Search results for: multilayer network

Authors: A. Dinelka Panagoda, Bathiya Bandara, Chamod Wijetunga, Chathura Malinda, Lakmal Rupasinghe, Chethana Liyanapathirana

Abstract:

This research takes us forward with the concepts of Federated Learning and Software-Defined Networking (SDN) to introduce an efficient malware detection technique and provide a mitigation mechanism to give birth to a resilient and automated healthcare sector network system by also adding the feature of extended privacy preservation. Due to the daily transformation of new malware attacks on hospital Integrated Clinical Environment (ICEs), the healthcare industry is at an undefinable peak of never knowing its continuity direction. The state of blindness by the array of indispensable opportunities that new medical device inventions and their connected coordination offer daily, a factor that should be focused driven is not yet entirely understood by most healthcare operators and patients. This solution has the involvement of four clients in the form of hospital networks to build up the federated learning experimentation architectural structure with different geographical participation to reach the most reasonable accuracy rate with privacy preservation. While the logistic regression with cross-entropy conveys the detection, SDN comes in handy in the second half of the research to stack up the initial development phases of the system with malware mitigation based on policy implementation. The overall evaluation sums up with a system that proves the accuracy with the added privacy. It is no longer needed to continue with traditional centralized systems that offer almost everything but not privacy.

Keywords: software-defined network, federated learning, privacy, integrated clinical environment, decentralized learning, malware detection, malware mitigation

Procedia PDF Downloads 163

Authors: Rafat Rob, Khaled Alotaibi, Dana Nour, Abdullah Albadrani, Abdulmohsen Mulhim

Abstract:

Power systems digitization solutions provides a comprehensive smart, cohesive, interconnected network, extensive connectivity between digital assets, physical power plants, and resources to form digital economies. However, digitization has exposed the classical air gapped power plants to the rapid spread of cyber threats and attacks in the process delaying and forcing many organizations to rethink their cyber security policies and standards before they can augment their operation the new advanced digital devices. Cyber Security requirements for power systems (and industry control systems therein) demand a new approach, unique methodology, and design process that is completely different to Cyber Security measures designed for the IT systems. In practice, Cyber Security strategy, as applied to power systems, tends to be closely aligned to those measures applied for IT system purposes. The differentiator for Cyber Security in terms of power systems are the physical assets and applications used, alongside the ever-growing rate of expansion within the industry controls sector (in comparison to the relatively saturated growth observed for corporate IT systems). These factors increase the magnitude of the cyber security risk within such systems. The introduction of smart devices and sensors along the grid initiate vulnerable entry points to the systems. Every installed Smart Meter is a target; the way these devices communicate with each other may instigate a Denial of Service (DoS) and Distributed Denial of Service (DDoS) attack. Attacking one sensor or meter has the potential to propagate itself throughout the power grid reaching the IT network, where it may manifest itself as a malware infiltration.

Keywords: supply chain, cybersecurity, maturity model, risk, smart grid

Procedia PDF Downloads 93

Authors: Nibal Al-Batsh, Issam A. Al-Khatib, Subha Ghannam

Abstract:

Yatta is a town in the Governorate of Hebron, located 9 km south of Hebron City in the West Bank. The town houses over 100,000 people, 49% of which are females; a population that doubles every 15 years. Yatta has been connected to a water network since 1974 serving nearly 85% of the households. The water network is old and inadequate to meet the needs of the population. The water supply made available to the area is also very limited, estimated to be around 20 l/c/d. Residents are thus forced to rely on water vendors which supply water with a lower quality compared to municipal water while being 400% more expensive. As a cheaper and more reliable alternative, rainwater harvesting is a common practice in the area, with the majority of the households owning at least one cistern. Rainwater harvesting is of great socioeconomic importance in areas where water sources are scarce or polluted. In this research, the quality of harvested rainwater used for drinking and domestic purposes in the Yatta area was assessed throughout a year. A total of 100 samples, were collected from (cisterns) with an average capacity of 69 m3, which are adjacent to cement-roof catchment areas with an average area of 145 m2. Samples were analyzed for a number of parameters including: pH, alkalinity, hardness, turbidity, Total Dissolved Solids (TDS), NO3, NH4, chloride and salinity. Biological and microbiological contents such as Total Coliforms (TCC) and Fecal Coliforms (FC) bacteria were also tested. Results showed that most of the rainwater samples were within WHO and EPA guidelines set for chemical parameters. The research also addressed the impact of different socioeconomic attributes on rainwater harvesting through questionnaire that was pre-tested before the actual statically sample is collected.

Keywords: rainwater, harvesting, water quality, socio-economic aspects

Procedia PDF Downloads 241

Authors: Ali Pourkazemi

Abstract:

The transient radar method (TRM) is one of the non-destructive methods that was introduced by authors a few years ago. The transient radar method can be classified as a wave-based non destructive testing (NDT) method that can be used in a wide frequency range. Nevertheless, it requires a narrow band, ranging from a few GHz to a few THz, depending on the application. As a time-of-flight and real-time method, TRM can measure the electromagnetic properties of the sample under test not only quickly and accurately, but also blindly. This means that it requires no prior knowledge of the sample under test. For multi-layer structures, TRM is not only able to detect changes related to any parameter within the multi-layer structure but can also measure the electromagnetic properties of each layer and its thickness individually. Although the temperature, humidity, and general environmental conditions may affect the sample under test, they do not affect the accuracy of the Blind TRM algorithm. In this paper, the electromagnetic properties as well as the thickness of the individual building insulation materials - as a single-layer structure - are measured experimentally. Finally, the correlation between the reflection coefficients and some other technical parameters such as sound insulation, thermal resistance, thermal conductivity, compressive strength, and density is investigated. The sample to be studied is 30 cm x 50 cm and the thickness of the samples varies from a few millimeters to 6 centimeters. This experiment is performed with both biostatic and differential hardware at 10 GHz. Since it is a narrow-band system, high-speed computation for analysis, free-space application, and real-time sensor, it has a wide range of potential applications, e.g., in the construction industry, rubber industry, piping industry, wind energy industry, automotive industry, biotechnology, food industry, pharmaceuticals, etc. Detection of metallic, plastic pipes wires, etc. through or behind the walls are specific applications for the construction industry.

