Search results for: graph-based feature filtering method

Authors: P. Michael Preetam Raj, Souri Banerjee, Souvik Kundu

Abstract:

Oxide based memristors were fabricated in order to establish its potential applications in analog/digital electronics. BaTiO₃-BiFeO₃ (BT-BFO) was employed as an active material, whereas platinum (Pt) and Nb-doped SrTiO₃ (Nb:STO) were served as a top and bottom electrodes, respectively. Piezoelectric force microscopy (PFM) was utilized to present the ferroelectricity and repeatable polarization inversion in the BT-BFO, demonstrating its effectiveness for resistive switching. The fabricated memristors exhibited excellent electrical characteristics, such as hysteresis current-voltage (I-V), high on/off ratio, high retention time, cyclic endurance, and low operating voltages. The band-alignment between the active material BT-BFO and the substrate Nb:STO was experimentally investigated using X-Ray photoelectron spectroscopy, and it attributed to staggered heterojunction alignment. An energy band diagram was proposed in order to understand the electrical transport in BT-BFO/Nb:STO heterojunction. It was identified that the I-V curves of these memristors have several discontinuities. Curve fitting technique was utilized to analyse the I-V characteristic, and the obtained I-V equations were found to be parabolic. Utilizing this analysis, a non-linear BT-BFO memristors equivalent circuit model was developed. Interestingly, the obtained equivalent circuit of the BT-BFO memristors mimics the identical electrical performance, those obtained in the fabricated devices. Based on the developed equivalent circuit, a finite state machine (FSM) design was proposed. Efforts were devoted to fabricate the same FSM, and the results were well matched with those in the simulated FSM devices. Its multilevel noise filtering and immunity to external noise characteristics were also studied. Further, the feature of variable negative resistance was established by controlling the current through the memristor.

Keywords: band alignment, finite state machine, polarization inversion, resistive switching

Procedia PDF Downloads 116

Authors: Rajkumar Kolangarakandy

Abstract:

Feature extraction and selection is the primary part of any mammogram classification algorithms. The choice of feature, attribute or measurements have an important influence in any classification system. Discrete Wavelet Transformation (DWT) coefficients are one of the prominent features for representing images in frequency domain. The features obtained after the decomposition of the mammogram images using wavelet transformations have higher dimension. Even though the features are higher in dimension, they were highly correlated and redundant in nature. The dimensionality reduction techniques play an important role in selecting the optimum number of features from the higher dimension data, which are highly correlated. PCA is a mathematical tool that reduces the dimensionality of the data while retaining most of the variation in the dataset. In this paper, a multilevel classification of mammogram images using reduced discrete wavelet transformation coefficients and lazy classifiers is proposed. The classification is accomplished in two different levels. In the first level, mammogram ROIs extracted from the dataset is classified as normal and abnormal types. In the second level, all the abnormal mammogram ROIs is classified into benign and malignant too. A further classification is also accomplished based on the variation in structure and intensity distribution of the images in the dataset. The Lazy classifiers called Kstar, IBL and LWL are used for classification. The classification results obtained with the reduced feature set is highly promising and the result is also compared with the performance obtained without dimension reduction.

Keywords: PCA, wavelet transformation, lazy classifiers, Kstar, IBL, LWL

Procedia PDF Downloads 318

Authors: Ho Yeon Park, Kyoung-Jae Kim

Abstract:

Recently, user’s interests for location-based social network service increases according to the advances of social web and location-based technologies. It may be easy to recommend preferred items if we can use user’s preference, context and social network information simultaneously. In this study, we propose context-aware POI (point-of-interests) recommender systems using location-based network analysis and sentiment analysis which consider context, social network information and implicit user’s preference score. We propose a context-aware POI recommendation system consisting of three sub-modules and an integrated recommendation system of them. First, we will develop a recommendation module based on network analysis. This module combines social network analysis and cluster-indexing collaboration filtering. Next, this study develops a recommendation module using social singular value decomposition (SVD) and implicit SVD. In this research, we will develop a recommendation module that can recommend preference scores based on the frequency of POI visits of user in POI recommendation process by using social and implicit SVD which can reflect implicit feedback in collaborative filtering. We also develop a recommendation module using them that can estimate preference scores based on the recommendation. Finally, this study will propose a recommendation module using opinion mining and emotional analysis using data such as reviews of POIs extracted from location-based social networks. Finally, we will develop an integration algorithm that combines the results of the three recommendation modules proposed in this research. Experimental results show the usefulness of the proposed model in relation to the recommended performance.

