Search results for: convolution neural network- long short term memory network (CNN-LSTM)

Authors: Pia Berner Hansson, Robert Murison Anders Lund, Hammar Åsa

Abstract:

Major Depressive Disorder (MDD) is often associated with high levels of stress and disturbances in the Hypothalamic Pituitary Adrenal (HPA) system, yielding high levels of cortisol, in addition to cognitive dysfunction. Previous studies in this patient group have shown a relationship between cortisol profile and cognitive functioning in the acute phase of MDD and that the patients had significantly less suppression after dexamethasone administration. However, few studies have investigated this relationship over time and in phases of symptom reduction. The aim of the present study was to examine the relationships between cortisol levels after the Dexamethasone Suppression Test (DST) and cognitive function in a long term perspective in MDD patients. Patients meeting the DSM-IV criteria for a MDD were included in the study and tested in symptom reduction. A control group was included. Cortisol was measured in saliva collected with Salivette sampling devices. Saliva samples were collected 4 times during a 24 hours period over two consecutive days: at awakening, after 45 minutes, after 7 hours and at 11 pm. Dexamethasone (1.0 mg) was given on Day 1 at 11 pm. The neuropsychological test battery consisted of standardized tests measuring memory and Executive Functioning (EF). Cortisol levels did not differ significantly between patients and controls on Day 1 or Day 2. Both groups showed significant suppression after Dexamethasone. There were no correlations between cortisol levels or suppression after Dexamethasone and cognitive measures. The results indicate that the HPA-axis functioning normalizes in phases of symptom reduction in MDD patients and that there no relation between cortisol profile and cognitive functioning in memory or EF.

Keywords: depression, MDD, cortisol, suppression, cognitive functioning

Procedia PDF Downloads 317

Authors: Juzhong Tan, William Kerr

Abstract:

Roasting is one critical procedure in chocolate processing, where special favors are developed, moisture content is decreased, and better processing properties are developed. Therefore, determination of roasting degree of cocoa bean is important for chocolate manufacturers to ensure the quality of chocolate products, and it also decides the commercial value of cocoa beans collected from cocoa farmers. The roasting degree of cocoa beans currently relies on human specialists, who sometimes are biased, and chemical analysis, which take long time and are inaccessible to many manufacturers and farmers. In this study, a self-made electronic nose system consists of gas sensors (TGS 800 and 2000 series) was used to detecting the gas generated by cocoa beans with a different roasting degree (0min, 20min, 30min, and 40min) and the signals collected by gas sensors were used to train a three-layers ANN. Chemical analysis of the graded beans was operated by traditional GC-MS system and the contents of volatile chemical compounds were used to train another ANN as a reference to electronic nosed signals trained ANN. Both trained ANN were used to predict cocoa beans with a different roasting degree for validation. The best accuracy of grading achieved by electronic nose signals trained ANN (using signals from TGS 813 826 820 880 830 2620 2602 2610) turned out to be 96.7%, however, the GC trained ANN got the accuracy of 83.8%.

Keywords: artificial neutron network, cocoa bean, electronic nose, roasting

Procedia PDF Downloads 215

Authors: Nopsarun Vannasirikul

Abstract:

The purposes of this research paper were to study the current situations of community development, women’s potentials, women’s participation, network, and government policy as well as to study the factors influencing women’s potentials, women’s participation, network, and government policy that have on the community development. The population included the women age of 18 years old who were living in the communities of Bangkok areas. This study was a mix research method of quantitative and qualitative method. A simple random sampling method was utilized to obtain 400 sample groups from 50 districts of Bangkok and to perform data collection by using questionnaire. Also, a purposive sampling method was utilized to obtain 12 informants for an in-depth interview to gain an in-sight information for quantitative method.

Keywords: community development, participation, network, women’s right, management

Procedia PDF Downloads 153

Authors: Hanbhin Ryu, Ilkwon Yun

Abstract:

Automotive industry targets to provide an improvement in safety and convenience through realizing fully autonomous vehicle. For partially realizing fully automated driving, Current vehicles already feature varieties of advanced driver assistance system (ADAS) for safety and infotainment systems for the driver’s convenience. This paper presents Cloud Connected Vehicle (CCV) which connected vehicles with cloud data center via the access network to control the vehicle for achieving next autonomous driving form and describes its features. This paper also describes the shortcoming of the existing In-Vehicle Network (IVN) to be a next generation IVN of CCV and organize the 802.3 Ethernet, the next generation of IVN, related research issue to verify the feasibility of using Ethernet. At last, this paper refers to additional considerations to adopting Ethernet-based IVN for CCV.