Keywords: transient radar method, blind electromagnetic geometrical parameter extraction technique, ultrafast nondestructive multilayer dielectric structure characterization, electronic measurement systems, illumination, data acquisition performance, submillimeter depth resolution, time-dependent reflected electromagnetic signal blind analysis method, EM signal blind analysis method, time domain reflectometer, microwave, milimeter wave frequencies

Procedia PDF Downloads 61

Authors: Xiaolei Hu, Enrico Ferrera, Riccardo Tomasi, Claudio Pastrone

Abstract:

Load Forecasting plays a key role in making today's and future's Smart Energy Grids sustainable and reliable. Accurate power consumption prediction allows utilities to organize in advance their resources or to execute Demand Response strategies more effectively, which enables several features such as higher sustainability, better quality of service, and affordable electricity tariffs. It is easy yet effective to apply Load Forecasting at larger geographic scale, i.e. Smart Micro Grids, wherein the lower available grid flexibility makes accurate prediction more critical in Demand Response applications. This paper analyses the application of short-term load forecasting in a concrete scenario, proposed within the EU-funded GreenCom project, which collect load data from single loads and households belonging to a Smart Micro Grid. Three short-term load forecasting techniques, i.e. linear regression, artificial neural networks, and radial basis function network, are considered, compared, and evaluated through absolute forecast errors and training time. The influence of weather conditions in Load Forecasting is also evaluated. A new definition of Gain is introduced in this paper, which innovatively serves as an indicator of short-term prediction capabilities of time spam consistency. Two models, 24- and 1-hour-ahead forecasting, are built to comprehensively compare these three techniques.

Keywords: short-term load forecasting, smart micro grid, linear regression, artificial neural networks, radial basis function network, gain

Procedia PDF Downloads 452

Authors: Nur Bella Octoria Bella, Ayomi Dita Rarasati

Abstract:

DKI Jakarta is Indonesia's capital city with the 10th highest congestion rate in the world based on the 2019 traffic index. Other than that based on World Air Quality Report in 2019 showed DKI Jakarta's air pollutant concentrate 49.4 µg and the 5th highest air pollutant in the world. In the urban city nowadays, the mobility rate is high enough and the efficiency for sustainability assessment in transport infrastructure development is needed. This efficiency is the important key for sustainable infrastructure development. DKI Jakarta is nowadays in the process of constructing the railway infrastructure to support the transportation system. The problems appearing are the railway infrastructure networks and the service in DKI Jakarta already planned based on sustainability factors or not. Therefore, the aim of this research is to make the evaluation of railways infrastructure networks performance and services in DKI Jakarta regards on the railway sustainability key factors. Further, this evaluation will be used to make the railway sustainability assessment framework and to offer some of the alternative solutions to improve railway transportation sustainability in DKI Jakarta. Firstly a very detailed literature review of papers that have focused on railway sustainability factors and their improvements of railway sustainability, published in the scientific journal in the period 2011 until 2021. Regarding the sustainability factors from the literature review, further, it is used to assess the current condition of railway infrastructure in DKI Jakarta. The evaluation will be using a Likert rate questionnaire and directed to the transportation railway expert and the passenger. Furthermore, the mapping and evaluation rate based on the sustainability factors will be compared to the effect factors using the Analytical Hierarchical Process (AHP). This research offers the network's performance and service rate impact on the sustainability aspect and the passenger willingness for using the rail public transportation in DKI Jakarta.

Keywords: transportation sustainability, railway transportation, sustainability, DKI Jakarta

Procedia PDF Downloads 150

Authors: Yu-Jia Jian, Emily Chia-Yu Su, Hui-Ling Hsu, Jian-Jhih Chen

Abstract:

Coronary artery is the major supplier of myocardial blood flow. When fat and cholesterol are deposit in the coronary arterial wall, narrowing and stenosis of the artery occurs, which may lead to myocardial ischemia and eventually infarction. According to the World Health Organization (WHO), estimated 740 million people have died of coronary heart disease in 2015. According to Statistics from Ministry of Health and Welfare in Taiwan, heart disease (except for hypertensive diseases) ranked the second among the top 10 causes of death from 2013 to 2016, and it still shows a growing trend. According to American Heart Association (AHA), the risk factors for coronary heart disease including: age (> 65 years), sex (men to women with 2:1 ratio), obesity, diabetes, hypertension, hyperlipidemia, smoking, family history, lack of exercise and more. We have collected a dataset of 421 patients from a hospital located in northern Taiwan who received coronary computed tomography (CT) angiography. There were 300 males (71.26%) and 121 females (28.74%), with age ranging from 24 to 92 years, and a mean age of 56.3 years. Prior to coronary CT angiography, basic data of the patients, including age, gender, obesity index (BMI), diastolic blood pressure, systolic blood pressure, diabetes, hypertension, hyperlipidemia, smoking, family history of coronary heart disease and exercise habits, were collected and used as input variables. The output variable of the prediction module is the degree of coronary artery stenosis. The output variable of the prediction module is the narrow constriction of the coronary artery. In this study, the dataset was randomly divided into 80% as training set and 20% as test set. Four machine learning algorithms, including logistic regression, stepwise regression, neural network and decision tree, were incorporated to generate prediction results. We used area under curve (AUC) / accuracy (Acc.) to compare the four models, the best model is neural network, followed by stepwise logistic regression, decision tree, and logistic regression, with 0.68 / 79 %, 0.68 / 74%, 0.65 / 78%, and 0.65 / 74%, respectively. Sensitivity of neural network was 27.3%, specificity was 90.8%, stepwise Logistic regression sensitivity was 18.2%, specificity was 92.3%, decision tree sensitivity was 13.6%, specificity was 100%, logistic regression sensitivity was 27.3%, specificity 89.2%. From the result of this study, we hope to improve the accuracy by improving the module parameters or other methods in the future and we hope to solve the problem of low sensitivity by adjusting the imbalanced proportion of positive and negative data.

Keywords: decision support, computed tomography, coronary artery, machine learning

Procedia PDF Downloads 217

Authors: S. Łęgowik-Świącik

Abstract:

This article is a contribution to the research into the determinants of value creation via digital platforms in variable operating conditions. The dynamics of the market environment caused by the COVID-19 pandemic have made enterprises built on digital platforms financially successful. While many classic companies are struggling with the uncertainty of conducting a business and difficulties in the process of value creation, digital platforms create value by modifying the existing business model to meet the changing needs of customers. Therefore, the objective of this publication is to understand and explain the relationship between value creation and the conversion of the business model built on digital platforms under pandemic conditions. The considerations relating to the conceptual framework and determining the research objective allowed for adopting the hypothesis, assuming that the processes of value creation are evolving, and the measurement of these processes allows for the protection of value created and enables its growth in changing circumstances. The research methods, such as critical literature analysis and case study, were applied to accomplish the objective pursued and verify the hypothesis formulated. The empirical research was carried out based on the data from enterprises listed on the Nasdaq Stock Exchange: Amazon, Alibaba, and Facebook. The research period was the years 2018-2021. The surveyed enterprises were chosen based on the targeted selection. The problem discussed is important and current since the lack of in-depth theoretical research results in few attempts to identify the determinants of value creation via digital platforms. The above arguments led to an attempt at theoretical analysis and empirical research to fill in the gap perceived by deepening the understanding of the process of value creation through network effects via digital platforms under pandemic conditions.