Keywords: sentiment analysis, network analysis, recommender systems, point-of-interests, business analytics

Procedia PDF Downloads 226

Authors: Armin Nabaei, M. Omair Ahmad, M. N. S. Swamy

Abstract:

Many facial expression and emotion recognition methods in the traditional approaches of using LDA, PCA, and EBGM have been proposed. In recent years deep learning models have provided a unique platform addressing by automatically extracting the features for the detection of facial expression and emotions. However, deep networks require large training datasets to extract automatic features effectively. In this work, we propose an efficient emotion detection algorithm using face images when only small datasets are available for training. We design a deep network whose feature extraction capability is enhanced by utilizing several parallel modules between the input and output of the network, each focusing on the extraction of different types of coarse features with fined grained details to break the symmetry of produced information. In fact, we leverage long range dependencies, which is one of the main drawback of CNNs. We develop this work by introducing a Dynamic Soft-Margin SoftMax.The conventional SoftMax suffers from reaching to gold labels very soon, which take the model to over-fitting. Because it’s not able to determine adequately discriminant feature vectors for some variant class labels. We reduced the risk of over-fitting by using a dynamic shape of input tensor instead of static in SoftMax layer with specifying a desired Soft- Margin. In fact, it acts as a controller to how hard the model should work to push dissimilar embedding vectors apart. For the proposed Categorical Loss, by the objective of compacting the same class labels and separating different class labels in the normalized log domain.We select penalty for those predictions with high divergence from ground-truth labels.So, we shorten correct feature vectors and enlarge false prediction tensors, it means we assign more weights for those classes with conjunction to each other (namely, “hard labels to learn”). By doing this work, we constrain the model to generate more discriminate feature vectors for variant class labels. Finally, for the proposed optimizer, our focus is on solving weak convergence of Adam optimizer for a non-convex problem. Our noteworthy optimizer is working by an alternative updating gradient procedure with an exponential weighted moving average function for faster convergence and exploiting a weight decay method to help drastically reducing the learning rate near optima to reach the dominant local minimum. We demonstrate the superiority of our proposed work by surpassing the first rank of three widely used Facial Expression Recognition datasets with 93.30% on FER-2013, and 16% improvement compare to the first rank after 10 years, reaching to 90.73% on RAF-DB, and 100% k-fold average accuracy for CK+ dataset, and shown to provide a top performance to that provided by other networks, which require much larger training datasets.

Keywords: computer vision, facial expression recognition, machine learning, algorithms, depp learning, neural networks

Procedia PDF Downloads 54

Authors: Edward K. Boahen, Brunel E. Bouya-Moko, Changda Wang

Abstract:

The major threat to online social network (OSN) users is account compromisation. Spammers now spread malicious messages by exploiting the trust relationship established between account owners and their friends. The challenge in detecting a compromised account by service providers is validating the trusted relationship established between the account owners, their friends, and the spammers. Another challenge is the increase in required human interaction with the feature selection. Research available on supervised learning (machine learning) has limitations with the feature selection and accounts that cannot be profiled, like application programming interface (API). Therefore, this paper discusses the various behaviours of the OSN users and the current approaches in detecting a compromised OSN account, emphasizing its limitations and challenges. We propose a deep learning approach that addresses and resolve the constraints faced by the previous schemes. We detailed our proposed optimized nonsymmetric deep auto-encoder (OPT_NDAE) for unsupervised feature learning, which reduces the required human interaction levels in the selection and extraction of features. We evaluated our proposed classifier using the NSL-KDD and KDDCUP'99 datasets in a graphical user interface enabled Weka application. The results obtained indicate that our proposed approach outperformed most of the traditional schemes in OSN compromised account detection with an accuracy rate of 99.86%.

Keywords: computer security, network security, online social network, account compromisation

Procedia PDF Downloads 96

Authors: Hanbey Hazar, Hakan Gul, Ugur Ozturk

Abstract:

In order to decrease the hazardous emissions of the internal combustion engines and to improve the combustion and thermal efficiency, thermal barrier coatings are applied. In this experimental study, cylinder, piston, exhaust, and inlet valves which are combustion chamber components have been coated with a ceramic material, and this earned the engine LHR feature. Cylinder, exhaust and inlet valves of the diesel engine used in the tests were coated with ekabor-2 commercial powder, which is a ceramic material, to a thickness of 50 µm, by using the boriding method. The piston of a diesel engine was coated in 300 µm thickness with bor-based powder by using plasma coating method. Pumpkin seeds oil methyl ester (PSME) was produced by the transesterification method. In addition, dimethoxymethane additive materials were used to improve the properties of diesel fuel, pumpkin seeds oil methyl ester (PSME) and its mixture. Dimethoxymethane was blended with test fuels, which was used as a pilot fuel, at the volumetric ratios of 4% and 8%. Due to thermal barrier coating, the diesel engine's CO, HC, and smoke density values decreased; but, NOx and exhaust gas temperature (EGT) increased.