Keywords: autonomous vehicle, cloud connected vehicle, ethernet, in-vehicle network

Procedia PDF Downloads 461

Authors: Alemayehu Muluneh, Saskia Keesstra, Leo Stroosnijder, Woldeamlak Bewket, Ashenafi Burka

Abstract:

There are some scientific evidences and a belief by many that forests attract rain and deforestation contributes to a decline of rainfall. However, there is still a lack of concrete scientific evidence on the role of forests in rainfall amount. In this paper, we investigate the forest-rainfall relationships in the environmentally hot spot area of the Central Rift Valley (CRV) of Ethiopia. Specifically, we evaluate long term (1970-2009) rainfall variability and its relationship with historical forest cover and the relationship between existing forest cover and topographical variables and rainfall distribution. The study used 16 long term and 15 short term rainfall stations. The Mann-Kendall test, bi variate and multiple regression models were used. The results show forest and wood land cover continuously declined over the 40 years period (1970-2009), but annual rainfall in the rift valley floor increased by 6.42 mm/year. But, on the escarpment and highlands, annual rainfall decreased by 2.48 mm/year. The increase in annual rainfall in the rift valley floor is partly attributable to the increase in evaporation as a result of increasing temperatures from the 4 existing lakes in the rift valley floor. Though, annual rainfall is decreasing on the escarpment and highlands, there was no significant correlation between this rainfall decrease and forest and wood land decline and also rainfall variability in the region was not explained by forest cover. Hence, the decrease in annual rainfall on the escarpment and highlands is likely related to the global warming of the atmosphere and the surface waters of the Indian Ocean. Spatial variability of number of rainy days from systematically observed two-year’s rainfall data (2012-2013) was significantly (R2=-0.63) explained by forest cover (distance from forest). But, forest cover was not a significant variable (R2=-0.40) in explaining annual rainfall amount. Generally, past deforestation and existing forest cover showed very little effect on long term and short term rainfall distribution, but a significant effect on number of rainy days in the CRV of Ethiopia.

Keywords: elevation, forest cover, rainfall, slope

Procedia PDF Downloads 525

Authors: Amrah Maryam

Abstract:

The emergence of online social networks (OSNs) has transformed the way we pursue and share information. On the one hand, OSNs provide great ease for the spreading of positive information while, on the other hand, they may also become a channel for the spreading of malicious rumors and misinformation throughout the social network. Thus, to assure the trustworthiness of OSNs to its users, it is of vital importance to detect the misinformation propagation in the network by placing network monitors. In this paper, we aim to place monitors near the suspected nodes with the intent to limit the diffusion of misinformation in the social network, and then we also detect the most significant nodes in the network for propagating true information in order to minimize the effect of already diffused misinformation. Thus, we initiate two heuristic monitor placement using articulation points and truth propagation using eigenvector centrality. Furthermore, to provide real-time workings of the system, we integrate both the monitor placement and truth propagation entities as well. To signify the effectiveness of the approaches, we have carried out the experiment and evaluation of Stanford datasets of online social networks.

Keywords: online social networks, monitor placement, independent cascade model, spread of misinformation

Procedia PDF Downloads 145

Authors: Tapan Jain, Davender Singh Saini

Abstract:

Clustering is a useful mechanism in wireless sensor networks which helps to cope with scalability and data transmission problems. The basic aim of our research work is to provide efficient clustering using Hierarchical agglomerative clustering (HAC). If the distance between the sensing nodes is calculated using their location then it’s quantitative HAC. This paper compares the various agglomerative clustering techniques applied in a wireless sensor network using the quantitative data. The simulations are done in MATLAB and the comparisons are made between the different protocols using dendrograms.

Keywords: routing, hierarchical clustering, agglomerative, quantitative, wireless sensor network

Procedia PDF Downloads 581

Authors: Wanderson Goncalves e Goncalves, José Alberto Silva de Sá

Abstract:

Human actions are a threat to the balance and conservation of the Amazon forest. Therefore the environmental monitoring services play an important role as the preservation and maintenance of this environment. This study classified forest types using data from a forest inventory provided by the 'Florestal e da Biodiversidade do Estado do Pará' (IDEFLOR-BIO), located between the municipalities of Santarém, Juruti and Aveiro, in the state of Pará, Brazil, covering an area approximately of 600,000 hectares, Bands 3, 4 and 5 of the TM-Landsat satellite image, and Self - Organizing Maps. The information from the satellite images was extracted using QGIS software 2.8.1 Wien and was used as a database for training the neural network. The midpoints of each sample of forest inventory have been linked to images. Later the Digital Numbers of the pixels have been extracted, composing the database that fed the training process and testing of the classifier. The neural network was trained to classify two forest types: Rain Forest of Lowland Emerging Canopy (Dbe) and Rain Forest of Lowland Emerging Canopy plus Open with palm trees (Dbe + Abp) in the Mamuru Arapiuns glebes of Pará State, and the number of examples in the training data set was 400, 200 examples for each class (Dbe and Dbe + Abp), and the size of the test data set was 100, with 50 examples for each class (Dbe and Dbe + Abp). Therefore, total mass of data consisted of 500 examples. The classifier was compiled in Orange Data Mining 2.7 Software and was evaluated in terms of the confusion matrix indicators. The results of the classifier were considered satisfactory, and being obtained values of the global accuracy equal to 89% and Kappa coefficient equal to 78% and F1 score equal to 0,88. It evaluated also the efficiency of the classifier by the ROC plot (receiver operating characteristics), obtaining results close to ideal ratings, showing it to be a very good classifier, and demonstrating the potential of this methodology to provide ecosystem services, particularly in anthropogenic areas in the Amazon.

