Search results for: neural network

Authors: Hooshang Mohammad Alikhani, Javad Nikoukar

Abstract:

Restructuring and deregulation of power industry have changed the objectives of transmission expansion planning and increased the uncertainties. Due to these changes, new approaches and criteria are needed for transmission planning in deregulated power systems. The objective of this research work is to present a new approach for transmission expansion planning with considering new objectives and uncertainties in deregulated power systems. The approach must take into account the desires of all stakeholders in transmission expansion planning. Market based criteria must be defined to achieve the new objectives. Combination of market based criteria, technical criteria and economical criteria must be used for measuring the goodness of expansion plans to achieve market requirements, technical requirements, and economical requirements altogether.

Keywords: deregulated power systems, transmission network, stakeholder, energy systems

Procedia PDF Downloads 654

Authors: Jindong Gu, Matthias Schubert, Volker Tresp

Abstract:

In many outlier detection tasks, only training data belonging to one class, i.e., the positive class, is available. The task is then to predict a new data point as belonging either to the positive class or to the negative class, in which case the data point is considered an outlier. For this task, we propose a novel corrupted Generative Adversarial Network (CorGAN). In the adversarial process of training CorGAN, the Generator generates outlier samples for the negative class, and the Discriminator is trained to distinguish the positive training data from the generated negative data. The proposed framework is evaluated using an image dataset and a real-world network intrusion dataset. Our outlier-detection method achieves state-of-the-art performance on both tasks.

Keywords: one-class classification, outlier detection, generative adversary networks, semi-supervised learning

Procedia PDF Downloads 151

Authors: Juan Pablo Botero Torres, Alba Avila, Luis Felipe Giraldo

Abstract:

Independent navigation and mobility through unknown spaces pose a challenge for the autonomy of visually impaired people (VIP), who have relied on the use of traditional assistive tools like the white cane and trained dogs. However, emerging visually assistive technologies (VAT) have proposed several human-machine interfaces (HMIs) that could improve VIP’s ability for self-guidance. Hereby, we introduce the design and implementation of a visually assistive device, Anajaa – Visual Substitution System (AVSS). This system integrates ultrasonic sensors with custom electronics, and computer vision models (convolutional neural networks), in order to achieve a robust system that acquires information of the surrounding space and transmits it to the user in an intuitive and efficient manner. AVSS consists of two modules: the sensing and the actuation module, which are fitted to a chest mount and belt that communicate via Bluetooth. The sensing module was designed for the acquisition and processing of proximity signals provided by an array of ultrasonic sensors. The distribution of these within the chest mount allows an accurate representation of the surrounding space, discretized in three different levels of proximity, ranging from 0 to 6 meters. Additionally, this module is fitted with an RGB-D camera used to detect potentially threatening obstacles, like staircases, using a convolutional neural network specifically trained for this purpose. Posteriorly, the depth data is used to estimate the distance between the stairs and the user. The information gathered from this module is then sent to the actuation module that creates an HMI, by the means of a 3x2 array of vibration motors that make up the tactile display and allow the system to deliver haptic feedback. The actuation module uses vibrational messages (tactones); changing both in amplitude and frequency to deliver different awareness levels according to the proximity of the obstacle. This enables the system to deliver an intuitive interface. Both modules were tested under lab conditions, and the HMI was additionally tested with a focal group of VIP. The lab testing was conducted in order to establish the processing speed of the computer vision algorithms. This experimentation determined that the model can process 0.59 frames per second (FPS); this is considered as an adequate processing speed taking into account that the walking speed of VIP is 1.439 m/s. In order to test the HMI, we conducted a focal group composed of two females and two males between the ages of 35-65 years. The subject selection was aided by the Colombian Cooperative of Work and Services for the Sightless (COOTRASIN). We analyzed the learning process of the haptic messages throughout five experimentation sessions using two metrics: message discrimination and localization success. These correspond to the ability of the subjects to recognize different tactones and locate them within the tactile display. Both were calculated as the mean across all subjects. Results show that the focal group achieved message discrimination of 70% and a localization success of 80%, demonstrating how the proposed HMI leads to the appropriation and understanding of the feedback messages, enabling the user’s awareness of its surrounding space.

Keywords: computer vision on embedded systems, electronic trave aids, human-machine interface, haptic feedback, visual assistive technologies, vision substitution systems

Procedia PDF Downloads 81

Authors: Li-Hsueh Chen

Abstract:

Efficient operation is more and more important for managers of hotels. Unlike the manufacturing industry, hotels cannot store their products. In addition, many hotels provide room service, and food and beverage service simultaneously. When efficiencies of hotels are evaluated, the internal structure should be considered. Hence, based on the operational characteristics of hotels, this study proposes a DEA model to simultaneously assess the efficiencies among the room production division, food and beverage production division, room service division and food and beverage service division. However, not only the enhancement of efficiency but also the adjustment of scale can improve the performance. In terms of the adjustment of scale, scale elasticity or returns to scale can help to managers to make decisions concerning expansion or contraction. In order to construct a reasonable approach to measure the efficiencies and scale elasticities of hotels, this study builds an alternative variable-returns-to-scale-based two-stage network DEA model with the combination of parallel and series structures to explore the scale elasticities of the whole system, room production division, food and beverage production division, room service division and food and beverage service division based on the data of international tourist hotel industry in Taiwan. The results may provide valuable information on operational performance and scale for managers and decision makers.