Keywords: business model, digital platforms, enterprise management, pandemic conditions, value creation process

Procedia PDF Downloads 113

Authors: Ryan Chalk, G. M. Shafiullah

Abstract:

Climate change is a pertinent issue facing governments and societies around the world. The industrial revolution has resulted in a steady increase in the average global temperature. The mining and energy production industries have been significant contributors to this change prompting government to intervene by promoting low emission technology within these sectors. This paper initially reviews the energy problem in Australia and the mining sector with a focus on the energy requirements and production methods utilised in Western Australia (WA). Renewable energy in the form of utility-scale solar photovoltaics (PV) provides a solution to these problems by providing emission-free energy which can be used to supplement the existing natural gas turbines in operation at the proposed site. This research presents a custom renewable solution for the mining site considering the specific township network, local weather conditions, and seasonal load profiles. A summary of the required PV output is presented to supply slightly over 50% of the towns power requirements during the peak (summer) period, resulting in close to full coverage in the trench (winter) period. Dig Silent Power Factory Software has been used to simulate the characteristics of the existing infrastructure and produces results of integrating PV. Large scale PV penetration in the network introduce technical challenges, that includes; voltage deviation, increased harmonic distortion, increased available fault current and power factor. Results also show that cloud cover has a dramatic and unpredictable effect on the output of a PV system. The preliminary analyses conclude that mitigation strategies are needed to overcome voltage deviations, unacceptable levels of harmonics, excessive fault current and low power factor. Mitigation strategies are proposed to control these issues predominantly through the use of high quality, made for purpose inverters. Results show that use of inverters with harmonic filtering reduces the level of harmonic injections to an acceptable level according to Australian standards. Furthermore, the configuration of inverters to supply active and reactive power assist in mitigating low power factor problems. Use of FACTS devices; SVC and STATCOM also reduces the harmonics and improve the power factor of the network, and finally, energy storage helps to smooth the power supply.

Keywords: climate change, mitigation strategies, photovoltaic (PV), power quality

Procedia PDF Downloads 155

Authors: Akber Oumer Abdurezak

Abstract:

Road traffic accident is a major problem in Ethiopia, resulting in the deaths of many people and potential injuries and crash every year and loss of properties. According to the Federal Transport Authority, one of the main causes of traffic accident and crash in Ethiopia is over speeding. Implementation of different technologies is used to monitor the speed of vehicles in order to minimize accidents and crashes. This research aimed at designing a speed monitoring system to monitor the speed of travelling vehicles and movements, reporting illegal speeds or overspeeding vehicles to the concerned bodies. The implementation of the system is through a wireless sensor network. The proposed system can sense and detect the movement of vehicles, process, and analysis the data obtained from the sensor and the cloud system. The data is sent to the central controlling server. The system contains accelerometer and gyroscope sensors to sense and collect the data of the vehicle. Arduino to process the data and Global System for Mobile Communication (GSM) module for communication purposes to send the data to the concerned body. When the speed of the vehicle exceeds the allowable speed limit, the system sends a message to database as “over speeding”. Both accelerometer and gyroscope sensors are used to collect acceleration data. The acceleration data then convert to speed, and the corresponding speed is checked with the speed limit, and those above the speed limit are reported to the concerned authorities to avoid frequent accidents. The proposed system decreases the occurrence of accidents and crashes due to overspeeding and can be used as an eye opener for the implementation of other intelligent transport system technologies. This system can also integrate with other technologies like GPS and Google Maps to obtain better output.

Keywords: accelerometer, IOT, GSM, gyroscope

Procedia PDF Downloads 60

Authors: Arfanara Najnin, Michael W. Roach, Jr., Dr. Jianhong Cecilia Xia

Abstract:

The SkyCar Rapid Transit System (SRT) is an innovative intelligent transport system for the sustainable urban transport system. This system will increase the urban area network connectivity and decrease urban area traffic congestion. The SRT system is designed as a suspended Personal Rapid Transit (PRT) system that travels under a guideway 5m above the ground. A driver-less passenger is via pod-cars that hang from slender beams supported by columns that replace existing lamp posts. The beams are setup in a series of interconnecting loops providing non-stop travel from beginning to end to assure journey time. The SRT forward movement is effected by magnetic motors built into the guideway. Passenger stops are at either at line level 5m above the ground or ground level via a spur guideway that curves off the main thoroughfare. The main objective of this paper is to propose an integrated Automated Transit Network (ATN) technology for the future intelligent transport system in the urban built environment. To fulfil the objective a 4D simulated model in the urban built environment has been proposed by using the concept of SRT-ATN system. The methodology for the design, construction and testing parameters of a Technology Demonstrator (TD) for proof of concept and a Simulator (S) has been demonstrated. The completed TD and S will provide an excellent proving ground for the next development stage, the SRT Prototype (PT) and Pilot System (PS). This paper covered by a 4D simulated model in the virtual built environment is to effectively show how the SRT-ATN system works. OpenSim software has been used to develop the model in a virtual environment, and the scenario has been simulated to understand and visualize the proposed SkyCar Rapid Transit Network model. The SkyCar system will be fabricated in a modular form which is easily transported. The system would be installed in increasingly congested city centers throughout the world, as well as in airports, tourist resorts, race tracks and other special purpose for the urban community. This paper shares the lessons learnt from the proposed innovation and provides recommendations on how to improve the future transport system in urban built environment. Safety and security of passengers are prime factors to be considered for this transit system. Design requirements to meet the safety needs to be part of the research and development phase of the project. Operational safety aspects would also be developed during this period. The vehicles, the track and beam systems and stations are the main components that need to be examined in detail for safety and security of patrons. Measures will also be required to protect columns adjoining intersections from errant vehicles in vehicular traffic collisions. The SkyCar Rapid Transit takes advantage of all current disruptive technologies; batteries, sensors and 4G/5G communication and solar energy technologies which will continue to reduce the costs and make the systems more profitable. SkyCar's energy consumption is extremely low compared to other transport systems.