Keywords: boriding, diesel engine, exhaust emission, thermal barrier coating

Procedia PDF Downloads 464

Authors: Zahra Akbari, Farzin Zokaee, Talat Ghomashchi

Abstract:

Whey is a by-product of the dairy industry whose major components are lactose (44–52 g/L), proteins (6–8 g/L) and mineral salts (4–9 g/L). Approximately 50% of 121 million tons of whey produced in the world in 1993 were disposed into rivers, lakes or other water bodies, treated in wastewater treatment plants or loaded onto land. This represents a significant loss of resources and causes serious pollution problems since whey is a heavy organic pollutant with high COD and BOD values, 40–60 g/L and 50–80 g/L, respectively. The removal of cheese whey proteins and minerals represent an important task both in environmental and in food sciences. The most important treatments which are considered in this study, have been done by using lime, Al2O3, FeCl3 and AlCl3 along with heating and also acidic-alkaline method. Results show that the best way for removal of protein is accomplished with adding HCl to decrease pH from 6 to 4, boiling for 20 min, and filtering protein aggregates. Also partial demineralization in whey solution for reducing ash is accomplished by adding NaOH to increase pH to 7.2 and heating solution for 20 min.

Keywords: whey treatment, dairy industry, precipitation, protein, mineral

Procedia PDF Downloads 392

Authors: Paulo Fernando Silva Filho, Elcio Hideiti Shiguemori

Abstract:

The selection of specific landmarks for an Unmanned Aerial Vehicles’ Visual Navigation systems based on Automatic Landmark Recognition has significant influence on the precision of the system’s estimated position. At the same time, manual selection of the landmarks does not guarantee a high recognition rate, which would also result on a poor precision. This work aims to develop an automatic landmark selection that will take the image of the flight area and identify the best landmarks to be recognized by the Visual Navigation Landmark Recognition System. The criterion to select a landmark is based on features detected by ORB or AKAZE and edges information on each possible landmark. Results have shown that disposition of possible landmarks is quite different from the human perception.

Keywords: clustering, edges, feature points, landmark selection, X-means

Procedia PDF Downloads 258

Authors: Çiğdem Bakır, Mecit Yüzkat

Abstract:

With the increasing emphasis on water quality worldwide, the search for and expanding the market for new and intelligent monitoring systems has increased. The current method is the laboratory process, where samples are taken from bodies of water, and tests are carried out in laboratories. This method is time-consuming, a waste of manpower, and uneconomical. To solve this problem, we used machine learning methods to detect water pollution in our study. We created decision trees with the Orange3 software we used in our study and tried to determine all the factors that cause water pollution. An automatic prediction model based on water quality was developed by taking many model inputs such as water temperature, pH, transparency, conductivity, dissolved oxygen, and ammonia nitrogen with machine learning methods. The proposed approach consists of three stages: preprocessing of the data used, feature detection, and classification. We tried to determine the success of our study with different accuracy metrics and the results. We presented it comparatively. In addition, we achieved approximately 98% success with the decision tree.

Keywords: decision tree, water quality, water pollution, machine learning

Procedia PDF Downloads 66

Authors: Mohamed El Morsy, Gabriela Achtenová

Abstract:

Condition monitoring is used to increase machinery availability and machinery performance, whilst reducing consequential damage, increasing machine life, reducing spare parts inventories, and reducing breakdown maintenance. An efficient condition monitoring system provides early warning of faults by predicting them at an early stage. When a localized fault occurs in gears, the vibration signals always exhibit non-stationary behavior. The periodic impulsive feature of the vibration signal appears in the time domain and the corresponding gear mesh frequency (GMF) emerges in the frequency domain. However, one limitation of frequency-domain analysis is its inability to handle non-stationary waveform signals, which are very common when machinery faults occur. Particularly at the early stage of gear failure, the GMF contains very little energy and is often overwhelmed by noise and higher-level macro-structural vibrations. An effective signal processing method would be necessary to remove such corrupting noise and interference. In this paper, a new hybrid method based on optimal Morlet wavelet filter and autocorrelation enhancement is presented. First, to eliminate the frequency associated with interferential vibrations, the vibration signal is filtered with a band-pass filter determined by a Morlet wavelet whose parameters are selected or optimized based on maximum Kurtosis. Then, to further reduce the residual in-band noise and highlight the periodic impulsive feature, an autocorrelation enhancement algorithm is applied to the filtered signal. The test stand is equipped with three dynamometers; the input dynamometer serves as the internal combustion engine, the output dynamometers induce a load on the output joint shaft flanges. The pitting defect is manufactured on the tooth side of a gear of the fifth speed on the secondary shaft. The gearbox used for experimental measurements is of the type most commonly used in modern small to mid-sized passenger cars with transversely mounted powertrain and front wheel drive: a five-speed gearbox with final drive gear and front wheel differential. The results obtained from practical experiments prove that the proposed method is very effective for gear fault diagnosis.