Keywords: efficiency, scale elasticity, network data envelopment analysis, international tourist hotel

Procedia PDF Downloads 225

Authors: J. M. Counsell, Y. Khalid, M. J. Stewart

Abstract:

Achieving nearly zero carbon heating continues to be identified by UK government analysis as an important feature of any lowest cost pathway to reducing greenhouse gas emissions. Heat currently accounts for 48% of UK energy consumption and approximately one third of UK’s greenhouse gas emissions. Heat Networks are being promoted by UK investment policies as one means of supporting hybrid heat pump based solutions. To this effect the RISE (Renewable Integrated and Sustainable Electric) heating system project is investigating how an all-electric heating sourceshybrid configuration could play a key role in long-term decarbonisation of heat.  For the purposes of this study, hybrid systems are defined as systems combining the technologies of an electric driven air source heat pump, electric powered thermal storage, a thermal vessel and micro-heat network as an integrated system.  This hybrid strategy allows for the system to store up energy during periods of low electricity demand from the national grid, turning it into a dynamic supply of low cost heat which is utilized only when required. Currently a prototype of such a system is being tested in a modern house integrated with advanced controls and sensors. This paper presents the virtual performance analysis of the system and its design for a micro heat network with multiple dwelling units. The results show that the RISE system is controllable and can reduce carbon emissions whilst being competitive in running costs with a conventional gas boiler heating system.

Keywords: gas boilers, heat pumps, hybrid heating and thermal storage, renewable integrated and sustainable electric

Procedia PDF Downloads 419

Authors: Wade Ghribi, Abdelmoty M. Ahmed, Ahmed Said Badawy, Belgacem Bouallegue

Abstract:

In higher education institutions, the most pressing priority is to improve student performance and retention. Large volumes of student data are used in Educational Data Mining techniques to find new hidden information from students' learning behavior, particularly to uncover the early symptom of at-risk pupils. On the other hand, data with noise, outliers, and irrelevant information may provide incorrect conclusions. By identifying features of students' data that have the potential to improve performance prediction results, comparing and identifying the most appropriate ensemble learning technique after preprocessing the data, and optimizing the hyperparameters, this paper aims to develop a reliable students' performance prediction model for Higher Education Institutions. Data was gathered from two different systems: a student information system and an e-learning system for undergraduate students in the College of Computer Science of a Saudi Arabian State University. The cases of 4413 students were used in this article. The process includes data collection, data integration, data preprocessing (such as cleaning, normalization, and transformation), feature selection, pattern extraction, and, finally, model optimization and assessment. Random Forest, Bagging, Stacking, Majority Vote, and two types of Boosting techniques, AdaBoost and XGBoost, are ensemble learning approaches, whereas Decision Tree, Support Vector Machine, and Artificial Neural Network are supervised learning techniques. Hyperparameters for ensemble learning systems will be fine-tuned to provide enhanced performance and optimal output. The findings imply that combining features of students' behavior from e-learning and students' information systems using Majority Vote produced better outcomes than the other ensemble techniques.

Keywords: educational data mining, student performance prediction, e-learning, classification, ensemble learning, higher education

Procedia PDF Downloads 108

Authors: Markus Mazzola, Gunther Schaaf

Abstract:

A vehicle driving with an Adaptive Cruise Control System (ACC) is usually controlled decentrally, based on the information of radar systems and in some publications based on C2X-Communication (CACC) to guarantee stable platoons. In this paper, we present a Model Predictive Control (MPC) design of a centralized, server-based ACC-System, whereby the vehicular platoon is modeled and controlled as a whole. It is then proven that the proposed MPC design guarantees asymptotic stability and hence string stability of the platoon. The Networked MPC design is chosen to be able to integrate system constraints optimally as well as to reduce the effects of communication delay and packet loss. The performance of the proposed controller is then simulated and analyzed in an LTE communication scenario using the LTE/EPC Network Simulator LENA, which is based on the ns-3 network simulator.

Keywords: adaptive cruise control, centralized server, networked model predictive control, string stability

Procedia PDF Downloads 515

Authors: Fuyuan Ma, Yuhan Wang, Junhe Zhang, Ying Wang

Abstract:

Identifying the sources of information dissemination is a pivotal task in the study of collective behaviors in networks, enabling us to discern and intercept the critical pathways through which information propagates from its origins. This allows for the control of the information’s dissemination impact in its early stages. Numerous methods for source detection rely on pre-existing, underlying propagation models as prior knowledge. Current models that eschew prior knowledge attempt to harness label propagation algorithms to model the statistical characteristics of propagation states or employ Graph Neural Networks (GNNs) for deep reverse modeling of the diffusion process. These approaches are either deficient in modeling the propagation patterns of information or are constrained by the over-smoothing problem inherent in GNNs, which limits the stacking of sufficient model depth to excavate global propagation patterns. Consequently, we introduce the ODESI model. Initially, the model employs a label propagation algorithm to delineate the distribution density of infected states within a graph structure and extends the representation of infected states from integers to state vectors, which serve as the initial states of nodes. Subsequently, the model constructs a deep architecture based on GNNs-coupled Ordinary Differential Equations (ODEs) to model the global propagation patterns of continuous propagation processes. Addressing the challenges associated with solving ODEs on graphs, we approximate the analytical solutions to reduce computational costs. Finally, we conduct simulation experiments on two real-world social network datasets, and the results affirm the efficacy of our proposed ODESI model in source identification tasks.

