and learnable graph to further improve the accuracy of autism diagnosis and recognition results. Based on the Autism Brain Imaging Data Exchange I dataset(ABIDE I), we reached a prediction accuracy of 74% between ASD and TD subjects. Besides, to study the biomarkers that can help doctors analyze diseases and interpretability, we used the features by extracting the top five maximum and minimum ROI weights. This work provides a meaningful way for brain disorder identification.

Keywords: autism spectrum disorder, brain map, supervised machine learning, graph network, multimodal data, model interpretability

Procedia PDF Downloads 71

Authors: Monireh Ahmadi Bani, Adel Khorramrouz, Lalenoor Morvarid, Bagheri Mahtab

Abstract:

Background:- Foot disorders are common in musculoskeletal problems. Plantar pressure distribution measurement is one the most important part of foot disorders diagnosis for quantitative analysis. However, the association of plantar pressure and foot disorders is not clear. With the growth of dataset and machine learning methods, the relationship between foot disorders and plantar pressures can be detected. Significance of the study:- The purpose of this study was to predict the probability of common foot disorders based on peak plantar pressure distribution and center of pressure during walking. Methodologies:- 2323 participants were assessed in a foot therapy clinic between 2015 and 2021. Foot disorders were diagnosed by an experienced physician and then they were asked to walk on a force plate scanner. After the data preprocessing, due to the difference in walking time and foot size, we normalized the samples based on time and foot size. Some of force plate variables were selected as input to a deep neural network (DNN), and the probability of any each foot disorder was measured. In next step, we used support vector machine (SVM) and run dataset for each foot disorder (classification of yes or no). We compared DNN and SVM for foot disorders prediction based on plantar pressure distributions and center of pressure. Findings:- The results demonstrated that the accuracy of deep learning architecture is sufficient for most clinical and research applications in the study population. In addition, the SVM approach has more accuracy for predictions, enabling applications for foot disorders diagnosis. The detection accuracy was 71% by the deep learning algorithm and 78% by the SVM algorithm. Moreover, when we worked with peak plantar pressure distribution, it was more accurate than center of pressure dataset. Conclusion:- Both algorithms- deep learning and SVM will help therapist and patients to improve the data pool and enhance foot disorders prediction with less expense and error after removing some restrictions properly.

Keywords: deep neural network, foot disorder, plantar pressure, support vector machine

Procedia PDF Downloads 359

Authors: Robert Rusinek, Marek Gancarz, Jolanta Wawrzyniak, Marzena Gawrysiak-Witulska, Dariusz Wiącek, Agnieszka Nawrocka

Abstract:

Investigations were performed to examine the possibility of using an electronic nose to monitor the development of fungal microflora during the first eighteen days of rapeseed storage. The Cyranose 320 device with polymer-composite sensors was used. Each sample of infected material was divided into three parts, and the degree of spoilage was measured in three ways: analysis of colony forming units (CFU), determination of ergosterol content (ERG), and measurement with the eNose. Principal component analysis (PCA) was performed on the generated patterns of signals, and six groups of different spoilage levels were isolated. The electronic nose with polymer-composite sensors under laboratory conditions distinguished between species of spoiled and unspoiled seeds with 100% accuracy. Despite some minor differences in the CFU and ergosterol content, the electronic nose provided responses correctly corresponding to the level of spoilage with 85% accuracy. Therefore, the main conclusion from the study is that the electronic nose is a promising tool for quick and non-destructive detection of the level of oil seed spoilage. The research was supported by the National Centre for Research and Development (NCBR), Grant No. PBS2/A8/22/2013.