Keywords: wavelet analysis, pitted gear, autocorrelation, gear fault diagnosis

Procedia PDF Downloads 369

Authors: Baptiste Bouvier

Abstract:

While we are constantly flooded with stimuli in daily life, attention allows us to select the ones we specifically process and ignore the others. Some salient stimuli may sometimes pass this filter independently of our will, in a "bottom-up" way. The role of the acoustic properties of the timbre of a sound on its salience, i.e., its ability to capture the attention of a listener, is still not well understood. We implemented a paradigm called the "additional singleton paradigm", in which participants have to discriminate targets according to their duration. This task is perturbed (higher error rates and longer response times) by the presence of an irrelevant additional sound, of which we can manipulate a feature of our choice at equal loudness. This allows us to highlight the influence of the timbre features of a sound stimulus on its salience at equal loudness. We have shown that a stimulus that is brighter than the others but not louder leads to an attentional capture phenomenon in this framework. This work opens the door to the study of the influence of any timbre feature on salience.

Keywords: attention, audition, bottom-up attention, psychoacoustics, salience, timbre

Procedia PDF Downloads 147

Authors: Misagh Irandoost Shahrestani, Mehdi Ashjaee, Shahrokh Zandieh Vakili

Abstract:

In this paper, partially premixed laminar methane/air co-flow flame is studied experimentally. Methane-air flame was established on an axisymmetric coannular burner. The fuel-air jet flows from the central tube while the secondary air flows from the region between the inner and the outer tube. The aim is to investigate the flame features and to develop a nonintrusive method for temperature measurement of methane/air partially premixed flame using Mach-Zehnder interferometry method. Different equivalence ratios and Reynolds numbers are considered. Flame generic visible appearance was also investigated and its various structures were studied. Three distinguished flame regimes were seen based on its appearance. A double flame structure can be seen for the equivalence ratio in the range of 1<Φ<2.1. By adding air to the mixture up to Φ=4 the flame has the characteristics of both premixed and non-premixed flames. Finally for 4<Φ<∞ the flame mainly becomes non-premixed like and the luminous sooting region on its tip is the obvious feature of this type of flames. The Mach-Zehnder method is used to obtain temperature field of a transparent fluid by means of index of refraction. Temperature obtained from optical techniques was compared with that of obtained from thermocouples in order to validate the results. Good agreement was observed for these two methods.

Keywords: flame structure, Mach-Zehnder interferometry, partially premixed flame, temperature field

Procedia PDF Downloads 445

Authors: Ronak Khosravi, Mahmood Abbasi Layegh, Siamak Haghipour, Avin Esmaili

Abstract:

In this paper, a study on classification of musical woodwind instruments using a small set of features selected from a broad range of extracted ones by the sequential forward selection method was carried out. Firstly, we extract 42 features for each record in the music database of 402 sound files belonging to five different groups of Flutes (end blown and internal duct), Single –reed, Double –reed (exposed and capped), Triple reed and Quadruple reed. Then, the sequential forward selection method is adopted to choose the best feature set in order to achieve very high classification accuracy. Two different classification techniques of support vector machines and relevance vector machines have been tested out and an accuracy of up to 96% can be achieved by using 21 time, frequency and coefficient features and relevance vector machine with the Gaussian kernel function.

Keywords: coefficient features, relevance vector machines, spectral features, support vector machines, temporal features

Procedia PDF Downloads 296

Authors: S. Rozhkova, O. Rozhkova, A. Harlova, V. Lasukov

Abstract:

For optimal unbiased filter as mean-square and in the case of functioning anomalous noises in the observation memory channel, we have proved insensitivity of filter to inaccurate knowledge of the anomalous noise intensity matrix and its equivalence to truncated filter plotted only by non anomalous components of an observation vector.