Keywords: SkyCar, rapid transit, Intelligent Transport System (ITS), Automated Transit Network (ATN), urban built environment, 4D Visualization, smart city

Procedia PDF Downloads 202

Authors: Samah Jawad Jabari

Abstract:

Water is the most important and valuable resource not only for human life but also for all living things on the planet. The water supply utilities should fulfill the water requirement quantitatively and qualitatively. Drinking water systems are exposed to both natural (hurricanes and flood) and manmade hazards (risks) that are common in Palestine. Non-Revenue Water (NRW) is a manmade risk which remains a major concern in Palestine, as the NRW levels are estimated to be at a high level. In this research, Hebron city water distribution network was taken as a case study to estimate and audit the NRW levels. The research also investigated the state of the existing water distribution system in the study area by investigating the water losses and obtained more information on NRW prevention and management practices. Data and information have been collected from the Palestinian Water Authority (PWA) and Hebron Municipality (HM) archive. In addition to that, a questionnaire has been designed and administered by the researcher in order to collect the necessary data for water auditing. The questionnaire also assessed the views of stakeholder in PWA and HM (staff) on the current status of the NRW in the Hebron water distribution system. The important result obtained by this research shows that NRW in Hebron city was high and in excess of 30%. The main factors that contribute to NRW were the inaccuracies in billing volumes, unauthorized consumption, and the method of estimating consumptions through faulty meters. Policy for NRW reduction is available in Palestine; however, it is clear that the number of qualified staff available to carry out the activities related to leak detection is low, and that there is a lack of appropriate technologies to reduce water losses and undertake sufficient system maintenance, which needs to be improved to enhance the performance of the network and decrease the level of NRW losses.

Keywords: non-revenue water, water auditing, leak detection, water meters

Procedia PDF Downloads 279

Authors: Abdel-Monem Sayed Mohamed

Abstract:

Lake Aswan is one of the largest man-made reservoirs in the world. The reservoir began to fill in 1964 and the level rose gradually, with annual irrigation cycles, until it reached a maximum water level of 181.5 m in November 1999, with a capacity of 160 km3. The filling of such large reservoir changes the stress system either through increasing vertical compressional stress by loading and/or increased pore pressure through the decrease of the effective normal stress. The resulted effect on fault zones changes stability depending strongly on the orientation of pre-existing stress and geometry of the reservoir/fault system. The main earthquake occurred on November 14, 1981, with magnitude 5.5. This event occurred after 17 years of the reservoir began to fill, along the active part of the Kalabsha fault and located not far from the High Dam. Numerous of small earthquakes follow this earthquake and continue till now. For this reason, 13 seismograph stations (radio-telemetry network short-period seismometers) were installed around the northern part of Lake Aswan. The main purpose of the network is to monitor the earthquake activity continuously within Aswan region. The data described here are obtained from the continuous record of earthquake activity and lake-water level variation through the period from 1982 to 2015. The seismicity is concentrated in the Kalabsha area, where there is an intersection of the easterly trending Kalabsha fault with the northerly trending faults. The earthquake foci are distributed in two seismic zones, shallow and deep in the crust. Shallow events have focal depths of less than 12 km while deep events extend from 12 to 28 km. Correlation between the seismicity and the water level variation in the lake provides great suggestion to distinguish the micro-earthquakes, particularly, those in shallow seismic zone in the reservoir–triggered seismicity category. The water loading is one factor from several factors, as an activating medium in triggering earthquakes. The common factors for all cases of induced seismicity seem to be the presence of specific geological conditions, the tectonic setting and water loading. The role of the water loading is as a supplementary source of earthquake events. So, the earthquake activity in the area originated tectonically (ML ≥ 4) and the water factor works as an activating medium in triggering small earthquakes (ML ≤ 3). Study of the inducing seismicity from the water level variation in Aswan Lake is of great importance and play great roles necessity for the safety of the High Dam body and its economic resources.

Keywords: Aswan lake, Aswan seismic network, seismicity, water level variation

Procedia PDF Downloads 360

Authors: Alessandro Andre Leme

Abstract:

To analyze three Brazilian think tanks: a) Fernando Henrique Foundation; b) Celso Furtado International Center; c) Millennium Institute and how they dispute interpretations about the type of development and State that should be adopted in Brazil. We will make use of Network and content analysis of the sites. The analyzes show a dispute that goes from a defense of ultraliberalism to developmentalism, going through a hybrid between State and Market voiced in each of the Think Tanks.

Keywords: sociopolitical and economic thinking, development, strategies, intellectuals, state

Procedia PDF Downloads 130

Authors: Zhiyong Zheng, Xu Li, Liang Huang, Zhengliang Sun, Jianhua Xu

Abstract:

Yaw angle plays a significant role in many vehicle safety applications, such as collision avoidance and lane-keeping system. Although the estimation of the yaw angle has been extensively studied in existing literature, it is still the main challenge to simultaneously achieve a reliable and cost-effective solution in complex urban environments. This paper proposes a multi-stage learning framework to estimate the yaw angle with a monocular camera, which can deal with the challenge in a more reliable manner. In the first stage, an efficient road detection network is designed to extract the road region, providing a highly reliable reference for the estimation. In the second stage, a variational auto-encoder (VAE) is proposed to learn the distribution patterns of road regions, which is particularly suitable for modeling the changing patterns of yaw angle under different driving maneuvers, and it can inherently enhance the generalization ability. In the last stage, a gated recurrent unit (GRU) network is used to capture the temporal correlations of the learned patterns, which is capable to further improve the estimation accuracy due to the fact that the changes of deflection angle are relatively easier to recognize among continuous frames. Afterward, the yaw angle can be obtained by combining the estimated deflection angle and the road direction stored in a roadway map. Through effective multi-stage learning, the proposed framework presents high reliability while it maintains better accuracy. Road-test experiments with different driving maneuvers were performed in complex urban environments, and the results validate the effectiveness of the proposed framework.

Keywords: gated recurrent unit, multi-stage learning, reliable estimation, variational auto-encoder, yaw angle

Procedia PDF Downloads 123

Authors: Gong Zhilin, Jing Yang, Jian Yin

Abstract:

The purpose of this paper is to develop a novel system for credit card fraud detection based on sequential modeling of data using hybrid deep learning models. The projected model encapsulates five major phases are pre-processing, imbalance-data handling, feature extraction, optimal feature selection, and fraud detection with an ensemble classifier. The collected raw data (input) is pre-processed to enhance the quality of the data through alleviation of the missing data, noisy data as well as null values. The pre-processed data are class imbalanced in nature, and therefore they are handled effectively with the K-means clustering-based SMOTE model. From the balanced class data, the most relevant features like improved Principal Component Analysis (PCA), statistical features (mean, median, standard deviation) and higher-order statistical features (skewness and kurtosis). Among the extracted features, the most optimal features are selected with the Self-improved Arithmetic Optimization Algorithm (SI-AOA). This SI-AOA model is the conceptual improvement of the standard Arithmetic Optimization Algorithm. The deep learning models like Long Short-Term Memory (LSTM), Convolutional Neural Network (CNN), and optimized Quantum Deep Neural Network (QDNN). The LSTM and CNN are trained with the extracted optimal features. The outcomes from LSTM and CNN will enter as input to optimized QDNN that provides the final detection outcome. Since the QDNN is the ultimate detector, its weight function is fine-tuned with the Self-improved Arithmetic Optimization Algorithm (SI-AOA).