Keywords: artificial neural network, computational intelligence, pattern recognition, unsupervised learning

Procedia PDF Downloads 348

Authors: Paweł Konieczka, Lech Raczyński, Wojciech Wiślicki, Oleksandr Fedoruk, Konrad Klimaszewski, Przemysław Kopka, Wojciech Krzemień, Roman Shopa, Jakub Baran, Aurélien Coussat, Neha Chug, Catalina Curceanu, Eryk Czerwiński, Meysam Dadgar, Kamil Dulski, Aleksander Gajos, Beatrix C. Hiesmayr, Krzysztof Kacprzak, łukasz Kapłon, Grzegorz Korcyl, Tomasz Kozik, Deepak Kumar, Szymon Niedźwiecki, Dominik Panek, Szymon Parzych, Elena Pérez Del Río, Sushil Sharma, Shivani Shivani, Magdalena Skurzok, Ewa łucja Stępień, Faranak Tayefi, Paweł Moskal

Abstract:

This paper develops the transformation of non-image data into 2-dimensional matrices, as a preparation stage for classification based on convolutional neural networks (CNNs). In positron emission tomography (PET) studies, CNN may be applied directly to the reconstructed distribution of radioactive tracers injected into the patient's body, as a pattern recognition tool. Nonetheless, much PET data still exists in non-image format and this fact opens a question on whether they can be used for training CNN. In this contribution, the main focus of this paper is the problem of processing vectors with a small number of features in comparison to the number of pixels in the output images. The proposed methodology was applied to the classification of PET coincidence events.

Keywords: convolutional neural network, kernel principal component analysis, medical imaging, positron emission tomography

Procedia PDF Downloads 119

Authors: Ying Li, Yan Li

Abstract:

A new algorithm for single hidden layer feedforward neural networks (SLFN), Orthogonal Basis Extreme Learning (OBEL) algorithm, is proposed and the algorithm derivation is given in the paper. The algorithm can decide both the NNs parameters and the neuron number of hidden layer(s) during training while providing extreme fast learning speed. It will provide a practical way to develop NNs. The simulation results of function approximation showed that the algorithm is effective and feasible with good accuracy and adaptability.

Keywords: neural network, orthogonal basis extreme learning, function approximation

Procedia PDF Downloads 519

Authors: Jameela Ali, Hamid A. Jalab, Loay E. George, Abdul Rahim Ahmad, Azizah Suliman, Karim Al-Jashamy

Abstract:

Red blood cells (RBC) are the most common types of blood cells and are the most intensively studied in cell biology. The lack of RBCs is a condition in which the amount of hemoglobin level is lower than normal and is referred to as “anemia”. Abnormalities in RBCs will affect the exchange of oxygen. This paper presents a comparative study for various techniques for classifyig the red blood cells as normal, or abnormal (anemic) using WEKA. WEKA is an open source consists of different machine learning algorithms for data mining applications. The algorithm tested are Radial Basis Function neural network, Support vector machine, and K-Nearest Neighbors algorithm. Two sets of combined features were utilized for classification of blood cells images. The first set, exclusively consist of geometrical features, was used to identify whether the tested blood cell has a spherical shape or non-spherical cells. While the second set, consist mainly of textural features was used to recognize the types of the spherical cells. We have provided an evaluation based on applying these classification methods to our RBCs image dataset which were obtained from Serdang Hospital-Malaysia, and measuring the accuracy of test results. The best achieved classification rates are 97%, 98%, and 79% for Support vector machines, Radial Basis Function neural network, and K-Nearest Neighbors algorithm respectively

Keywords: red blood cells, classification, radial basis function neural networks, suport vector machine, k-nearest neighbors algorithm

Procedia PDF Downloads 466

Authors: Seunghee Park, Junkyeong Kim, Eun-Seok Shin, Sang-Hun Han

Abstract:

In this study, two kinds of nondestructive evaluation (NDE) techniques (rebound hardness and ultrasonic pulse velocity methods) are investigated for the effective maintenance of underwater concrete structures. A new methodology to estimate the underwater concrete strengths more effectively, named “artificial neural network (ANN) – based concrete strength estimation with the combination of rebound hardness and ultrasonic pulse velocity methods” is proposed and verified throughout a series of experimental works.