Keywords: colony forming units, electronic nose, ergosterol, rapeseed

Procedia PDF Downloads 323

Authors: Bin Feng, Zhulin Zong

Abstract:

Radar systems are critical in the modern military, aviation, and maritime operations, and their proper functioning is essential for the success of these operations. However, due to the complexity and sensitivity of radar systems, they are susceptible to various faults that can significantly affect their performance. Traditional radar fault diagnosis strategies rely on expert knowledge and rule-based approaches, which are often limited in effectiveness and require a lot of time and resources. Deep learning has recently emerged as a promising approach for fault diagnosis due to its ability to learn features and patterns from large amounts of data automatically. In this paper, we propose a radar fault diagnosis strategy based on deep learning that can accurately identify and classify faults in radar systems. Our approach uses convolutional neural networks (CNN) to extract features from radar signals and fault classify the features. The proposed strategy is trained and validated on a dataset of measured radar signals with various types of faults. The results show that it achieves high accuracy in fault diagnosis. To further evaluate the effectiveness of the proposed strategy, we compare it with traditional rule-based approaches and other machine learning-based methods, including decision trees, support vector machines (SVMs), and random forests. The results demonstrate that our deep learning-based approach outperforms the traditional approaches in terms of accuracy and efficiency. Finally, we discuss the potential applications and limitations of the proposed strategy, as well as future research directions. Our study highlights the importance and potential of deep learning for radar fault diagnosis. It suggests that it can be a valuable tool for improving the performance and reliability of radar systems. In summary, this paper presents a radar fault diagnosis strategy based on deep learning that achieves high accuracy and efficiency in identifying and classifying faults in radar systems. The proposed strategy has significant potential for practical applications and can pave the way for further research.

Keywords: radar system, fault diagnosis, deep learning, radar fault

Procedia PDF Downloads 92

Authors: Sung Duk Kim, Kwan U. Kim, Jin Sim, Ryong Jin Jang

Abstract:

The phenomenon of seasonal variations in the Earth’s rotation velocity was discovered in the 1930s when a crystal clock was developed and analyzed in a quantitative way for the first time between 1955 and 1968 when observation data of the seasonal variations was analyzed by an atomic clock. According to the previous investigation, atmospheric circulation is supposed to be a factor affecting the seasonal variations in the Earth’s rotation velocity in many cases, but the problem has not been solved yet. In order to solve the problem, it is necessary to apply dynamics to consider the Earth’s spatial motion, rotation, and change of shape of the Earth (movement of materials in and out of the Earth and change of the Earth’s figure) at the same time and in interrelation to the accuracy of post-Newtonian approximation regarding the Earth body as a system of mass points because the stability of the Earth’s rotation angular velocity is in the range of 10⁻⁸~10⁻⁹. For it, the equation was derived, which can consider the 3 kinds of motion above mentioned at the same time by taking the effect of the resultant external force on the Earth’s rotation into account in a relativistic way to the accuracy of post-Newtonian approximation. Therefore, the equation has been solved to obtain the theoretical values of periodic change in the Earth’s rotation velocity, and they have been compared with the astronomical observation data so to reveal the cause for the periodic change in the Earth’s rotation velocity.

Keywords: Earth rotation, moment function, periodic change, seasonal variation, relativistic change

Procedia PDF Downloads 77

Authors: Muhammad Maisam Abbas, Salman Tariq, Usama Riaz, Muhammad Tanveer, Humaira Abdul Ghafoor

Abstract:

Autism Spectrum Disorder (ASD) refers to a developmental disorder that impairs an individual's communication and interaction ability. Individuals feel difficult to read facial expressions while communicating or interacting. Facial Expression Recognition (FER) is a unique method of classifying basic human expressions, i.e., happiness, fear, surprise, sadness, disgust, neutral, and anger through static and dynamic sources. This paper conducts a comprehensive comparison and proposed optimal method for a continued research project—a system that can assist people who have Autism Spectrum Disorder (ASD) in recognizing facial expressions. Comparison has been conducted on three supervised learning algorithms EigenFace, FisherFace, and LBPH. The JAFFE, CK+, and TFEID (I&II) datasets have been used to train and test the algorithms. The results were then evaluated based on variance, standard deviation, and accuracy. The experiments showed that FisherFace has the highest accuracy for all datasets and is considered the best algorithm to be implemented in our system.