Keywords: credit card, data mining, fraud detection, money transactions

Procedia PDF Downloads 115

Authors: Yuning Guan

Abstract:

Under the background of rapid urbanization, mountainous counties need to break through mountain barriers for urban expansion due to undulating topography, resulting in ecological problems such as landscape fragmentation and reduced biodiversity. Green infrastructure networks are constructed to alleviate the contradiction between urban expansion and ecological protection, promoting the healthy and sustainable development of urban ecosystems. This study applies the MSPA model, the MCR model and Linkage Mapper Tools to identify eco-sources and eco-corridors in the Shapingba District of Chongqing and combined with landscape connectivity assessment and circuit theory to delineate the importance levels to extract ecological pinch point areas on the corridors. The results show that: (1) 20 ecological sources are identified, with a total area of 126.47 km², accounting for 31.88% of the study area, and showing a pattern of ‘one core, three corridors, multi-point distribution’. (2) 37 ecological corridors are formed in the area, with a total length of 62.52km, with a ‘more in the west, less in the east’ pattern. (3) 42 ecological pinch points are extracted, accounting for 25.85% of the length of the corridors, which are mainly distributed in the eastern new area. Accordingly, this study proposes optimization strategies for sub-area protection of ecological sources, grade-level construction of ecological corridors, and precise restoration of ecological pinch points.

Keywords: green infrastructure network, morphological spatial pattern, minimal cumulative resistance, mountainous counties, circuit theory, shapingba district

Procedia PDF Downloads 28

Authors: Tamer ElSerafi

Abstract:

In recent approaches to heritage conservation, the whole context of historic areas becomes as important as the single historic building. This makes the provision of infrastructure and network of mobility an effective element in the urban conservation. Sustainable urban conservation projects consider the high density of activities, the need for a good quality access system to the transit system, and the importance of the configuration of the mobility network by identifying the best way to connect the different districts of the urban area through a complex unique system that helps the synergic development to achieve a sustainable mobility system. A sustainable urban mobility is a key factor in maintaining the integrity between socio-cultural aspects and functional aspects. This paper illustrates the mobility aspects, mobility problems in historic districts, and the needs of the mobility systems in the first part. The second part is a practical analysis for different mobility plans. It is challenging to find innovative and creative conservation solutions fitting modern uses and needs without risking the loss of inherited built resources. Urban mobility management is becoming an essential and challenging issue in the urban conservation projects. Depending on literature review and practical analysis, this paper tries to define and clarify the guidelines for mobility management in historic districts as a key element in sustainability of urban conservation and development projects. Such rules and principles could control the conflict between the socio–cultural and economic activities, and the different needs for mobility in these districts in a sustainable way. The practical analysis includes a comparison between mobility plans which have been implemented in four different cities; Freiburg in Germany, Zurich in Switzerland and Bray Town in Ireland. This paper concludes with a matrix of guidelines that considers both principles of sustainability and livability factors in urban historic districts.

Keywords: sustainable mobility, urban mobility, mobility management, historic districts

Procedia PDF Downloads 147

Authors: Kamal KC, Zhendong Yin, Dasen Li, Zhilu Wu

Abstract:

Intra-plant disease classification based on leaf images is a challenging computer vision task due to similarities in texture, color, and shape of leaves with a slight variation of leaf spot; and external environmental changes such as lighting and background noises. Deep convolutional neural network (DCNN) has proven to be an effective tool for binary classification. In this paper, two methods for binary classification of diseased plant leaves using DCNN are presented; model created from scratch and transfer learning. Our main contribution is a thorough evaluation of 4 networks created from scratch and transfer learning of 5 pre-trained models. Training and testing of these models were performed on a plant leaf images dataset belonging to 16 distinct classes, containing a total of 22,265 images from 8 different plants, consisting of a pair of healthy and diseased leaves. We introduce a deep CNN model, Optimized MobileNet. This model with depthwise separable CNN as a building block attained an average test accuracy of 99.77%. We also present a fine-tuning method by introducing the concept of a convolutional block, which is a collection of different deep neural layers. Fine-tuned models proved to be efficient in terms of accuracy and computational cost. Fine-tuned MobileNet achieved an average test accuracy of 99.89% on 8 pairs of [healthy, diseased] leaf ImageSet.

Keywords: deep convolution neural network, depthwise separable convolution, fine-grained classification, MobileNet, plant disease, transfer learning

Procedia PDF Downloads 169

Authors: Thomas Rowan, Mohammed Seaid

Abstract:

A fast finite volume solver for multi-layered shallow water flows with mass exchange and an erodible bed is developed. This enables the user to solve a number of complex sediment-based problems including (but not limited to), dam-break over an erodible bed, recirculation currents and bed evolution as well as levy and dyke failure. This research develops methodologies crucial to the under-standing of multi-sediment fluvial mechanics and waterway design. In this model mass exchange between the layers is allowed and, in contrast to previous models, sediment and fluid are able to transfer between layers. In the current study we use a two-step finite volume method to avoid the solution of the Riemann problem. Entrainment and deposition rates are calculated for the first time in a model of this nature. In the first step the governing equations are rewritten in a non-conservative form and the intermediate solutions are calculated using the method of characteristics. In the second stage, the numerical fluxes are reconstructed in conservative form and are used to calculate a solution that satisfies the conservation property. This method is found to be considerably faster than other comparative finite volume methods, it also exhibits good shock capturing. For most entrainment and deposition equations a bed level concentration factor is used. This leads to inaccuracies in both near bed level concentration and total scour. To account for diffusion, as no vertical velocities are calculated, a capacity limited diffusion coefficient is used. The additional advantage of this multilayer approach is that there is a variation (from single layer models) in bottom layer fluid velocity: this dramatically reduces erosion, which is often overestimated in simulations of this nature using single layer flows. The model is used to simulate a standard dam break. In the dam break simulation, as expected, the number of fluid layers utilised creates variation in the resultant bed profile, with more layers offering a higher deviation in fluid velocity . These results showed a marked variation in erosion profiles from standard models. The overall the model provides new insight into the problems presented at minimal computational cost.