Keywords: source identification, ordinary differential equations, label propagation, complex networks

Procedia PDF Downloads 20

Authors: Marius Gavrilescu, Sabina-Adriana Floria, Florin Leon, Silvia Curteanu, Costel Anton

Abstract:

Natural sciences provide a wide range of experimental data whose related problems require study and modeling beyond the capabilities of conventional methodologies. Such problems have solution spaces whose complexity and high dimensionality require correspondingly complex regression methods for proper characterization. In this context, we propose an optimization method which consists in a hybrid dual optimizer setup: a global optimizer based on a modified variant of the popular Imperialist Competitive Algorithm (ICA), and a local optimizer based on a gradient descent approach. The ICA is modified such that intermediate solution populations are more quickly and efficiently pruned of low-fitness individuals by appropriately altering the assimilation, revolution and competition phases, which, combined with an initialization strategy based on low-discrepancy sampling, allows for a more effective exploration of the corresponding solution space. Subsequently, gradient-based optimization is used locally to seek the optimal solution in the neighborhoods of the solutions found through the modified ICA. We use this combined approach to find the optimal configuration and weights of a fully-connected neural network, resulting in regression models used to characterize the process of obtained bricks using silicon-based materials. Installations in the raw ceramics industry, i.e., bricks, are characterized by significant energy consumption and large quantities of emissions. Thus, the purpose of our approach is to determine by simulation the working conditions, including the manufacturing mix recipe with the addition of different materials, to minimize the emissions represented by CO and CH4. Our approach determines regression models which perform significantly better than those found using the traditional ICA for the aforementioned problem, resulting in better convergence and a substantially lower error.

Keywords: optimization, biologically inspired algorithm, regression models, bricks, emissions

Procedia PDF Downloads 82

Authors: Hadj Sahraoui Omar, Kebir Lahcen Wahib, Bennia Ahmed

Abstract:

Interferometric Synthetic Aperture Radar (InSAR) coherence is a crucial parameter for accurately monitoring ground deformation and environmental changes. However, coherence can be degraded by various factors such as temporal decorrelation, atmospheric disturbances, and geometric misalignments, limiting the reliability of InSAR measurements (Omar Hadj‐Sahraoui and al. 2019). To address this challenge, we propose an innovative hybrid approach that combines artificial intelligence (AI) with advanced filtering techniques to optimize interferometric coherence in InSAR data. Specifically, we introduce a Convolutional Neural Network (CNN) integrated with the Lee filter to enhance the performance of radar interferometry. This hybrid method leverages the strength of CNNs to automatically identify and mitigate the primary sources of decorrelation, while the Lee filter effectively reduces speckle noise, improving the overall quality of interferograms. We develop a deep learning-based model trained on multi-temporal and multi-frequency SAR datasets, enabling it to predict coherence patterns and enhance low-coherence regions. This hybrid CNN-SAR with Lee filtering significantly reduces noise and phase unwrapping errors, leading to more precise deformation maps. Experimental results demonstrate that our approach improves coherence by up to 30% compared to traditional filtering techniques, making it a robust solution for challenging scenarios such as urban environments, vegetated areas, and rapidly changing landscapes. Our method has potential applications in geohazard monitoring, urban planning, and environmental studies, offering a new avenue for enhancing InSAR data reliability through AI-powered optimization combined with robust filtering techniques.

Keywords: CNN-SAR, Lee Filter, hybrid optimization, coherence, InSAR phase unwrapping, speckle noise reduction

Procedia PDF Downloads 12

Authors: Hediyeh Talebi, Shokoofeh Ghiam, Changiz Eslahchi

Abstract:

Traditionally, Alzheimer’s disease research has focused on genes with significant fold changes, potentially neglecting subtle but biologically important alterations. Our study introduces an integrative approach that highlights genes crucial to underlying biological processes, regardless of their fold change magnitude. Alzheimer's Single-cell RNA-seq data related to the peripheral blood mononuclear cells (PBMC) was extracted from the Gene Expression Omnibus (GEO). After quality control, normalization, scaling, batch effect correction, and clustering, differentially expressed genes (DEGs) were identified with adjusted p-values less than 0.05. These DEGs were categorized based on cell-type, resulting in four datasets, each corresponding to a distinct cell type. To distinguish between cells from healthy individuals and those with Alzheimer's, an adversarial autoencoder with a classifier was employed. This allowed for the separation of healthy and diseased samples. To identify the most influential genes in this classification, the weight matrices in the network, which includes the encoder and classifier components, were multiplied, and focused on the top 20 genes. The analysis revealed that while some of these genes exhibit a high fold change, others do not. These genes, which may be overlooked by previous methods due to their low fold change, were shown to be significant in our study. The findings highlight the critical role of genes with subtle alterations in diagnosing Alzheimer's disease, a facet frequently overlooked by conventional methods. These genes demonstrate remarkable discriminatory power, underscoring the need to integrate biological relevance with statistical measures in gene prioritization. This integrative approach enhances our understanding of the molecular mechanisms in Alzheimer’s disease and provides a promising direction for identifying potential therapeutic targets.