Keywords: mathematical expectation, filtration, anomalous noise, memory

Procedia PDF Downloads 342

Authors: Paiboon Jeamponk, Tasanee Ponglaa, Patchapon Srisanguan

Abstract:

The aim of this research paper was based on an examination of sources of water supply and water quality for local consumption, conducted at eco-tourism villages of Suan Luang Sub- District Municipality of Amphawa District, Samut Songkram Province. The study incorporated both questionnaire and field work of water testing as the research tool and method. The sample size of 288 households was based on the population of the district, whereas the selected sample water sources were from 60 households: 30 samples were ground water and another 30 were surface water. Degree of heavy metal contamination in the water including copper, iron, manganese, zinc, cadmium and lead was investigated utilizing the Atomic Absorption- Direct Aspiration method. The findings unveiled that 96.0 percent of household water consumption was based on water supply, while the rest on canal, river and rain water. The household behaviour of consumption revealed that 47.2 percent of people routinely consumed water without boiling or filtering prior to consumption. The investigation of water supply quality found that the degree of heavy metal contamination including metal, lead, iron, copper, manganese and cadmium met the standards of the Department of Health.

Keywords: sources of water supply, water quality, water supply, Thailand

Procedia PDF Downloads 278

Authors: Sudha T., Naveen C.

Abstract:

Dengue is a serious public health issue that causes significant annual economic and welfare burdens on nations. However, enhanced optimization techniques and quantitative modeling approaches can predict the incidence of dengue. By advocating for a data-driven approach, public health officials can make informed decisions, thereby improving the overall effectiveness of sudden disease outbreak control efforts. The National Oceanic and Atmospheric Administration and the Centers for Disease Control and Prevention are two of the U.S. Federal Government agencies from which this study uses environmental data. Based on environmental data that describe changes in temperature, precipitation, vegetation, and other factors known to affect dengue incidence, many predictive models are constructed that use different machine learning methods to estimate weekly dengue cases. The first step involves preparing the data, which includes handling outliers and missing values to make sure the data is prepared for subsequent processing and the creation of an accurate forecasting model. In the second phase, multiple feature selection procedures are applied using various machine learning models and optimization techniques. During the third phase of the research, machine learning models like the Huber Regressor, Support Vector Machine, Gradient Boosting Regressor (GBR), and Support Vector Regressor (SVR) are compared with several optimization techniques for feature selection, such as Harmony Search and Genetic Algorithm. In the fourth stage, the model's performance is evaluated using Mean Square Error (MSE), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) as assistance. Selecting an optimization strategy with the least number of errors, lowest price, biggest productivity, or maximum potential results is the goal. In a variety of industries, including engineering, science, management, mathematics, finance, and medicine, optimization is widely employed. An effective optimization method based on harmony search and an integrated genetic algorithm is introduced for input feature selection, and it shows an important improvement in the model's predictive accuracy. The predictive models with Huber Regressor as the foundation perform the best for optimization and also prediction.

Keywords: deep learning model, dengue fever, prediction, optimization

Procedia PDF Downloads 34

Authors: Xiaocong Liu, Huazhen Wang, Ting He, Xiaozheng Li, Weihan Zhang, Jian Chen

Abstract:

The utilization of electronic medical record (EMR) data to establish the disease diagnosis model has become an important research content of biomedical informatics. Deep learning can automatically extract features from the massive data, which brings about breakthroughs in the study of EMR data. The challenge is that deep learning lacks semantic knowledge, which leads to impracticability in medical science. This research proposes a method of incorporating lexical-semantic knowledge from abundant entities into a convolutional neural network (CNN) framework for pediatric disease diagnosis. Firstly, medical terms are vectorized into Lexical Semantic Vectors (LSV), which are concatenated with the embedded word vectors of word2vec to enrich the feature representation. Secondly, the semantic distribution of medical terms serves as Semantic Decision Guide (SDG) for the optimization of deep learning models. The study evaluate the performance of LSV-SDG-CNN model on four kinds of Chinese EMR datasets. Additionally, CNN, LSV-CNN, and SDG-CNN are designed as baseline models for comparison. The experimental results show that LSV-SDG-CNN model outperforms baseline models on four kinds of Chinese EMR datasets. The best configuration of the model yielded an F1 score of 86.20%. The results clearly demonstrate that CNN has been effectively guided and optimized by lexical-semantic knowledge, and LSV-SDG-CNN model improves the disease classification accuracy with a clear margin.

Keywords: convolutional neural network, electronic medical record, feature representation, lexical semantics, semantic decision

Procedia PDF Downloads 112

Authors: Hae-Yeoun Lee

Abstract:

Mosaic refers to a technique that makes image by gathering lots of small materials in various colours. This paper presents an automatic algorithm that makes the photomosaic image using photos. The algorithm is composed of four steps: Partition and feature extraction, block matching, redundancy removal and colour adjustment. The input image is partitioned in the small block to extract feature. Each block is matched to find similar photo in database by comparing similarity with Euclidean difference between blocks. The intensity of the block is adjusted to enhance the similarity of image by replacing the value of light and darkness with that of relevant block. Further, the quality of image is improved by minimizing the redundancy of tiles in the adjacent blocks. Experimental results support that the proposed algorithm is excellent in quantitative analysis and qualitative analysis.