Keywords: underwater concrete, rebound hardness, Schmidt hammer, ultrasonic pulse velocity, ultrasonic sensor, artificial neural networks, ANN

Procedia PDF Downloads 516

Authors: Cameron Hamilton, Walter Potter, Gerrit Hoogenboom, Ronald McClendon, Will Hobbs

Abstract:

A model was constructed to predict the amount of solar radiation that will make contact with the surface of the earth in a given location an hour into the future. This project was supported by the Southern Company to determine at what specific times during a given day of the year solar panels could be relied upon to produce energy in sufficient quantities. Due to their ability as universal function approximators, an artificial neural network was used to estimate the nonlinear pattern of solar radiation, which utilized measurements of weather conditions collected at the Griffin, Georgia weather station as inputs. A number of network configurations and training strategies were utilized, though a multilayer perceptron with a variety of hidden nodes trained with the resilient propagation algorithm consistently yielded the most accurate predictions. In addition, a modeled DNI field and adjacent weather station data were used to bolster prediction accuracy. In later trials, the solar radiation field was preprocessed with a discrete wavelet transform with the aim of removing noise from the measurements. The current model provides predictions of solar radiation with a mean square error of 0.0042, though ongoing efforts are being made to further improve the model’s accuracy.

Keywords: artificial neural networks, resilient propagation, solar radiation, time series forecasting

Procedia PDF Downloads 370

Authors: Mostafa Sefidgar, Kaamran Raahemifar, Hossein Bazmara, Madjid Soltani

Abstract:

The computational methods provide condition for investigation related to the process of drug delivery, such as convection and diffusion of drug in extracellular matrices, and drug extravasation from microvascular. The information of this process clarifies the mechanisms of drug delivery from the injection site to absorption by a solid tumor. In this study, an advanced numerical method is used to solve fluid flow and solute transport equations simultaneously to show how capillary network structure induced by tumor affects drug delivery. The effect of heterogeneous capillary network induced by tumor on interstitial fluid flow and drug delivery is investigated by this multi scale method. The sprouting angiogenesis model is used for generating capillary network induced by tumor. Fluid flow governing equations are implemented to calculate blood flow through the tumor-induced capillary network and fluid flow in normal and tumor tissues. The Starling’s law is used for closing this system of equations and coupling the intravascular and extravascular flows. Finally, convection-diffusion-reaction equation is used to simulate drug delivery. The dynamic approach which changes the capillary network structure based on signals sent by hemodynamic and metabolic stimuli is used in this study for more realistic assumption. The study indicates that drug delivery to solid tumors depends on the tumor induced capillary network structure. The dynamic approach generates the irregular capillary network around the tumor and predicts a higher interstitial pressure in the tumor region. This elevated interstitial pressure with irregular capillary network leads to a heterogeneous distribution of drug in the tumor region similar to in vivo observations. The investigation indicates that the drug transport properties have a significant role against the physiological barrier of drug delivery to a solid tumor.

Keywords: solid tumor, physiological barriers to drug delivery, angiogenesis, microvascular network, solute transport

Procedia PDF Downloads 295

Authors: Pouya Taraghi, Yong Li, Nader Yoosef-Ghodsi, Muntaseer Kainat, Samer Adeeb

Abstract:

Buried pipelines play a crucial role in the transportation of energy products such as oil, gas, and various chemical fluids, ensuring their efficient and safe distribution. However, these pipelines are often susceptible to ground movements caused by geohazards like landslides, fault movements, lateral spreading, and more. Such ground movements can lead to strain-induced failures in pipes, resulting in leaks or explosions, leading to fires, financial losses, environmental contamination, and even loss of human life. Therefore, it is essential to study how buried pipelines respond when traversing geohazard-prone areas to assess the potential impact of ground movement on pipeline design. As such, this study introduces an approach called the Physics-Informed Neural Network (PINN) to predict the strain demand in inelastic pipes subjected to permanent ground displacement (PGD). This method uses a deep learning framework that does not require training data and makes it feasible to consider more realistic assumptions regarding existing nonlinearities. It leverages the underlying physics described by differential equations to approximate the solution. The study analyzes various scenarios involving different geohazard types, PGD values, and crossing angles, comparing the predictions with results obtained from finite element methods. The findings demonstrate a good agreement between the results of the proposed method and the finite element method, highlighting its potential as a simulation-free, data-free, and meshless alternative. This study paves the way for further advancements, such as the simulation-free reliability assessment of pipes subjected to PGD, as part of ongoing research that leverages the proposed method.