Keywords: alzheimer's disease, single-cell RNA-seq, neural networks, blood biomarkers

Procedia PDF Downloads 66

Authors: Pius Marthin, Nihal Ata Tutkun

Abstract:

Survival analysis has become one of the paramount procedures in the modeling of time-to-event data. When we encounter complex survival problems, the traditional approach remains limited in accounting for the complex correlational structure between the covariates and the outcome due to the strong assumptions that limit the inference and prediction ability of the resulting models. Several studies exist on the deep learning approach to survival modeling; moreover, the application for the case of complex survival problems still needs to be improved. In addition, the existing models need to address the data structure's complexity fully and are subject to noise and redundant information. In this study, we design a deep learning technique (CmpXRnnSurv_AE) that obliterates the limitations imposed by traditional approaches and addresses the above issues to jointly predict the risk-specific probabilities and survival function for recurrent events with competing risks. We introduce the component termed Risks Information Weights (RIW) as an attention mechanism to compute the weighted cumulative incidence function (WCIF) and an external auto-encoder (ExternalAE) as a feature selector to extract complex characteristics among the set of covariates responsible for the cause-specific events. We train our model using synthetic and real data sets and employ the appropriate metrics for complex survival models for evaluation. As benchmarks, we selected both traditional and machine learning models and our model demonstrates better performance across all datasets.

Keywords: cumulative incidence function (CIF), risk information weight (RIW), autoencoders (AE), survival analysis, recurrent events with competing risks, recurrent neural networks (RNN), long short-term memory (LSTM), self-attention, multilayers perceptrons (MLPs)

Procedia PDF Downloads 90

Authors: Dibakar Chakrabarty, Mebada Suiting

Abstract:

Rapid urbanization coupled with changes in land use pattern results in increasing peak discharge and shortening of catchment time of concentration. The consequence is floods, which often inundate roads and inhabited areas of cities and towns. Management of storm water resulting from rainfall has, therefore, become an important issue for the municipal bodies. Proper management of storm water obviously includes adequate design of storm water drainage networks. The design of storm water network is a costly exercise. Least cost design of storm water networks assumes significance, particularly when the fund available is limited. Optimal design of a storm water system is a difficult task as it involves the design of various components, like, open or closed conduits, storage units, pumps etc. In this paper, a methodology for least cost design of storm water drainage systems is proposed. The methodology proposed in this study consists of coupling a storm water simulator with an optimization method. The simulator used in this study is EPA’s storm water management model (SWMM), which is linked with Genetic Algorithm (GA) optimization method. The model proposed here is a mixed integer nonlinear optimization formulation, which takes care of minimizing the sectional areas of the open conduits of storm water networks, while satisfactorily conveying the runoff resulting from rainfall to the network outlet. Performance evaluations of the developed model show that the proposed method can be used for cost effective design of open conduit based storm water networks.

Keywords: genetic algorithm (GA), optimal design, simulation-optimization, storm water network, SWMM

Procedia PDF Downloads 248

Authors: Cyrillus Prabandana

Abstract:

Supply chain optimisation requires multiple factors as consideration or constraints. These factors are including but not limited to demand forecasting, raw material fulfilment, production capacity, inventory level, facilities locations, transportation means, and manpower availability. By knowing all manageable factors involved and assuming the uncertainty with pre-defined percentage factors, an integrated supply chain model could be developed to manage various business scenarios. This paper analyse the utilisation of geographical point of view to develop an integrated supply chain network model to optimise the distribution of finished product appropriately according to forecasted demand and available supply. The supply chain optimisation model shows that small change in one supply chain constraint is possible to largely impact other constraints, and the new information from the model should be able to support the decision making process. The model was focused on three areas, i.e. raw material fulfilment, production capacity and finished products transportation. To validate the model suitability, it was implemented in a project aimed to optimise the concrete supply chain in a mining location. The high level of operations complexity and involvement of multiple stakeholders in the concrete supply chain is believed to be sufficient to give the illustration of the larger scope. The implementation of this geographical supply chain network modeling resulted an optimised concrete supply chain from raw material fulfilment until finished products distribution to each customer, which indicated by lower percentage of missed concrete order fulfilment to customer.