Keywords: autism spectrum disorder, ASD, EigenFace, facial expression recognition, FisherFace, local binary pattern histogram, LBPH

Procedia PDF Downloads 176

Authors: Liyun Chang, Cheng-Hsiang Tsai

Abstract:

To ensure the radiation precisely delivering to the target of cancer patients, the linear accelerator equipped with the pretreatment on-board imaging system is introduced and through it the patient setup is verified before the daily treatment. New generation radiotherapy using beam-intensity modulation, usually associated the treatment with steep dose gradients, claimed to have achieved both a higher degree of dose conformation in the targets and a further reduction of toxicity in normal tissues. However, this benefit is counterproductive if the beam is delivered imprecisely. To avoid shooting critical organs or normal tissues rather than the target, it is very important to carry out the quality assurance (QA) of this on-board imaging system. The QA of the On-Board Imager® (OBI) system of one Varian Clinac-iX linear accelerator was performed through our procedures modified from a relevant report and AAPM TG142. Two image modalities, 2D radiography and 3D cone-beam computed tomography (CBCT), of the OBI system were examined. The daily and monthly QA was executed for five months in the categories of safety, geometrical accuracy and image quality. A marker phantom and a blade calibration plate were used for the QA of geometrical accuracy, while the Leeds phantom and Catphan 504 phantom were used in the QA of radiographic and CBCT image quality, respectively. The reference images were generated through a GE LightSpeed CT simulator with an ADAC Pinnacle treatment planning system. Finally, the image quality was analyzed via an OsiriX medical imaging system. For the geometrical accuracy test, the average deviations of the OBI isocenter in each direction are less than 0.6 mm with uncertainties less than 0.2 mm, while all the other items have the displacements less than 1 mm. For radiographic image quality, the spatial resolution is 1.6 lp/cm with contrasts less than 2.2%. The spatial resolution, low contrast, and HU homogenous of CBCT are larger than 6 lp/cm, less than 1% and within 20 HU, respectively. All tests are within the criteria, except the HU value of Teflon measured with the full fan mode exceeding the suggested value that could be due to itself high HU value and needed to be rechecked. The OBI system in our facility was then demonstrated to be reliable with stable image quality. The QA of OBI system is really necessary to achieve the best treatment for a patient.

Keywords: CBCT, image quality, quality assurance, OBI

Procedia PDF Downloads 300

Authors: Sirilak Areerachakul

Abstract:

Water quality has initiated serious management efforts in many countries. Artificial Neural Network (ANN) models are developed as forecasting tools in predicting water quality trend based on historical data. This study endeavors to automatically classify water quality. The water quality classes are evaluated using 6 factor indices. These factors are pH value (pH), Dissolved Oxygen (DO), Biochemical Oxygen Demand (BOD), Nitrate Nitrogen (NO3N), Ammonia Nitrogen (NH3N) and Total Coliform (T-Coliform). The methodology involves applying data mining techniques using multilayer perceptron (MLP) neural network models. The data consisted of 11 sites of Saen Saep canal in Bangkok, Thailand. The data is obtained from the Department of Drainage and Sewerage Bangkok Metropolitan Administration during 2007-2011. The results of multilayer perceptron neural network exhibit a high accuracy multilayer perception rate at 94.23% in classifying the water quality of Saen Saep canal in Bangkok. Subsequently, this encouraging result could be combined with GIS data improves the classification accuracy significantly.