Keywords: strain demand, inelastic pipe, permanent ground displacement, machine learning, physics-informed neural network

Procedia PDF Downloads 47

Authors: Sayed Hadi Sadeghi

Abstract:

This research study examines the current state of support services for e-network practice in an Australian and an American university. It identifies information that will be of assistance to Australian and American universities to improve their existing online programs. The study investigated the two universities using a quantitative methodological approach. Participants were students, lecturers and admins of universities engaged with online courses and learning management systems. The support services for e-network practice variables, namely academic support services, administrative support and technical support, were investigated for e-practice. Evaluations of e-network support service and its sub factors were above average and excellent in both countries, although the American admins and lecturers tended to evaluate this factor higher than others did. Support practice was evaluated higher by all participants of an American university than by Australians. One explanation for the results may be that most suppliers of the Australian university e-learning system were from eastern Asian cultural backgrounds with a western networking support perspective about e-learning.

Keywords: support services, e-Network practice, Australian universities, United States universities

Procedia PDF Downloads 149

Authors: Sujata Kulkarni, Udhav Bhosle, Vijaykumar T.

Abstract:

Due to the closeness between seismic signals and non-seismic signals, it is vital to detect earthquakes using conventional methods. In order to distinguish between seismic events and non-seismic events depending on their amplitude, our study processes the data that come from seismic sensors. The authors suggest a robust noise suppression technique that makes use of a bandpass filter, an IIR Wiener filter, recursive short-term average/long-term average (STA/LTA), and Carl short-term average (STA)/long-term average for event identification (LTA). The trigger ratio used in the proposed study to differentiate between seismic and non-seismic activity is determined. The proposed work focuses on significant feature extraction for machine learning-based seismic event detection. This serves as motivation for compiling a dataset of all features for the identification and forecasting of seismic signals. We place a focus on feature vector dimension reduction techniques due to the temporal complexity. The proposed notable features were experimentally tested using a machine learning model, and the results on unseen data are optimal. Finally, a presentation using a hybrid dataset (captured by different sensors) demonstrates how this model may also be employed in a real-time setting while lowering false alarm rates. The planned study is based on the examination of seismic signals obtained from both individual sensors and sensor networks (SN). A wideband seismic signal from BSVK and CUKG station sensors, respectively located near Basavakalyan, Karnataka, and the Central University of Karnataka, makes up the experimental dataset.

Keywords: Carl STA/LTA, features extraction, real time, dataset, machine learning, seismic detection

Procedia PDF Downloads 100

Authors: Elif D. Oguz Erkal, David Krackhardt, Erica Cochran-Hameen

Abstract:

Even though social network analysis (SNA) is an abundantly studied concept for many organizations and industries, a clear SNA approach to the project teams has not yet been adopted by the construction industry. The main challenges for performing SNA in construction and the apparent reason for this gap is the unique and complex structure of each construction project, the comparatively high circulation of project team members/contributing parties and the variety of authentic problems for each project. Additionally, there are stakeholders from a variety of professional backgrounds collaborating in a high-stress environment fueled by time and cost constraints. Within this case study on Project RE, a design & build project performed at the Urban Design Build Studio of Carnegie Mellon University, social network analysis of the project design team will be performed with the main goal of applying social network theory to construction project environments. The research objective is to determine a correlation between the network of how individuals relate to each other on one’s perception of their own professional strengths and weaknesses and the communication patterns within the team and the group dynamics. Data is collected through a survey performed over four rounds conducted monthly, detailed follow-up interviews and constant observations to assess the natural alteration in the network with the effect of time. The data collected is processed by the means of network analytics and in the light of the qualitative data collected with observations and individual interviews. This paper presents the full ethnography of this construction design team of fourteen architecture students based on an elaborate social network data analysis over time. This study is expected to be used as an initial step to perform a refined, targeted and large-scale social network data collection in construction projects in order to deduce the impacts of social networks on project performance and suggest better collaboration structures for construction project teams henceforth.

Keywords: construction design teams, construction project management, social network analysis, team collaboration, network analytics

Procedia PDF Downloads 185

Authors: B. R. Campomanes-Alvarez, P. Quiros, B. Fernandez

Abstract:

Automatic Speech Recognition (ASR) is a machine-based process of decoding and transcribing oral speech. A typical ASR system receives acoustic input from a speaker or an audio file, analyzes it using algorithms, and produces an output in the form of a text. Some speech recognition systems use Hidden Markov Models (HMMs) to deal with the temporal variability of speech and Gaussian Mixture Models (GMMs) to determine how well each state of each HMM fits a short window of frames of coefficients that represents the acoustic input. Another way to evaluate the fit is to use a feed-forward neural network that takes several frames of coefficients as input and produces posterior probabilities over HMM states as output. Deep neural networks (DNNs) that have many hidden layers and are trained using new methods have been shown to outperform GMMs on a variety of speech recognition systems. Acoustic models for state-of-the-art ASR systems are usually training on massive amounts of data. However, audio files with their corresponding transcriptions can be difficult to obtain, especially in the Spanish language. Hence, in the case of these low-resource scenarios, building an ASR model is considered as a complex task due to the lack of labeled data, resulting in an under-trained system. Semi-supervised learning approaches arise as necessary tasks given the high cost of transcribing audio data. The main goal of this proposal is to develop a procedure based on acoustic semi-supervised learning for Spanish ASR systems by using DNNs. This semi-supervised learning approach consists of: (a) Training a seed ASR model with a DNN using a set of audios and their respective transcriptions. A DNN with a one-hidden-layer network was initialized; increasing the number of hidden layers in training, to a five. A refinement, which consisted of the weight matrix plus bias term and a Stochastic Gradient Descent (SGD) training were also performed. The objective function was the cross-entropy criterion. (b) Decoding/testing a set of unlabeled data with the obtained seed model. (c) Selecting a suitable subset of the validated data to retrain the seed model, thereby improving its performance on the target test set. To choose the most precise transcriptions, three confidence scores or metrics, regarding the lattice concept (based on the graph cost, the acoustic cost and a combination of both), was performed as selection technique. The performance of the ASR system will be calculated by means of the Word Error Rate (WER). The test dataset was renewed in order to extract the new transcriptions added to the training dataset. Some experiments were carried out in order to select the best ASR results. A comparison between a GMM-based model without retraining and the DNN proposed system was also made under the same conditions. Results showed that the semi-supervised ASR-model based on DNNs outperformed the GMM-model, in terms of WER, in all tested cases. The best result obtained an improvement of 6% relative WER. Hence, these promising results suggest that the proposed technique could be suitable for building ASR models in low-resource environments.

Keywords: automatic speech recognition, deep neural networks, machine learning, semi-supervised learning

Procedia PDF Downloads 325

Authors: Taehan Lee

Abstract:

We consider a network design problem which has shortest routing restriction based on the values determined by the installed facilities on each arc. In conventional multicommodity network design problem, a commodity can be routed through any possible path when the capacity is available. But, we consider a problem in which the commodity between two nodes must be routed on a path which has shortest metric value and the link metric value is determined by the installed facilities on the link. By this routing restriction, the problem has a distinct characteristic. We present an integer programming formulation containing the primal-dual optimality conditions to the shortest path routing. We give some computational results for the model.

Keywords: integer programming, multicommodity network design, routing, shortest path

Procedia PDF Downloads 408

Authors: Gabriela Tutuanu

Abstract:

The economic, social and environmental undesirable impact of the existing development pattern pushes to the adoption and use of a new development paradigm that of sustainable development. This paper intends to substantiate how the marketing can help the sustainable development. It begins with the subjects of sustainable development and sustainable marketing as they are discussed in literature. The sustainable development is a three dimensional concept which embeds the economic dimension, the social dimension and the environmental dimension that ask to have in view the simultaneous pursuit of economic prosperity, social equity and environmental quality. A major challenge to achieve these goals at business level and to integrate all three dimensions of sustainability is the sustainable marketing. The sustainable marketing is a relationship marketing that aims at building lasting relationships with the social and natural environment on a long-term thinking and futurity and this philosophy allows helping all three dimensions of sustainability. As marketing solutions that could contribute to the sustainable development. We advance the stimulation of sustainable demand, the constant innovation and improvement of sustainable products, the design and use of customized communication, a multichannel distribution network and the sale of sustainable products and services at fair prices. Their implementation will increase the economic, social and environmental sustainability at a large extent in the future if they are supported by political, governmental and legal authorities.

Keywords: sustainable development, sustainable marketing, sustainable demand, sustainable product, credible communication, multi-channel distribution network, fair price

Procedia PDF Downloads 461

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 63

Authors: Mst Shapna Akter, Hossain Shahriar

Abstract:

One of the most important challenges in the field of software code audit is the presence of vulnerabilities in software source code. Every year, more and more software flaws are found, either internally in proprietary code or revealed publicly. These flaws are highly likely exploited and lead to system compromise, data leakage, or denial of service. C and C++ open-source code are now available in order to create a largescale, machine-learning system for function-level vulnerability identification. We assembled a sizable dataset of millions of opensource functions that point to potential exploits. We developed an efficient and scalable vulnerability detection method based on deep neural network models that learn features extracted from the source codes. The source code is first converted into a minimal intermediate representation to remove the pointless components and shorten the dependency. Moreover, we keep the semantic and syntactic information using state-of-the-art word embedding algorithms such as glove and fastText. The embedded vectors are subsequently fed into deep learning networks such as LSTM, BilSTM, LSTM-Autoencoder, word2vec, BERT, and GPT-2 to classify the possible vulnerabilities. Furthermore, we proposed a neural network model which can overcome issues associated with traditional neural networks. Evaluation metrics such as f1 score, precision, recall, accuracy, and total execution time have been used to measure the performance. We made a comparative analysis between results derived from features containing a minimal text representation and semantic and syntactic information. We found that all of the deep learning models provide comparatively higher accuracy when we use semantic and syntactic information as the features but require higher execution time as the word embedding the algorithm puts on a bit of complexity to the overall system.