Keywords: decision making, geographical supply chain modeling, supply chain optimisation, supply chain

Procedia PDF Downloads 346

Authors: Radhia Toujani, Jalel Akaichi

Abstract:

In recent years, the study of community detection in social networks has received great attention. The hierarchical structure of the network leads to the emergence of the convergence to a locally optimal community structure. In this paper, we aim to avoid this local optimum in the introduced hybrid hierarchical method. To achieve this purpose, we present an objective function where we incorporate the value of structural and semantic similarity based modularity and a metaheuristic namely bees colonies algorithm to optimize our objective function on both hierarchical level divisive and agglomerative. In order to assess the efficiency and the accuracy of the introduced hybrid bee colony model, we perform an extensive experimental evaluation on both synthetic and real networks.

Keywords: social network, community detection, agglomerative hierarchical clustering, divisive hierarchical clustering, similarity, modularity, metaheuristic, bee colony

Procedia PDF Downloads 379

Authors: Najiba Ouled Omar, Azza Harbaoui, Henda Ben Ghezala

Abstract:

Current research practices sentiment analysis with a focus on social networks, DEfi Fouille de Texte (DEFT) (Text Mining Challenge) evaluation campaign focuses on opinion mining and sentiment analysis on social networks, especially social network Twitter. It aims to confront the systems produced by several teams from public and private research laboratories. DEFT offers participants the opportunity to work on regularly renewed themes and proposes to work on opinion mining in several editions. The purpose of this article is to scrutinize and analyze the works relating to opinions mining and sentiment analysis in the Twitter social network realized by DEFT. It examines the tasks proposed by the organizers of the challenge and the methods used by the participants.

Keywords: opinion mining, sentiment analysis, emotion, polarity, annotation, OSEE, figurative language, DEFT, Twitter, Tweet

Procedia PDF Downloads 139

Authors: Andrew Laghos

Abstract:

The Internet is a big part of our everyday lives. People can now access the internet from a variety of places including home, college, and work. Many airports, hotels, restaurants and cafeterias, provide free wireless internet to their visitors. Using technologies like computers, tablets, and mobile phones, we spend a lot of our time online getting entertained, getting informed, and communicating with each other. This study deals with the latter part, namely, human communication through the Internet. People can communicate with each other using social media, social network sites (SNS), e-mail, messengers, chatrooms, and so on. By connecting with each other they form virtual communities. Regarding SNS, types of connections that can be studied include friendships and cliques. Analyzing these connections is important to help us understand online user behavior. The method of Social Network Analysis (SNA) was used on a case study, and results revealed the existence of some useful patterns of interactivity between the participants. The study ends with implications of the results and ideas for future research.

Keywords: human communication, internet communities, online user behavior, psychology

Procedia PDF Downloads 497

Authors: Mujahid Muhammad, Paul Kearney, Adel Aneiba

Abstract:

The ability for vehicles to communicate with other vehicles (V2V), the physical (V2I) and network (V2N) infrastructures, pedestrians (V2P), etc. – collectively known as V2X (Vehicle to Everything) – will enable a broad and growing set of applications and services within the intelligent transport domain for improving road safety, alleviate traffic congestion and support autonomous driving. The telecommunication research and industry communities and standardization bodies (notably 3GPP) has finally approved in Release 14, cellular communications connectivity to support V2X communication (known as LTE – V2X). LTE – V2X system will combine simultaneous connectivity across existing LTE network infrastructures via LTE-Uu interface and direct device-to-device (D2D) communications. In order for V2X services to function effectively, a robust security mechanism is needed to ensure legal and safe interaction among authenticated V2X entities in the LTE-based V2X architecture. The characteristics of vehicular networks, and the nature of most V2X applications, which involve human safety makes it significant to protect V2X messages from attacks that can result in catastrophically wrong decisions/actions include ones affecting road safety. Attack vectors include impersonation attacks, modification, masquerading, replay, MiM attacks, and Sybil attacks. In this paper, we focus our attention on LTE-based V2X security and access control mechanisms. The current LTE-A security framework provides its own access authentication scheme, the AKA protocol for mutual authentication and other essential cryptographic operations between UEs and the network. V2N systems can leverage this protocol to achieve mutual authentication between vehicles and the mobile core network. However, this protocol experiences technical challenges, such as high signaling overhead, lack of synchronization, handover delay and potential control plane signaling overloads, as well as privacy preservation issues, which cannot satisfy the adequate security requirements for majority of LTE-based V2X services. This paper examines these challenges and points to possible ways by which they can be addressed. One possible solution, is the implementation of the distributed peer-to-peer LTE security mechanism based on the Bitcoin/Namecoin framework, to allow for security operations with minimal overhead cost, which is desirable for V2X services. The proposed architecture can ensure fast, secure and robust V2X services under LTE network while meeting V2X security requirements.