Keywords: artificial neural network, geographic information system, water quality, computer science

Procedia PDF Downloads 344

Authors: Hsin-Yang Wang, K. M. A. Salam, Ying-Jia Lin, Daniel Tan, Tzu-Hsuan Chou, Hung-Yu Kao

Abstract:

Detecting user intent from natural language expression has a wide variety of use cases in different natural language processing applications. Recently few-shot training has a spike of usage on commercial domains. Due to the lack of significant sample features, the downstream task performance has been limited or leads to an unstable result across different domains. As a state-of-the-art method, the pre-trained BERT model gathering the sentence-level information from a large text corpus shows improvement on several NLP benchmarks. In this research, we are proposing a method to change multi-class classification tasks into binary classification tasks, then use the confidence score to rank the results. As a language model, BERT performs well on sequence data. In our experiment, we change the objective from predicting labels into finding the relations between words in sequence data. Our proposed method achieved 71.0% accuracy in the internal intent detection dataset and 63.9% accuracy in the HuffPost dataset. Acknowledgment: This work was supported by NCKU-B109-K003, which is the collaboration between National Cheng Kung University, Taiwan, and SoftBank Corp., Tokyo.

Keywords: OSML, BERT, text classification, one shot

Procedia PDF Downloads 101

Authors: Alaa El-Din Rezk

Abstract:

In Egypt, management of crops still away from what is being used today by utilizing the advances of mechanical design capabilities, sensing and electronics technology. These technologies have been introduced in many places and recorm, for Straight Path, Curved Path, Sine Wave ded high accuracy in different field operations. So, an autonomous robotic platform based on machine vision has been developed and constructed to be implemented in Egyptian conditions as self-propelled mobile vehicle for carrying tools for inter/intra-row crop management based on different control modules. The experiments were carried out at plant protection research institute (PPRI) during 2014-2015 to optimize the accuracy of agricultural robotic platform control using machine vision in term of the autonomous navigation and performance of the robot’s guidance system. Results showed that the robotic platform' guidance system with machine vision was able to adequately distinguish the path and resisted image noise and did better than human operators for getting less lateral offset error. The average error of autonomous was 2.75, 19.33, 21.22, 34.18, and 16.69 mm. while the human operator was 32.70, 4.85, 7.85, 38.35 and 14.75 mm Path, Offset Discontinuity and Angle Discontinuity respectively.

Keywords: autonomous robotic, Hough transform, image processing, machine vision

Procedia PDF Downloads 316

Authors: Aileen F. Wang

Abstract:

Mammography has been one of the most reliable methods for early detection and diagnosis of breast cancer. However, mammography misses about 17% and up to 30% of breast cancers due to the subtle and unstable appearances of breast cancer in their early stages. Recent computer-aided diagnosis (CADx) technology using Zernike moments has improved detection accuracy. However, it has several drawbacks: it uses manual segmentation, Zernike moments are not robust, and it still has a relatively high false negative rate (FNR)–17.6%. This project will focus on the development of a novel breast cancer detection algorithm to automatically segment the breast mass and further reduce FNR. The algorithm consists of automatic segmentation of a single breast mass using Point Region Growing Segmentation, reconstruction of the segmented breast mass using Pseudo-Zernike moments, and classification of the breast mass using the root mean square (RMS). A comparative study among the various algorithms on the segmentation and reconstruction of breast masses was performed on randomly selected mammographic images. The results demonstrated that the newly developed algorithm is the best in terms of accuracy and cost effectiveness. More importantly, the new classifier RMS has the lowest FNR–6%.

Keywords: computer aided diagnosis, mammography, point region growing segmentation, pseudo-zernike moments, root mean square

Procedia PDF Downloads 453

Authors: Laidi Maamar, Achwak Madani, Abdellah El Ahdj Abdellah

Abstract:

The aim of this work is to use Support Vector regression (SVR) combined with dragonfly, firefly, Bee Colony and particle swarm Optimization algorithm to predict global solar radiation on horizontal surfaces in some cities in Algeria. Combining these optimization algorithms with SVR aims principally to enhance accuracy by fine-tuning the parameters, speeding up the convergence of the SVR model, and exploring a larger search space efficiently; these parameters are the regularization parameter (C), kernel parameters, and epsilon parameter. By doing so, the aim is to improve the generalization and predictive accuracy of the SVR model. Overall, the aim is to leverage the strengths of both SVR and optimization algorithms to create a more powerful and effective regression model for various cities and under different climate conditions. Results demonstrate close agreement between predicted and measured data in terms of different metrics. In summary, SVM has proven to be a valuable tool in modeling global solar radiation, offering accurate predictions and demonstrating versatility when combined with other algorithms or used in hybrid forecasting models.