Keywords: cyber security, vulnerability detection, neural networks, feature extraction

Procedia PDF Downloads 71

Authors: Muhammad Arif Khan, Haroon Idrees, Imran Aziz, Sidra Mushtaq

Abstract:

Social platforms like online discussion and mailing groups are well aligned with academic as well as professional learning spaces. Professional communities are increasingly moving to online forums for sharing and capturing the intellectual abilities. This study investigated dynamics of social connectivity of yahoo mailing groups of Pakistani Library and Information Science (LIS) professionals using Graph Theory technique. Design/Methodology: Social Network Analysis is the increasingly concerned domain for scientists in identifying whether people grow together through online social interaction or, whether they just reflect connectivity. We have conducted a longitudinal study using Network Graph Theory technique to analyze the large data-set of email communication. The data was collected from three yahoo mailing groups using network analysis software over a period of six months i.e. January to June 2016. Findings of the network analysis were reviewed through focus group discussion with LIS experts and selected respondents of the study. Data were analyzed in Microsoft Excel and network diagrams were visualized using NodeXL and ORA-Net Scene package. Findings: Findings demonstrate that professionals and students exhibit intellectual growth the more they get tied within a network by interacting and participating in communication through online forums. The study reports on dynamics of the large network by visualizing the email correspondence among group members in a network consisting vertices (members) and edges (randomized correspondence). The model pair wise relationship between group members was illustrated to show characteristics, reasons, and strength of ties. Connectivity of nodes illustrated the frequency of communication among group members through examining node coupling, diffusion of networks, and node clustering has been demonstrated in-depth. Network analysis was found to be a useful technique in investigating the dynamics of the large network.

Keywords: emailing networks, network graph theory, online social platforms, yahoo mailing groups

Procedia PDF Downloads 223

Authors: M. Caramihai, Irina Severin

Abstract:

This paper deals with the design of an intelligent control structure for a bioprocess of Hansenula polymorpha yeast cultivation. The objective of the process control is to produce biomass in a desired physiological state. The work demonstrates that the designed Hybrid Control Techniques (HCT) are able to recognize specific evolution bioprocess trajectories using neural networks trained specifically for this purpose, in order to estimate the model parameters and to adjust the overall bioprocess evolution through an expert system and a fuzzy structure. The design of the control algorithm as well as its tuning through realistic simulations is presented. Taking into consideration the synergism of different paradigms like fuzzy logic, neural network, and symbolic artificial intelligence (AI), in this paper we present a real and fulfilled intelligent control architecture with application in bioprocess control.

Keywords: bioprocess, intelligent control, neural nets, fuzzy structure, hybrid techniques

Procedia PDF Downloads 398

Authors: Christian Kapuku, Seung-Young Kho

Abstract:

An accurate representation of the transportation system serving the region is one of the important aspects of transportation modeling. Such representation often requires developing an abstract model of the system elements, which also requires important amount of data, surveys and time. However, in some cases such as in developing countries, data deficiencies, time and budget constraints do not always allow such accurate representation, leaving opportunities to assumptions that may negatively affect the quality of the analysis. With the emergence of Internet open source data especially in the mapping technologies as well as the advances in Geography Information System, opportunities to tackle these issues have raised. Therefore, the objective of this paper is to demonstrate such application through a practical case of the development of the transportation network for the city of Kinshasa. The GIS geo-referencing was used to construct the digitized map of Transportation Analysis Zones using available scanned images. Centroids were then dynamically placed at the center of activities using an activities density map. Next, the road network with its characteristics was built using OpenStreet data and other official road inventory data by intersecting their layers and cleaning up unnecessary links such as residential streets. The accuracy of the final network was then checked, comparing it with satellite images from Google and Bing. For the validation, the final network was exported into Emme3 to check for potential network coding issues. Results show a high accuracy between the built network and satellite images, which can mostly be attributed to the use of open source data.

Keywords: geographic information system (GIS), network construction, transportation database, open source data

Procedia PDF Downloads 153

Authors: Z. Nougrara, J. Meunier

Abstract:

In this paper, we present a new approach for extracting roads, there road network and its nodes from satellite image representing regions in Algeria. Our approach is related to our previous research work. It is founded on the information theory and the mathematical morphology. We therefore have to define objects as sets of pixels and to study the shape of these objects and the relations that exist between them. The main interest of this study is to solve the problem of the automatic mapping from satellite images. This study is thus applied for that the geographical representation of the images is as near as possible to the reality.