Keywords: authentication, long term evolution, security, vehicle-to-everything

Procedia PDF Downloads 167

Authors: Xu Jie

Abstract:

As a key network security method, cryptographic services must fully cope with problems such as the wide variety of cryptographic algorithms, high concurrency requirements, random job crossovers, and instantaneous surges in workloads. Its complexity and dynamics also make it difficult for traditional static security policies to cope with the ever-changing situation. Cyber Threats and Environment. Traditional resource scheduling algorithms are inadequate when facing complex decisionmaking problems in dynamic environments. A network cryptographic resource allocation algorithm based on reinforcement learning is proposed, aiming to optimize task energy consumption, migration cost, and fitness of differentiated services (including user, data, and task security). By modeling the multi-job collaborative cryptographic service scheduling problem as a multiobjective optimized job flow scheduling problem, and using a multi-agent reinforcement learning method, efficient scheduling and optimal configuration of cryptographic service resources are achieved. By introducing reinforcement learning, resource allocation strategies can be adjusted in real time in a dynamic environment, improving resource utilization and achieving load balancing. Experimental results show that this algorithm has significant advantages in path planning length, system delay and network load balancing, and effectively solves the problem of complex resource scheduling in cryptographic services.

Keywords: multi-agent reinforcement learning, non-stationary dynamics, multi-agent systems, cooperative and competitive agents

Procedia PDF Downloads 24

Authors: Altilio Rosa, Chirico Francesco, Foglia Goffredo

Abstract:

In recent years, the number of Unmanned Aerial Vehicles (UAVs) has significantly increased. The rapid development of commercial and recreational drones makes them an important part of our society. Despite the growing list of their applications, these vehicles pose a huge threat to civil and military installations: detection, classification and neutralization of such flying objects become an urgent need. Radar is an effective remote sensing tool for detecting and tracking flying objects, but scenarios characterized by the presence of a high number of tracks related to flying birds make especially challenging the drone detection task: operator PPI is cluttered with a huge number of potential threats and his reaction time can be severely affected. Flying birds compared to UAVs show similar velocity, RADAR cross-section and, in general, similar characteristics. Building from the absence of a single feature that is able to distinguish UAVs and birds, this paper uses a multiple features approach where an original feature selection technique is developed to feed binary classifiers trained to distinguish birds and UAVs. RADAR tracks acquired on the field and related to different UAVs and birds performing various trajectories were used to extract specifically designed target movement-related features based on velocity, trajectory and signal strength. An optimization strategy based on a genetic algorithm is also introduced to select the optimal subset of features and to estimate the performance of several classification algorithms (Neural network, SVM, Logistic regression…) both in terms of the number of selected features and misclassification error. Results show that the proposed methods are able to reduce the dimension of the data space and to remove almost all non-drone false targets with a suitable classification accuracy (higher than 95%).

Keywords: birds, classification, machine learning, UAVs

Procedia PDF Downloads 222

Authors: Dongwon Lee, Sunmyeng Kim

Abstract:

Underwater acoustic networks (UANs) have become a very active research area in recent years. Compared with wireless networks, UANs are characterized by the limited bandwidth, long propagation delay and high channel dynamic in acoustic modems, which pose challenges to the design of medium access control (MAC) protocol. The characteristics severely affect network performance. In this paper, we study a MS-MAC (Multi-Sender MAC) protocol in order to improve network performance. The proposed protocol exploits temporal reuse by learning the propagation delays to neighboring nodes. A source node locally calculates the transmission schedules of its neighboring nodes and itself based on the propagation delays to avoid collisions. Performance evaluation is conducted using simulation, and confirms that the proposed protocol significantly outperforms the previous protocol in terms of throughput.

Keywords: acoustic channel, MAC, temporal reuse, UAN

Procedia PDF Downloads 350

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

Abstract:

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

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

Procedia PDF Downloads 74

Authors: E. O. E. Nnadi

Abstract:

Transportation is a critical factor in the economic growth and development of any nation, region or state. Good transportation network supports every sector of the economy like the manufacturing, transportation and encourages investors thereby affect the overall economic prosperity. The paper evaluates the impact of transportation on the economic growth of Nigeria using south eastern states as a case study. The choice of the case study is its importance as the commercial and industrial nerve of the country. About 200 respondents who are of different professions such as dealers in goods, transporters, contractors, consultants, bankers were selected and a set of questionnaire were administered to using the systematic sampling technique in the five states of the region. Descriptive statistics and relative importance index (RII) technique was employed for the analysis of the data gathered. The findings of the analysis reveal that Nigeria has the least effective ratio per population in Africa of 949.91 km/Person. Conclusion was drawn to improve road network in the area and the country as a whole to enhance the economic activities of the people.

Keywords: economic growth, south-east, transportation, transportation cost, Nigeria

Procedia PDF Downloads 273

Authors: Somdatta Saha

Abstract:

The main purpose was to provide an implementable design scenario for underground coal mines using wireless sensor networks (WSNs). The main reason being that given the intricacies in the physical structure of a coal mine, only low power WSN nodes can produce accurate surveillance and accident detection data. The work mainly concentrated on designing and simulating various alternate scenarios for a typical mine and comparing them based on the obtained results to arrive at a final design. In the Era of embedded technology, the Zigbee protocols are used in more and more applications. Because of the rapid development of sensors, microcontrollers, and network technology, a reliable technological condition has been provided for our automatic real-time monitoring of coal mine. The underground system collects temperature, humidity and methane values of coal mine through sensor nodes in the mine; it also collects the number of personnel inside the mine with the help of an IR sensor, and then transmits the data to information processing terminal based on ARM.