Keywords: support vector regression (SVR), optimization algorithms, global solar radiation prediction, hybrid forecasting models

Procedia PDF Downloads 38

Authors: E. Irankhah, M. Zarif, E. Mazrooei Rad, K. Ghandehari

Abstract:

Alzheimer's prevalence is on the rise, and the disease comes with problems like cessation of treatment, high cost of treatment, and the lack of early detection methods. The pathology of this disease causes the formation of protein deposits in the brain of patients called plaque amyloid. Generally, the diagnosis of this disease is done by performing tests such as a cerebrospinal fluid, CT scan, MRI, and spinal cord fluid testing, or mental testing tests and eye tracing tests. In this paper, we tried to use the Medial Temporal Atrophy (MTA) method and the Leave One Out (LOO) cycle to extract the statistical properties of the three Fz, Pz, and Cz channels of ERP signals for early diagnosis of this disease. In the process of CT scan images, the accuracy of the results is 81% for the healthy person and 88% for the severe patient. After the process of ERP signaling, the accuracy of the results for a healthy person in the delta band in the Cz channel is 81% and in the alpha band the Pz channel is 90%. In the results obtained from the signal processing, the results of the severe patient in the delta band of the Cz channel were 89% and in the alpha band Pz channel 92%.

Keywords: Alzheimer's disease, image and signal processing, LOO cycle, medial temporal atrophy

Procedia PDF Downloads 198

Authors: Ali Reza Sattarzadeh, Ronny J. Kutadinata, Pubudu N. Pathirana, Van Thanh Huynh

Abstract:

Traffic congestion has become a severe worldwide problem, affecting everyday life, fuel consumption, time, and air pollution. The primary causes of these issues are inadequate transportation infrastructure, poor traffic signal management, and rising population. Traffic flow forecasting is one of the essential and effective methods in urban congestion and traffic management, which has attracted the attention of researchers. With the development of technology, undeniable progress has been achieved in existing methods. However, there is a possibility of improvement in the extraction of temporal and spatial features to determine the importance of traffic flow sequences and extraction features. In the proposed model, we implement the convolutional neural network (CNN) and long short-term memory (LSTM) deep learning models for mining nonlinear correlations and their effectiveness in increasing the accuracy of traffic flow prediction in the real dataset. According to the experiments, the results indicate that implementing Conv-LSTM networks increases the productivity and accuracy of deep learning models for traffic flow prediction.

Keywords: deep learning algorithms, intelligent transportation systems, spatiotemporal features, traffic flow prediction

Procedia PDF Downloads 173

Authors: Bowen Ding, Yihao Kuang

Abstract:

In recent years, large-scale pre-trained language models have achieved state-of-the-art performance on a variety of natural language processing tasks. The word vectors produced by these language models can be viewed as dense encoded presentations of natural language that in text form. However, it is unknown how much linguistic information is encoded and how. In this paper, we construct several corresponding probing tasks for multiple linguistic information to clarify the encoding capabilities of different language models and performed a visual display. We firstly obtain word presentations in vector form from different language models, including BERT, ELMo, RoBERTa and GPT. Classifiers with a small scale of parameters and unsupervised tasks are then applied on these word vectors to discriminate their capability to encode corresponding linguistic information. The constructed probe tasks contain both semantic and syntactic aspects. The semantic aspect includes the ability of the model to understand semantic entities such as numbers, time, and characters, and the grammatical aspect includes the ability of the language model to understand grammatical structures such as dependency relationships and reference relationships. We also compare encoding capabilities of different layers in the same language model to infer how linguistic information is encoded in the model.