Keywords: photomosaic, Euclidean distance, block matching, intensity adjustment

Procedia PDF Downloads 265

Authors: Hu Ding, Kai Liu, Guoan Tang

Abstract:

The Chinese Loess Plateau suffers from serious gully erosion induced by natural and human causes. Gully features detection including gully-affected area and its two dimension parameters (length, width, area et al.), is a significant task not only for researchers but also for policy-makers. This study aims at gully-affected area detection in three catchments of Chinese Loess Plateau, which were selected in Changwu, Ansai, and Suide by using unmanned aerial vehicle (UAV). The methodology includes a sequence of UAV data generation, image segmentation, feature calculation and selection, and random forest classification. Two experiments were conducted to investigate the influences of segmentation strategy and feature selection. Results showed that vertical and horizontal root-mean-square errors were below 0.5 and 0.2 m, respectively, which were ideal for the Loess Plateau region. The segmentation strategy adopted in this paper, which considers the topographic information, and optimal parameter combination can improve the segmentation results. Besides, the overall extraction accuracy in Changwu, Ansai, and Suide achieved was 84.62%, 86.46%, and 93.06%, respectively, which indicated that the proposed method for detecting gully-affected area is more objective and effective than traditional methods. This study demonstrated that UAV can bridge the gap between field measurement and satellite-based remote sensing, obtaining a balance in resolution and efficiency for catchment-scale gully erosion research.

Keywords: unmanned aerial vehicle (UAV), object-analysis image analysis, gully erosion, gully-affected area, Loess Plateau, random forest

Procedia PDF Downloads 196

Authors: Ahmed Machmoum, Malika El Kyal

Abstract:

We consider iterative parallel multi-splitting method for differential algebraic equations. The main feature of the proposed idea is to use the asynchronous form. We prove that the multi-splitting technique can effectively accelerate the convergent performance of the iterative process. The main characteristic of an asynchronous mode is that the local algorithm not have to wait at predetermined messages to become available. We allow some processors to communicate more frequently than others, and we allow the communication delays tobe substantial and unpredictable. Note that synchronous algorithms in the computer science sense are particular cases of our formulation of asynchronous one.

Keywords: computer, multi-splitting methods, asynchronous mode, differential algebraic systems

Procedia PDF Downloads 528

Authors: Nishant Rodrigues, Nicole Spanedda, Chilukuri K. Mohan, Arindam Chakraborty

Abstract:

A crucial objective in quantum chemistry is the computation of the energy levels of chemical systems. This task requires electron-repulsion integrals as inputs, and the steep computational cost of evaluating these integrals poses a major numerical challenge in efficient implementation of quantum chemical software. This work presents a moment-based machine-learning approach for the efficient evaluation of electron-repulsion integrals. These integrals were approximated using linear combinations of a small number of moments. Machine learning algorithms were applied to estimate the coefficients in the linear combination. A random forest approach was used to identify promising features using a recursive feature elimination approach, which performed best for learning the sign of each coefficient but not the magnitude. A neural network with two hidden layers were then used to learn the coefficient magnitudes along with an iterative feature masking approach to perform input vector compression, identifying a small subset of orbitals whose coefficients are sufficient for the quantum state energy computation. Finally, a small ensemble of neural networks (with a median rule for decision fusion) was shown to improve results when compared to a single network.

Keywords: quantum energy calculations, atomic orbitals, electron-repulsion integrals, ensemble machine learning, random forests, neural networks, feature extraction

Procedia PDF Downloads 85

Authors: Rui Mao, Heming Ji, Xiaoyu Wang

Abstract:

Keystroke dynamics are widely used in identity recognition. It has the advantage that the individual typing rhythm is difficult to imitate. It also supports continuous authentication through the keyboard without extra devices. The existing keystroke dynamics authentication methods based on machine learning have a drawback in supporting relatively complex scenarios with massive data. There are drawbacks to both feature extraction and model optimization in these methods. To overcome the above weakness, an authentication model of keystroke dynamics based on deep learning is proposed. The model uses feature vectors formed by keystroke content and keystroke time. It ensures efficient continuous authentication by cooperating attention mechanisms with the combination of CNN and Bi-LSTM. The model has been tested with Open Data Buffalo dataset, and the result shows that the FRR is 3.09%, FAR is 3.03%, and EER is 4.23%. This proves that the model is efficient and accurate on continuous authentication.