Keywords: erosion, finite volume method, sediment transport, shallow water equations

Procedia PDF Downloads 208

Authors: Guneet Saini, Shahrukh, Sunil Sharma

Abstract:

Traffic congestion is the most critical issue faced by those in the transportation profession today. Over the past few years, roundabouts have been recognized as a measure to promote efficiency at intersections globally. In developing countries like India, this type of intersection still faces a lot of issues, such as bottleneck situations, long queues and increased waiting times, due to increasing traffic which in turn affect the performance of the entire urban network. This research is a case study of a non-conventional roundabout, in terms of geometric design, in a small town in India. These types of roundabouts should be analyzed for their functionality in mixed traffic conditions, prevalent in many developing countries. Microscopic traffic simulation is an effective tool to analyze traffic conditions and estimate various measures of operational performance of intersections such as capacity, vehicle delay, queue length and Level of Service (LOS) of urban roadway network. This study involves analyzation of an unsymmetrical non-circular 6-legged roundabout known as “Kala Aam Chauraha” in a small town Bulandshahr in Uttar Pradesh, India using VISSIM simulation package which is the most widely used software for microscopic traffic simulation. For coding in VISSIM, data are collected from the site during morning and evening peak hours of a weekday and then analyzed for base model building. The model is calibrated on driving behavior and vehicle parameters and an optimal set of calibrated parameters is obtained followed by validation of the model to obtain the base model which can replicate the real field conditions. This calibrated and validated model is then used to analyze the prevailing operational traffic performance of the roundabout which is then compared with a proposed alternative to improve efficiency of roundabout network and to accommodate pedestrians in the geometry. The study results show that the alternative proposed is an advantage over the present roundabout as it considerably reduces congestion, vehicle delay and queue length and hence, successfully improves roundabout performance without compromising on pedestrian safety. The study proposes similar designs for modification of existing non-conventional roundabouts experiencing excessive delays and queues in order to improve their efficiency especially in the case of developing countries. From this study, it can be concluded that there is a need to improve the current geometry of such roundabouts to ensure better traffic performance and safety of drivers and pedestrians negotiating the intersection and hence this proposal may be considered as a best fit.

Keywords: operational performance, roundabout, simulation, VISSIM

Procedia PDF Downloads 127

Authors: Teesta Dey

Abstract:

Integrated Transport development with proper traffic management leads to sustainable growth of any urban sphere. Appropriate mass transport planning is essential for the populous cities in third world countries like India. The exponential growth of motor vehicles with unplanned road network is now the common feature of major urban centres in India. Kolkata, the third largest mega city in India, is not an exception of it. The imbalance between demand and supply of unplanned transport services in this city is manifested in the high economic and environmental costs borne by the associated society. With the passage of time, the growth and extent of passenger demand for rapid urban transport has outstripped proper infrastructural planning and causes severe transport problems in the overall urban realm. Hence Kolkata stands out in the world as one of the most crisis-ridden metropolises. The urban transport crisis of this city involves severe traffic congestion, the disparity in mass transport services on changing peripheral land uses, route overlapping, lowering of travel speed and faulty implementation of governmental plans as mostly induced by rapid growth of private vehicles on limited road space with huge carbon footprint. Therefore the paper will critically analyze the extant road network pattern for improving regional connectivity and accessibility, assess the degree of congestion, identify the deviation from demand and supply balance and finally evaluate the emerging alternate transport options as promoted by the government. For this purpose, linear, nodal and spatial transport network have been assessed based on certain selected indices viz. Road Degree, Traffic Volume, Shimbel Index, Direct Bus Connectivity, Average Travel and Waiting Tine Indices, Route Variety, Service Frequency, Bus Intensity, Concentration Analysis, Delay Rate, Quality of Traffic Transmission, Lane Length Duration Index and Modal Mix. Total 20 Traffic Intersection Points (TIPs) have been selected for the measurement of nodal accessibility. Critical Congestion Zones (CCZs) are delineated based on one km buffer zones of each TIP for congestion pattern analysis. A total of 480 bus routes are assessed for identifying the deficiency in network planning. Apart from bus services, the combined effects of other mass and para transit modes, containing metro rail, auto, cab and ferry services, are also analyzed. Based on systematic random sampling method, a total of 1500 daily urban passengers’ perceptions were studied for checking the ground realities. The outcome of this research identifies the spatial disparity among the 15 boroughs of the city with severe route overlapping and congestion problem. North and Central Kolkata-based mass transport services exceed the transport strength of south and peripheral Kolkata. Faulty infrastructural condition, service inadequacy, economic loss and workers’ inefficiency are the most dominant reasons behind the defective mass transport network plan. Hence there is an urgent need to revive the extant road based mass transport system of this city by implementing a holistic management approach by upgrading traffic infrastructure, designing new roads, better cooperation among different mass transport agencies, better coordination of transport and changing land use policies, large increase in funding and finally general passengers’ awareness.

Keywords: carbon footprint, critical congestion zones, direct bus connectivity, integrated transport development

Procedia PDF Downloads 260

Authors: Amani Abdallah, Isam Shahrour

Abstract:

The quality of drinking water is a major concern of public health. The control of this quality is generally performed in the laboratory, which requires a long time. This type of control is not adapted for accidental pollution from sudden events, which can have serious consequences on population health. Therefore, it is of major interest to develop real-time innovative solutions for the detection of accidental contamination in drinking water systems This paper presents researches conducted within the SunRise Demonstrator for ‘Smart and Sustainable Cities’ with a particular focus on the supervision of the water quality. This work aims at (i) implementing a smart water system in a large water network (Campus of the University Lille1) including innovative equipment for real-time detection of abnormal events, such as those related to the contamination of drinking water and (ii) develop a numerical modeling of the contamination diffusion in the water distribution system. The first step included verification of the water quality sensors and their effectiveness on a network prototype of 50m length. This part included the evaluation of the efficiency of these sensors in the detection both bacterial and chemical contamination events in drinking water distribution systems. An on-line optical sensor integral with a laboratory-scale distribution system (LDS) was shown to respond rapidly to changes in refractive index induced by injected loads of chemical (cadmium, mercury) and biological contaminations (Escherichia coli). All injected substances were detected by the sensor; the magnitude of the response depends on the type of contaminant introduced and it is proportional to the injected substance concentration.