Keywords: language models, probing task, text presentation, linguistic information

Procedia PDF Downloads 114

Authors: Preston Nabors, Nasseh Tabrizi

Abstract:

Portable Document Format (PDF) files have gained significant popularity for sharing and distributing documents due to their universal compatibility. However, the widespread use of PDF files has made them attractive targets for cybercriminals, who exploit vulnerabilities to deliver malware and compromise the security of end-user systems. This paper reviews notable contributions in PDF malware detection, including static, dynamic, signature-based, and hybrid analysis. It presents a comprehensive examination of PDF malware detection techniques, focusing on the emerging threat of adversarial sampling and the need for robust defense mechanisms. The paper highlights the vulnerability of machine learning classifiers to evasion attacks. It explores adversarial sampling techniques in PDF malware detection to produce mimicry and reverse mimicry evasion attacks, which aim to bypass detection systems. Improvements for future research are identified, including accessible methods, applying adversarial sampling techniques to malicious payloads, evaluating other models, evaluating the importance of features to malware, implementing adversarial defense techniques, and conducting comprehensive examination across various scenarios. By addressing these opportunities, researchers can enhance PDF malware detection and develop more resilient defense mechanisms against adversarial attacks.

Keywords: adversarial attacks, adversarial defense, adversarial machine learning, intrusion detection, PDF malware, malware detection, malware detection evasion

Procedia PDF Downloads 40

Authors: Nazmul Shahadat, Anthony S. Maida

Abstract:

In recent years, hypercomplex inspired neural networks improved deep CNN architectures due to their ability to share weights across input channels and thus improve cohesiveness of representations within the layers. The work described herein studies the effect of replacing existing layers in an Axial Attention ResNet with their quaternion variants that use cross-channel weight sharing to assess the effect on image classification. We expect the quaternion enhancements to produce improved feature maps with more interlinked representations. We experiment with the stem of the network, the bottleneck layer, and the fully connected backend by replacing them with quaternion versions. These modifications lead to novel architectures which yield improved accuracy performance on the ImageNet300k classification dataset. Our baseline networks for comparison were the original real-valued ResNet, the original quaternion-valued ResNet, and the Axial Attention ResNet. Since improvement was observed regardless of which part of the network was modified, there is a promise that this technique may be generally useful in improving classification accuracy for a large class of networks.

Keywords: axial attention, representational networks, weight sharing, cross-channel correlations, quaternion-enhanced axial attention, deep networks

Procedia PDF Downloads 84

Authors: Emmanuel Kamal Aziz Saba, Sarah Sayed El-Tawab

Abstract:

The present study was conducted to assess the involvement of ulnar sensory and/or motor nerve fibers in patients with carpal tunnel syndrome (CTS) and whether this affects the accuracy of the median versus ulnar comparative tests. The present study included 145 CTS hands and 71 asymptomatic control hands. Clinical examination was done. The following tests were done: Sensory conduction studies: median, ulnar and dorsal ulnar cutaneous nerves; and median versus ulnar digit (D) four sensory comparative study; and motor conduction studies: median nerve, ulnar nerve and median versus ulnar motor comparative study. It was found that 17 CTS hands (11.7%) had ulnar sensory abnormalities in 17 different patients. The median versus ulnar sensory and motor comparative studies were abnormal among all these 17 CTS hands. There were significant negative correlations between median motor latency and both ulnar sensory amplitudes recording D5 and D4. In conclusion, there is ulnar sensory nerve abnormality among CTS patients. This abnormality affects the amplitude of ulnar sensory nerve action potential. This does not affect the median versus ulnar sensory and motor comparative tests accuracy for use in CTS.