Keywords: keystroke dynamics, identity authentication, deep learning, CNN, LSTM

Procedia PDF Downloads 131

Authors: Victor Breux, Jérôme Boutet, Alain Goret, Viviane Cattin

Abstract:

Early detection of anomalies in data centers is important to reduce downtimes and the costs of periodic maintenance. However, there is little research on this topic and even fewer on the fusion of sensor data for the detection of abnormal events. The goal of this paper is to propose a method for anomaly detection in data centers by combining sensor data (temperature, humidity, power) and deep learning models. The model described in the paper uses one autoencoder per sensor to reconstruct the inputs. The auto-encoders contain Long-Short Term Memory (LSTM) layers and are trained using the normal samples of the relevant sensors selected by correlation analysis. The difference signal between the input and its reconstruction is then used to classify the samples using feature extraction and a random forest classifier. The data measured by the sensors of a data center between January 2019 and May 2020 are used to train the model, while the data between June 2020 and May 2021 are used to assess it. Performances of the model are assessed a posteriori through F1-score by comparing detected anomalies with the data center’s history. The proposed model outperforms the state-of-the-art reconstruction method, which uses only one autoencoder taking multivariate sequences and detects an anomaly with a threshold on the reconstruction error, with an F1-score of 83.60% compared to 24.16%.

Keywords: anomaly detection, autoencoder, data centers, deep learning

Procedia PDF Downloads 167

Authors: Abobaker Mohammed Qasem Farhan, Shangming Yang, Mohammed Al-Nehari

Abstract:

Pneumonia is a kind of lung disease that creates congestion in the chest. Such pneumonic conditions lead to loss of life of the severity of high congestion. Pneumonic lung disease is caused by viral pneumonia, bacterial pneumonia, or Covidi-19 induced pneumonia. The early prediction and classification of such lung diseases help to reduce the mortality rate. We propose the automatic Computer-Aided Diagnosis (CAD) system in this paper using the deep learning approach. The proposed CAD system takes input from raw computerized tomography (CT) scans of the patient's chest and automatically predicts disease classification. We designed the Hybrid Deep Learning Algorithm (HDLA) to improve accuracy and reduce processing requirements. The raw CT scans have pre-processed first to enhance their quality for further analysis. We then applied a hybrid model that consists of automatic feature extraction and classification. We propose the robust 2D Convolutional Neural Network (CNN) model to extract the automatic features from the pre-processed CT image. This CNN model assures feature learning with extremely effective 1D feature extraction for each input CT image. The outcome of the 2D CNN model is then normalized using the Min-Max technique. The second step of the proposed hybrid model is related to training and classification using different classifiers. The simulation outcomes using the publically available dataset prove the robustness and efficiency of the proposed model compared to state-of-art algorithms.

Keywords: CT scan, Covid-19, deep learning, image processing, lung disease classification

Procedia PDF Downloads 124

Authors: Othman Maklouf, Ahmed Adwaib

Abstract:

This paper introduces a comparative study between the main GPS/INS coupling schemes, this will include the loosely coupled and tightly coupled configurations, several types of situations and operational conditions, in which the data fusion process is done using Kalman filtering. This will include the importance of sensors calibration as well as the alignment of the strap down inertial navigation system. The limitations of the inertial navigation systems are investigated.

Keywords: GPS, INS, Kalman filter, sensor calibration, navigation system

Procedia PDF Downloads 567

Authors: Mohamad Kassab, Amal El Fallah Seghrouchni, Frederic Barbaresco, Raed Abu Zitar

Abstract:

Detecting objects such as drones is a challenging task as their relative size and maneuvering capabilities deceive machine learning models and cause them to misclassify drones as birds or other objects. In this work, we investigate applying several deep learning techniques to benchmark real data sets of flying drones. A deep learning paradigm is proposed for the purpose of mitigating the complexity of those systems. The proposed paradigm consists of a hybrid between the AdderNet deep learning paradigm and the Single Shot Detector (SSD) paradigm. The goal was to minimize multiplication operations numbers in the filtering layers within the proposed system and, hence, reduce complexity. Some standard machine learning technique, such as SVM, is also tested and compared to other deep learning systems. The data sets used for training and testing were either complete or filtered in order to remove the images with mall objects. The types of data were RGB or IR data. Comparisons were made between all these types, and conclusions were presented.