Keywords: distribution system, drinking water, refraction index, sensor, real-time

Procedia PDF Downloads 338

Authors: Atour Taghipour

Abstract:

In the real world each manufacturing company is an independent business unit. These business units are linked to each other through upstream and downstream linkages. The management of these linkages is called coordination which, could be considered as a difficult engineering task. The degree of difficulty of coordination depends on the type and the nature of information exchanged between partners as well as the structure of relationship from mutual to the network structure. The literature of manufacturing systems comprises a wide range of varieties of methods and approaches of coordination. In fact, two main streams of research can be distinguished: central coordination versus decentralized coordination. In the centralized systems a high degree of information exchanges is required. The high degree of information exchanges sometimes leads to difficulties when independent members do not want to share information. In order to address these difficulties, decentralized approaches of coordination of operations planning decisions based on some minimal information sharing have been proposed in many academic disciplines. This paper first proposes a framework of analysis in order to analyze the proposed approaches in the literature, based on this framework which includes the similarities between approaches we categorize the existing approaches. This classification can be used as a research map for future researches. The result of our paper highlights several opportunities for future research. First, it is proposed to develop more dynamic and stochastic mechanisms of planning coordination of manufacturing units. Second, in order to exploit the complementarities of approaches proposed by diverse science discipline, we propose to integrate the techniques of coordination. Finally, based on our approach we proposed to develop coordination standards to guaranty both the complementarity of these approaches as well as the freedom of companies to adopt any planning tools.

Keywords: network coordination, manufacturing, operations planning, supply chain

Procedia PDF Downloads 264

Authors: R. Barros, J. Azevedo, P. Padrão, M. Gregório, I. Pádua, C. Almeida, C. Rodrigues, P. Fontes, A. Coelho

Abstract:

Recent evidence shows a positive association between nutritional literacy and healthy eating. Traditional nutrition education strategies for childhood obesity prevention have shown weak effect. The Nutriscience project aims to create and evaluate an innovative and multidisciplinary strategy for promoting effective and accessible nutritional information to children, their families, and educators. Nutriscience is a one-year prospective follow-up evaluation study including pre-school children (3-5 y), who attend national schools’ network (29). The project is structured around a web-based intervention, using an on-line interactive platform, and focus on increasing fruit and vegetable consumption, and reducing sugar and salt intake. The platform acts as a social network where educational materials, games, and nutritional challenges are proposed in a gamification approach that promotes family and community social ties. A nutrition Massive Online Open Course is developed for educators, and a national healthy culinary contest will be promoted on TV channel. A parental self-reported questionnaire assessing sociodemographic and nutritional literacy (knowledge, attitudes, skills) is administered (baseline and end of the intervention). We expect that results on nutritional literacy from the presented strategy intervention will give us important information about the best practices for health intervention with kindergarten families. This intervention program using a digital interactive platform could be an educational tool easily adapted and disseminated for childhood obesity prevention.

Keywords: childhood obesity, educational tool, nutritional literacy, web-based intervention

Procedia PDF Downloads 325

Authors: Indunil Jayatilake, Warna Karunasena, Weena Lokuge

Abstract:

An Australian manufacturer has fabricated an innovative GFRP sandwich panel made from E-glass fiber skin and a modified phenolic core for structural applications. Debonding, which refers to separation of skin from the core material in composite sandwiches, is one of the most common types of damage in composites. The presence of debonding is of great concern because it not only severely affects the stiffness but also modifies the dynamic behaviour of the structure. Generally, it is seen that the majority of research carried out has been concerned about the delamination of laminated structures whereas skin-core debonding has received relatively minor attention. Furthermore, it is observed that research done on composite slabs having multiple skin-core debonding is very limited. To address this gap, a comprehensive research investigating dynamic behaviour of composite panels with single and multiple debonding is presented. The study uses finite-element modelling and analyses for investigating the influence of debonding on free vibration behaviour of single and multilayer composite sandwich panels. A broad parametric investigation has been carried out by varying debonding locations, debonding sizes and support conditions of the panels in view of both single and multiple debonding. Numerical models were developed with Strand7 finite element package by innovatively selecting the suitable elements to diligently represent their actual behavior. Three-dimensional finite element models were employed to simulate the physically real situation as close as possible, with the use of an experimentally and numerically validated finite element model. Comparative results and conclusions based on the analyses are presented. For similar extents and locations of debonding, the effect of debonding on natural frequencies appears greatly dependent on the end conditions of the panel, giving greater decrease in natural frequency when the panels are more restrained. Some modes are more sensitive to debonding and this sensitivity seems to be related to their vibration mode shapes. The fundamental mode seems generally the least sensitive mode to debonding with respect to the variation in free vibration characteristics. The results indicate the effectiveness of the developed three-dimensional finite element models in assessing debonding damage in composite sandwich panels

Keywords: debonding, free vibration behaviour, GFRP sandwich panels, three dimensional finite element modelling

Procedia PDF Downloads 300

Authors: Dong Zhao

Abstract:

Even though the recent claim of room temperature superconductivity by LK-99 was confirmed an unsuccessful attempt, this work reawakened people’s century striving to get applicable superconductors with Tc of room temperature or higher and under ambient pressure. One of the efforts was focusing on exploring the thorium salts. This is because certain thorium compounds revealed an unusual property of having both high electrical conductivity and diamagnetism or the so-called “coexistence of high electrical conductivity and diamagnetism.” It is well known that this property of the coexistence of high electrical conductivity and diamagnetism is held by superconductors because of the electron pairings. Consequently, the likelihood for these thorium compounds to have superconducting properties becomes great. However, as a surprise, these thorium salts possess this property at room temperature and atmosphere pressure. This gives rise to solid evidence for these thorium compounds to be room-temperature superconductors without a need for external pressure. Among these thorium compound superconductors claimed in that work, thorium di-iodide (ThI₂) is a unique one and has received comprehensive discussion. ThI₂ was synthesized and structurally analyzed by the single crystal diffraction method in the 1960s. Its special property of coexistence of high electrical conductivity and diamagnetism was revealed. Because of this unique property, a special molecular configuration was sketched. Except for an ordinary oxidation of +2 for the thorium cation, the thorium’s oxidation state in ThI₂ is +4. According to the experimental results, ThI₂‘s actual molecular configuration was determined as an unusual one of [Th4+(e-)2](I-)2. This means that the ThI₂ salt’s cation is composed of a [Th4+(e-)2]2+ cation core. In other words, the cation of ThI₂ is constructed by combining an oxidation state +4 of the thorium atom and a pair of electrons or an electron lone pair located on the thorium atom. This combination of the thorium atom and the electron lone pair leads to an oxidation state +2 for the [Th4+(e-)2]2+ cation core. This special construction of the thorium cation is very distinctive, which is believed to be the factor that grants ThI₂ the room temperature superconductivity. Actually, the key for ThI₂ to become a room-temperature superconductor is this characteristic electron lone pair residing on the thorium atom along with the formation of a network constructed by the thorium atoms. This network specializes in a way that allows the electron lone pairs to hop over it and, thus, to generate the supercurrent. This work will discuss, in detail, the special electrical and magnetic properties of ThI₂ as well as its structural features at ambient conditions. The exploration of how the electron pairing in combination with the structurally specialized network works together to bring ThI₂ into a superconducting state. From the experimental results, strong evidence has definitely pointed out that the ThI₂ should be a superconductor, at least at room temperature and under atmosphere pressure.