Keywords: median nerve, motor comparative study, sensory comparative study, ulnar nerve

Procedia PDF Downloads 432

Authors: Yang Yang, Dan Liu

Abstract:

Anomaly detection models for network flow based on machine learning have poor detection performance under extremely unbalanced training data conditions and also have slow detection speed and large resource consumption when deploying on network edge devices. Embedding multi-teacher knowledge distillation (mKD) in anomaly detection can transfer knowledge from multiple teacher models to a single model. Inspired by this, we proposed a state-of-the-art model, mKDNAD, to improve detection performance. mKDNAD mine and integrate the knowledge of one-dimensional sequence and two-dimensional image implicit in network flow to improve the detection accuracy of small sample classes. The multi-teacher knowledge distillation method guides the train of the student model, thus speeding up the model's detection speed and reducing the number of model parameters. Experiments in the CICIDS2017 dataset verify the improvements of our method in the detection speed and the detection accuracy in dealing with the small sample classes.

Keywords: network flow anomaly detection (NAD), multi-teacher knowledge distillation, machine learning, deep learning

Procedia PDF Downloads 124

Authors: Shibei Huang

Abstract:

In deep energy retrofits, the thermal performance of existing building envelopes is often difficult to determine with a high level of accuracy. For older buildings, the records of existing assemblies are often incomplete or inaccurate. To obtain greater baseline performance accuracy for energy models, in-field measurement tools can be used to obtain data on the thermal performance of the existing assemblies. For a known assembly, these field measurements assist in validating the U-factor estimates. If the field-measured U-factor consistently varies from the calculated prediction, those measurements prompt further study. For an unknown assembly, successful field measurements can provide approximate U-factor evaluation, validate assumptions, or identify anomalies requiring further investigation. Using case studies, this presentation will focus on the non-destructive methods utilizing a set of various field tools to validate the baseline U-factors for a range of existing buildings with various wall assemblies. The lessons learned cover what can be achieved, the limitations of these approaches and tools, and ideas for improving the validity of measurements. Key factors include the weather conditions, the interior conditions, the thermal mass of the measured assemblies, and the thermal profiles of the assemblies in question.

Keywords: existing building, sensor, thermal analysis, retrofit

Procedia PDF Downloads 64

Authors: Xiang Shijie, Zhou Dong, Tian Dan

Abstract:

Speaker recognition is a crucial task in the field of speech processing, aimed at identifying individuals based on their vocal characteristics. However, existing speaker recognition methods face numerous challenges. Traditional methods primarily rely on audio signals, which often suffer from limitations in noisy environments, variations in speaking style, and insufficient sample sizes. Additionally, relying solely on audio features can sometimes fail to capture the unique identity of the speaker comprehensively, impacting recognition accuracy. To address these issues, we propose a multi-modal network architecture that simultaneously processes both audio and text signals. By gradually integrating audio and text features, we leverage the strengths of both modalities to enhance the robustness and accuracy of speaker recognition. Our experiments demonstrate significant improvements with this multi-modal approach, particularly in complex environments, where recognition performance has been notably enhanced. Our research not only highlights the limitations of current speaker recognition methods but also showcases the effectiveness of multi-modal fusion techniques in overcoming these limitations, providing valuable insights for future research.

Keywords: feature fusion, memory network, multimodal input, speaker recognition

Procedia PDF Downloads 39

Authors: Apoorva Vinod, Archana Mathur, Snehanshu Saha

Abstract:

Though there exists a plethora of Activation Functions (AFs) used in single and multiple hidden layer Neural Networks (NN), their behavior always raised curiosity, whether used in combination or singly. The popular AFs –Sigmoid, ReLU, and Tanh–have performed prominently well for shallow and deep architectures. Most of the time, AFs are used singly in multi-layered NN, and, to the best of our knowledge, their performance is never studied and analyzed deeply when used in combination. In this manuscript, we experiment with multi-layered NN architecture (both on shallow and deep architectures; Convolutional NN and VGG16) and investigate how well the network responds to using two different AFs (Sigmoid-Tanh, Tanh-ReLU, ReLU-Sigmoid) used alternately against a traditional, single (Sigmoid-Sigmoid, Tanh-Tanh, ReLUReLU) combination. Our results show that using two different AFs, the network achieves better accuracy, substantially lower loss, and faster convergence on 4 computer vision (CV) and 15 Non-CV (NCV) datasets. When using different AFs, not only was the accuracy greater by 6-7%, but we also accomplished convergence twice as fast. We present a case study to investigate the probability of networks suffering vanishing and exploding gradients when using two different AFs. Additionally, we theoretically showed that a composition of two or more AFs satisfies Universal Approximation Theorem (UAT).