Keywords: drones detection, deep learning, birds versus drones, precision of detection, AdderNet

Procedia PDF Downloads 156

Authors: Mona Hejazi, Ali Motie Nasrabadi

Abstract:

Epilepsy is a persistent neurological disorder that affects more than 50 million people worldwide. Hence, there is a necessity to introduce an efficient prediction model for making a correct diagnosis of the epileptic seizure and accurate prediction of its type. In this study we consider how the Effective Connectivity (EC) patterns obtained from intracranial Electroencephalographic (EEG) recordings reveal information about the dynamics of the epileptic brain and can be used to predict imminent seizures, as this will enable the patients (and caregivers) to take appropriate precautions. We use this definition because we believe that effective connectivity near seizures begin to change, so we can predict seizures according to this feature. Results are reported on the standard Freiburg EEG dataset which contains data from 21 patients suffering from medically intractable focal epilepsy. Six channels of EEG from each patients are considered and effective connectivity using Directed Transfer Function (DTF) and Granger Causality (GC) methods is estimated. We concentrate on effective connectivity standard deviation over time and feature changes in five brain frequency sub-bands (Alpha, Beta, Theta, Delta, and Gamma) are compared. The performance obtained for the proposed scheme in predicting seizures is: average prediction time is 50 minutes before seizure onset, the maximum sensitivity is approximate ~80% and the false positive rate is 0.33 FP/h. DTF method is more acceptable to predict epileptic seizures and generally we can observe that the greater results are in gamma and beta sub-bands. The research of this paper is significantly helpful for clinical applications, especially for the exploitation of online portable devices.

Keywords: effective connectivity, Granger causality, directed transfer function, epilepsy seizure prediction, EEG

Procedia PDF Downloads 442

Authors: Yan Lyu, Yiqun Pan, Zhizhong Huang

Abstract:

In the design stage of a new building, the energy model of this building is often required for the analysis of the performance on energy efficiency. In practice, a certain degree of geometric simplification should be done in the establishment of building energy models, since the detailed geometric features of a real building are hard to be described perfectly in most energy simulation engine, such as ESP-r, eQuest or EnergyPlus. Actually, the detailed description is not necessary when the result with extremely high accuracy is not demanded. Therefore, this paper analyzed the relationship between the error of the simulation result from building energy models and the geometric simplification of the models. Finally, the following two parameters are selected as the indices to characterize the geometric feature of in building energy simulation: the southward projected area and total side surface area of the building, Based on the parameterization method, the simplification from an arbitrary column building to a typical shape (a cuboid) building can be made for energy modeling. The result in this study indicates that this simplification would only lead to the error that is less than 7% for those buildings with the ratio of southward projection length to total perimeter of the bottom of 0.25~0.35, which can cover most situations.

Keywords: building energy model, simulation, geometric simplification, design, regression

Procedia PDF Downloads 160

Authors: Adrià Arbués-Sangüesa, Coloma Ballester, Gloria Haro

Abstract:

Tracking sports players is a widely challenging scenario, specially in single-feed videos recorded in tight courts, where cluttering and occlusions cannot be avoided. This paper presents an analysis of several geometric and semantic visual features to detect and track basketball players. An ablation study is carried out and then used to remark that a robust tracker can be built with Deep Learning features, without the need of extracting contextual ones, such as proximity or color similarity, nor applying camera stabilization techniques. The presented tracker consists of: (1) a detection step, which uses a pretrained deep learning model to estimate the players pose, followed by (2) a tracking step, which leverages pose and semantic information from the output of a convolutional layer in a VGG network. Its performance is analyzed in terms of MOTA over a basketball dataset with more than 10k instances.

Keywords: basketball, deep learning, feature extraction, single-camera, tracking

Procedia PDF Downloads 117

Authors: Ratul Das

Abstract:

The present experimental study insights the decontamination of instantaneous velocity fluctuations captured by Acoustic Doppler Velocimeter (ADV) in gravel-bed streams to ascertain near-bed turbulence for low Reynolds number. The interference between incidental and reflected pulses produce spikes in the ADV data especially in the near-bed flow zone and therefore filtering the data are very essential. Nortek’s Vectrino four-receiver ADV probe was used to capture the instantaneous three-dimensional velocity fluctuations over a non-cohesive bed. A spike removal algorithm based on the acceleration threshold method was applied to note the bed roughness and its influence on velocity fluctuations and velocity power spectra in the carrier fluid. The velocity power spectra of despiked signals with a best combination of velocity threshold (VT) and acceleration threshold (AT) are proposed which ascertained velocity power spectra a satisfactory fit with the Kolmogorov “–5/3 scaling-law” in the inertial sub-range. Also, velocity distributions below the roughness crest level fairly follows a third-degree polynomial series.

Keywords: acoustic doppler velocimeter, gravel-bed, spike removal, reynolds shear stress, near-bed turbulence, velocity power spectra

Procedia PDF Downloads 285