Keywords: co-existence of high electrical conductivity and diamagnetism, electron lone pair, room temperature superconductor, special molecular configuration of thorium di-iodide ThI₂

Procedia PDF Downloads 44

Authors: Anjula Negi, D. T. V. Raghu Ramaswamy, Rajneesh Sareen

Abstract:

Road infrastructure is emphatically one of the top most critical infrastructure to the Indian economy. Road network in the country of around 3.3 million km is the second largest in the world. Nationwide statistics released by Ministry of Road, Transport and Highways reveal that every minute an accident happens and one death every 3.7 minutes. This reported scale in terms of safety is a matter of grave concern, and economically represents a national loss of 3% to the GDP. Union Budget 2016-17 has allocated USD 12 billion annually for development and strengthening of roads, an increase of 56% from last year. Thus, highlighting the importance of roads as critical infrastructure. National highway alone represent only 1.7% of the total road linkages, however, carry over 40% of traffic. Further, trends analysed from 2002 -2011 on national highways, indicate that in less than a decade, a 22 % increase in accidents have been reported, but, 68% increase in death fatalities. Paramount inference is that accident severity has increased with time. Over these years many measures to increase road safety, lessening damage to physical assets, reducing vulnerabilities leading to a build-up for resilient road infrastructure have been taken. In the context of national highway development program, policy makers proposed implementation of around 20 % of such road length on PPP mode. These roads were taken up on high-density traffic considerations and for qualitative implementation. In order to understand resilience impacts and safety parameters, enshrined in various PPP concession agreements executed with the private sector partners, such highway specific projects would be appraised. This research paper would attempt to assess such safety measures taken and the possible reasons behind an increase in accident severity through these PPP case study projects. Delving further on safety features to understand policy measures adopted in these cases and an introspection on reasons of severity, whether an outcome of increased speeds, faulty road design and geometrics, driver negligence, or due to lack of discipline in following lane traffic with increased speed. Assessment exercise would study these aspects hitherto to PPP and post PPP project structures, based on literature review and opinion surveys with sectoral experts. On the way forward, it is understood that the Ministry of Road, Transport and Highway’s estimate for strengthening the national highway network is USD 77 billion within next five years. The outcome of this paper would provide an understanding of resilience measures adopted, possible options for accessible and safe road network and its expansion to policy makers for possible policy initiatives and funding allocation in securing critical infrastructure.

Keywords: national highways, policy, PPP, safety

Procedia PDF Downloads 248

Authors: Shayan Mohajer Hamidi

Abstract:

Federated learning has emerged as a promising approach for distributed model training while preserving data privacy. However, the challenges of communication overhead, limited network resources, and slow convergence hinder its widespread adoption. On the other hand, knowledge distillation has shown great potential in compressing large models into smaller ones without significant loss in performance. In this paper, we propose an innovative framework that combines federated learning and knowledge distillation to address these challenges and enhance the efficiency of distributed learning. Our approach, called Federated Knowledge Distillation (FKD), enables multiple clients in a federated learning setting to collaboratively distill knowledge from a teacher model. By leveraging the collaborative nature of federated learning, FKD aims to improve model compression while maintaining privacy. The proposed framework utilizes a coded teacher model that acts as a reference for distilling knowledge to the client models. To demonstrate the effectiveness of FKD, we conduct extensive experiments on various datasets and models. We compare FKD with baseline federated learning methods and standalone knowledge distillation techniques. The results show that FKD achieves superior model compression, faster convergence, and improved performance compared to traditional federated learning approaches. Furthermore, FKD effectively preserves privacy by ensuring that sensitive data remains on the client devices and only distilled knowledge is shared during the training process. In our experiments, we explore different knowledge transfer methods within the FKD framework, including Fine-Tuning (FT), FitNet, Correlation Congruence (CC), Similarity-Preserving (SP), and Relational Knowledge Distillation (RKD). We analyze the impact of these methods on model compression and convergence speed, shedding light on the trade-offs between size reduction and performance. Moreover, we address the challenges of communication efficiency and network resource utilization in federated learning by leveraging the knowledge distillation process. FKD reduces the amount of data transmitted across the network, minimizing communication overhead and improving resource utilization. This makes FKD particularly suitable for resource-constrained environments such as edge computing and IoT devices. The proposed FKD framework opens up new avenues for collaborative and privacy-preserving distributed learning. By combining the strengths of federated learning and knowledge distillation, it offers an efficient solution for model compression and convergence speed enhancement. Future research can explore further extensions and optimizations of FKD, as well as its applications in domains such as healthcare, finance, and smart cities, where privacy and distributed learning are of paramount importance.

Keywords: federated learning, knowledge distillation, knowledge transfer, deep learning

Procedia PDF Downloads 55

Authors: Judith Stein, Elfriede Friedmann

Abstract:

The injection of a drug into the vitreous body for the treatment of retinal diseases like wet aged-related macular degeneration (AMD) is the most common medical intervention worldwide. We develop mathematical models for drug transport in the vitreous body of a human eye to analyse the impact of different rheological models of the vitreous on drug distribution. In addition to the convection diffusion equation characterizing the drug spreading, we use porous media modeling for the healthy vitreous with a dense collagen network and include the steady permeating flow of the aqueous humor described by Darcy's law driven by a pressure drop. Additionally, the vitreous body in a healthy human eye behaves like a viscoelastic gel through the collagen fibers suspended in the network of hyaluronic acid and acts as a drug depot for the treatment of retinal diseases. In a completely liquefied vitreous, we couple the drug diffusion with the classical Navier-Stokes flow equations. We prove the global existence and uniqueness of the weak solution of the developed initial-boundary value problem describing the drug distribution in the healthy vitreous considering the permeating aqueous humor flow in the realistic three-dimensional setting. In particular, for the drug diffusion equation, results from the literature are extended from homogeneous Dirichlet boundary conditions to our mixed boundary conditions that describe the eye with the Galerkin's method using Cauchy-Schwarz inequality and trace theorem. Because there is only a small effective drug concentration range and higher concentrations may be toxic, the ability to model the drug transport could improve the therapy by considering patient individual differences and give a better understanding of the physiological and pathological processes in the vitreous.

Keywords: coupled PDE systems, drug diffusion, mixed boundary conditions, vitreous body

Procedia PDF Downloads 121