Keywords: activation function, universal approximation function, neural networks, convergence

Procedia PDF Downloads 160

Authors: Shreya Borthakur, Aastha Vartak

Abstract:

This Electroencephalogram (EEG) project on Rapid Visual Serial Processing (RSVP) paradigm explores the temporal dynamics of visual attention processing in response to rapidly presented visual stimuli. The study builds upon previous research that used real-world images in RSVP tasks to understand the emergence of object representations in the human brain. The objectives of the research include investigating the differences in accuracy and reaction times between 5 Hz and 20 Hz presentation rates, as well as examining the prominent brain waves, particularly alpha and beta waves, associated with the attention task. The pre-processing and data analysis involves filtering EEG data, creating epochs for target stimuli, and conducting statistical tests using MATLAB, EEGLAB, Chronux toolboxes, and R. The results support the hypotheses, revealing higher accuracy at a slower presentation rate, faster reaction times for less complex targets, and the involvement of alpha and beta waves in attention and cognitive processing. This research sheds light on how short-term memory and cognitive control affect visual processing and could have practical implications in fields like education.

Keywords: RSVP, attention, visual processing, attentional blink, EEG

Procedia PDF Downloads 71

Authors: Samad Sajjadi

Abstract:

Machine translation is a new subject increasingly being used by academic writers, especially students and researchers whose native language is not English. There are numerous studies conducted on machine translation, but few investigations have assessed the accuracy of machine translation from Persian to English at lexical, semantic, and syntactic levels. Using Groves and Mundt’s (2015) Model of error taxonomy, the current study evaluated Persian-to-English translations produced by two famous online translators, Google Translate and Bing Translator. A total of 240 texts were randomly selected from different academic fields (law, literature, medicine, and mass media), and 60 texts were considered for each domain. All texts were rendered by the two translation systems and then by four human translators. All statistical analyses were applied using SPSS. The results indicated that Google translations were more accurate than the translations produced by the Bing Translator, especially in the domains of medicine (lexis: 186 vs. 225; semantic: 44 vs. 48; syntactic: 148 vs. 264 errors) and mass media (lexis: 118 vs. 149; semantic: 25 vs. 32; syntactic: 110 vs. 220 errors), respectively. Nonetheless, both machines are reasonably accurate in Persian-to-English translation of lexicons and syntactic structures, particularly from mass media and medical texts.

Keywords: machine translations, accuracy, human translation, efficiency

Procedia PDF Downloads 78

Authors: Sangguk Ahn, Youngyu Kim, Jaheon Gu, Gyoutae Park

Abstract:

In this paper, we developed a dual gas leak instrument to detect Hydrocarbon (HC) and Monoxide (CO) gases. To two kinds of gases, it is necessary to design compact structure for sensors. And then it is important to draw sensing circuits such as measuring, amplifying and filtering. After that, it should be well programmed with robust, systematic and module coding methods. In center of them, improvement of accuracy and initial response time are a matter of vital importance. To manufacture distinguished gas leak detector, we applied intrinsically safe explosion-proof structure to lithium ion battery, main circuits, a pump with motor, color LCD interfaces and sensing circuits. On software, to enhance measuring accuracy we used numerical analysis such as Lagrange and Neville interpolation. Performance test result is conducted by using standard Methane with seven different concentrations with three other products. We want raise risk prevention and efficiency of gas safe management through distributing to the field of gas safety. Acknowledgment: This study was supported by Small and Medium Business Administration under the research theme of ‘Commercialized Development of a portable intrinsically safe explosion-proof type dual gas leak detector’, (task number S2456036).

Keywords: gas leak, dual gas detector, intrinsically safe, explosion proof

Procedia PDF Downloads 228