Keywords: nodes, road network, satellite image, updating a road map

Procedia PDF Downloads 405

Authors: Chan Hui-Ya, Ding Shin-Tan

Abstract:

Background and Aim: The discharge planning is considered helpful to reduce the hospital length of stay and readmission rate, and then increased satisfaction with healthcare for patients and professionals. In order to decrease the waiting time of long-term care and boost the care quality of patients after discharge from the hospital, the Ministry of Health and Welfare department in Taiwan initiates a program “discharge planning connects with long-term care 2.0 services” in 2017. The purpose of this study is to investigate the outcome of the pilot of this program in a medical center. Methods: By purpose sampling, the study chose five wards in a medical center as pilot units. The researchers compared the beds of service, the numbers of cases which were transferred to the long-term care center and transferred rates per month between the pilot units and the other units, and analyze the basic data, the long-term care service needs and the approval service items of cases transfer to the long-term care center in pilot units. Results: From June to September 2017, a total of 92 referrals were made, and 51 patients were enrolled into the pilot program. There is a significant difference of transferring rate between the pilot units and the other units (χ = 702.6683, p < 0.001). Only 20 cases (39.2% success rate) were approved to accept the parts of service items of long-term care in the pilot units. The most approval item was respite care service (n = 13; 65%), while it was third at needs ranking of service lists during linking services process. Among the reasons of patients who cancelled the request, 38.71% reasons were related to the services which could not match the patients’ needs and expectation. Conclusion: The results indicate there is a requirement to modify the long-term care services to fit the needs of cases. The researchers suggest estimating the potential cases by screening data from hospital informatics systems and to hire more case manager according the service time of potential cases. Meanwhile, the strategies shortened the assessment scale and authorized hospital case managers to approve some items of long-term care should be considered.

Keywords: discharge planning, long-term care, case manager, patient care

Procedia PDF Downloads 271

Authors: Anoop T. R., Otman Basir, Robert F. Hess, Eileen E. Birch, Brooke A. Koritala, Reed M. Jost, Becky Luu, David Stager, Ben Thompson

Abstract:

Strabismus refers to a misalignment of the eyes. Early detection and treatment of strabismus in childhood can prevent the development of permanent vision loss due to abnormal development of visual brain areas. We developed a two-stage method for strabismus detection and classification based on photographs of the face. The first stage detects the presence or absence of strabismus, and the second stage classifies the type of strabismus. The first stage comprises face detection using Haar cascade, facial landmark estimation, face alignment, aligned face landmark detection, segmentation of the eye region, and detection of strabismus using VGG 16 convolution neural networks. Face alignment transforms the face to a canonical pose to ensure consistency in subsequent analysis. Using facial landmarks, the eye region is segmented from the aligned face and fed into a VGG 16 CNN model, which has been trained to classify strabismus. The CNN determines whether strabismus is present and classifies the type of strabismus (exotropia, esotropia, and vertical deviation). If stage 1 detects strabismus, the eye region image is fed into stage 2, which starts with the estimation of pupil center coordinates using mask R-CNN deep neural networks. Then, the distance between the pupil coordinates and eye landmarks is calculated along with the angle that the pupil coordinates make with the horizontal and vertical axis. The distance and angle information is used to characterize the degree and direction of the strabismic eye misalignment. This model was tested on 100 clinically labeled images of children with (n = 50) and without (n = 50) strabismus. The True Positive Rate (TPR) and False Positive Rate (FPR) of the first stage were 94% and 6% respectively. The classification stage has produced a TPR of 94.73%, 94.44%, and 100% for esotropia, exotropia, and vertical deviations, respectively. This method also had an FPR of 5.26%, 5.55%, and 0% for esotropia, exotropia, and vertical deviation, respectively. The addition of one more feature related to the location of corneal light reflections may reduce the FPR, which was primarily due to children with pseudo-strabismus (the appearance of strabismus due to a wide nasal bridge or skin folds on the nasal side of the eyes).

Keywords: strabismus, deep neural networks, face detection, facial landmarks, face alignment, segmentation, VGG 16, mask R-CNN, pupil coordinates, angle deviation, horizontal and vertical deviation

Procedia PDF Downloads 64

Authors: M. Olmedilla, M. R. Martinez-Torres, S. L. Toral

Abstract:

Electronic Word-Of-Mouth (eWOM) communities represent today an important source of information in which more and more customers base their purchasing decisions. They include thousands of reviews concerning very different products and services posted by many individuals geographically distributed all over the world. Due to their massive audience, eWOM communities can help users to find the product they are looking for even if they are less popular or rare. This is known as the long tail effect, which leads to a larger number of lower-selling niche products. This paper analyzes the long tail effect in a well-known eWOM community and defines a tool for finding niche products unavailable through conventional channels.

Keywords: eWOM, online user reviews, long tail theory, product categorization, social network analysis

Procedia PDF Downloads 404