Keywords: ARM, embedded board, wireless sensor network (Zigbee)

Procedia PDF Downloads 340

Authors: Wan Jen Chang

Abstract:

Back ground: Taiwan's current adult lifelong learning services have expanded from vocational training to universal lifelong learning. However, both the professional knowledge training of learning guidance and consulting services and the provision of adult online learning consulting service systems still need to be established. Purpose: The purposes of this study are as follows: 1. Analyze the professional training mechanism for cultivating adult lifelong learning consultation and coaching; 2. Explore the feasibility of constructing a system that uses network technology to provide adult learning consultation services. Research design: This study conducts a literature analysis of counseling and coaching policy reports on lifelong learning in European countries and the United States. There are two focus discussions were conducted with 15 lifelong learning scholars, experts and practitioners as research subjects. The following two topics were discussed and suggested: 1. The current situation, needs and professional ability training mechanism of "Adult Lifelong Learning Consulting and Services"; 2. Strategies for establishing an "Adult Lifelong Learning Consulting and Service internet System". Conclusion: 1.Based on adult lifelong learning consulting and service needs, plan a professional knowledge training and certification system.2.Adult lifelong learning consulting and service professional knowledge and skills training should include the use of network technology to provide consulting service skills.3.To establish an adult lifelong learning consultation and service system, the Ministry of Education should promulgate policies and measures at the central level and entrust local governments or private organizations to implement them.4.The adult lifelong learning consulting and service system can combine the national qualifications framework, private sector and NPO to expand learning consulting service partners.

Keywords: adult lifelong learning, profesional knowledge, consulting and service, network system

Procedia PDF Downloads 67

Authors: Roshanak Khodabakhsh Jolfaei, Javad Akbari Torkestani

Abstract:

As a cooperative environment for problem-solving, it is necessary that grids develop efficient job scheduling patterns with regard to their goals, domains and structure. Since the Grid environments facilitate distributed calculations, job scheduling appears in the form of a critical problem for the management of Grid sources that influences severely on the efficiency for the whole Grid environment. Due to the existence of some specifications such as sources dynamicity and conditions of the network in Grid, some algorithm should be presented to be adjustable and scalable with increasing the network growth. For this purpose, in this paper a job scheduling algorithm has been presented on the basis of learning automata in computational Grid which the performance of its results were compared with FPSO algorithm (Fuzzy Particle Swarm Optimization algorithm) and GJS algorithm (Grid Job Scheduling algorithm). The obtained numerical results indicated the superiority of suggested algorithm in comparison with FPSO and GJS. In addition, the obtained results classified FPSO and GJS in the second and third position respectively after the mentioned algorithm.

Keywords: computational grid, job scheduling, learning automata, dynamic scheduling

Procedia PDF Downloads 343

Authors: Richard Ren

Abstract:

Air pollution is a serious danger to international well-being and economies - it will kill an estimated 7 million people every year, costing world economies $2.6 trillion by 2060 due to sick days, healthcare costs, and reduced productivity. In the United States alone, 60,000 premature deaths are caused by poor air quality. For this reason, there is a crucial need to develop effective methods to forecast air quality, which can mitigate air pollution’s detrimental public health effects and associated costs by helping people plan ahead and avoid exposure. The goal of this study is to propose an artificial intelligence framework for predicting future air quality based on timing variables (i.e. season, weekday/weekend), future weather forecasts, as well as past pollutant and air quality measurements. The proposed framework utilizes multiple machine learning algorithms (logistic regression, random forest, neural network) with different specifications and averages the results of the three top-performing models to eliminate inaccuracies, weaknesses, and biases from any one individual model. Over time, the proposed framework uses new data to self-adjust model parameters and increase prediction accuracy. To demonstrate its applicability, a prototype of this framework was created to forecast air quality in Los Angeles, California using datasets from the RP4 weather data repository and EPA pollutant measurement data. The results showed good agreement between the framework’s predictions and real-life observations, with an overall 92% model accuracy. The combined model is able to predict more accurately than any of the individual models, and it is able to reliably forecast season-based variations in air quality levels. Top air quality predictor variables were identified through the measurement of mean decrease in accuracy. This study proposed and demonstrated the efficacy of a comprehensive air quality prediction framework leveraging multiple machine learning algorithms to overcome individual algorithm shortcomings. Future enhancements should focus on expanding and testing a greater variety of modeling techniques within the proposed framework, testing the framework in different locations, and developing a platform to automatically publish future predictions in the form of a web or mobile application. Accurate predictions from this artificial intelligence framework can in turn be used to save and improve lives by allowing individuals to protect their health and allowing governments to implement effective pollution control measures.Air pollution is a serious danger to international wellbeing and economies - it will kill an estimated 7 million people every year, costing world economies $2.6 trillion by 2060 due to sick days, healthcare costs, and reduced productivity. In the United States alone, 60,000 premature deaths are caused by poor air quality. For this reason, there is a crucial need to develop effective methods to forecast air quality, which can mitigate air pollution’s detrimental public health effects and associated costs by helping people plan ahead and avoid exposure. The goal of this study is to propose an artificial intelligence framework for predicting future air quality based on timing variables (i.e. season, weekday/weekend), future weather forecasts, as well as past pollutant and air quality measurements. The proposed framework utilizes multiple machine learning algorithms (logistic regression, random forest, neural network) with different specifications and averages the results of the three top-performing models to eliminate inaccuracies, weaknesses, and biases from any one individual model. Over time, the proposed framework uses new data to self-adjust model parameters and increase prediction accuracy. To demonstrate its applicability, a prototype of this framework was created to forecast air quality in Los Angeles, California using datasets from the RP4 weather data repository and EPA pollutant measurement data. The results showed good agreement between the framework’s predictions and real-life observations, with an overall 92% model accuracy. The combined model is able to predict more accurately than any of the individual models, and it is able to reliably forecast season-based variations in air quality levels. Top air quality predictor variables were identified through the measurement of mean decrease in accuracy. This study proposed and demonstrated the efficacy of a comprehensive air quality prediction framework leveraging multiple machine learning algorithms to overcome individual algorithm shortcomings. Future enhancements should focus on expanding and testing a greater variety of modeling techniques within the proposed framework, testing the framework in different locations, and developing a platform to automatically publish future predictions in the form of a web or mobile application. Accurate predictions from this artificial intelligence framework can in turn be used to save and improve lives by allowing individuals to protect their health and allowing governments to implement effective pollution control measures.Air pollution is a serious danger to international wellbeing and economies - it will kill an estimated 7 million people every year, costing world economies $2.6 trillion by 2060 due to sick days, healthcare costs, and reduced productivity. In the United States alone, 60,000 premature deaths are caused by poor air quality. For this reason, there is a crucial need to develop effective methods to forecast air quality, which can mitigate air pollution’s detrimental public health effects and associated costs by helping people plan ahead and avoid exposure. The goal of this study is to propose an artificial intelligence framework for predicting future air quality based on timing variables (i.e. season, weekday/weekend), future weather forecasts, as well as past pollutant and air quality measurements. The proposed framework utilizes multiple machine learning algorithms (logistic regression, random forest, neural network) with different specifications and averages the results of the three top-performing models to eliminate inaccuracies, weaknesses, and biases from any one individual model. Over time, the proposed framework uses new data to self-adjust model parameters and increase prediction accuracy. To demonstrate its applicability, a prototype of this framework was created to forecast air quality in Los Angeles, California using datasets from the RP4 weather data repository and EPA pollutant measurement data. The results showed good agreement between the framework’s predictions and real-life observations, with an overall 92% model accuracy. The combined model is able to predict more accurately than any of the individual models, and it is able to reliably forecast season-based variations in air quality levels. Top air quality predictor variables were identified through the measurement of mean decrease in accuracy. This study proposed and demonstrated the efficacy of a comprehensive air quality prediction framework leveraging multiple machine learning algorithms to overcome individual algorithm shortcomings. Future enhancements should focus on expanding and testing a greater variety of modeling techniques within the proposed framework, testing the framework in different locations, and developing a platform to automatically publish future predictions in the form of a web or mobile application. Accurate predictions from this artificial intelligence framework can in turn be used to save and improve lives by allowing individuals to protect their health and allowing governments to implement effective pollution control measures.

Keywords: air quality prediction, air pollution, artificial intelligence, machine learning algorithms

Procedia PDF Downloads 127

Authors: B. S. Yadav, A. Mani, S. Srivastava

Abstract:

Chickpea (Cicer arietinum L.) is an annual, self-pollinating, diploid (2n = 2x = 16) pulse crop that ranks second in world legume production after common bean (Phaseolus vulgaris). ICC 4958 flowers approximately 39 days after sowing under peninsular Indian conditions and the crop matures in less than 90 days in rained environments. The estimated collective yield losses due to abiotic stresses (6.4 million t) have been significantly higher than for biotic stresses (4.8 million t). Most legumes are known to be salt sensitive, and therefore, it is becoming increasingly important to produce cultivars tolerant to high-salinity in addition to other abiotic and biotic stresses for sustainable chickpea production. Our aim was to identify the genes that are involved in the defence mechanism against heavy metal toxicity in chickpea and establish the biological network of heavy metal toxicity in chickpea. ICC4958 variety of chick pea was taken and grown in normal condition and 150µM concentration of different heavy metal salt like CdCl₂, K₂Cr2O₇, NaAsO₂. At 15th day leave samples were collected and stored in RNA Later solution microarray was performed for checking out differential gene expression pattern. Our studies revealed that 111 common genes that involved in defense mechanism were up regulated and 41 genes were commonly down regulated during treatment of 150µM concentration of CdCl₂, K₂Cr₂O₇, and NaAsO₂. Biological network study shows that the genes which are differentially expressed are highly connected and having high betweenness and centrality.

Keywords: abiotic stress, biological network, chickpea, microarray

Procedia PDF Downloads 197

Authors: Gholba Niranjan Dilip, Anil Kumar

Abstract:

Conduct of accurate reliable mapping of oil palm plantations and census of individual palm trees is a huge challenge. This study addresses this challenge and developed an optimized solution implemented deep learning techniques on remote sensing data. The oil palm is a very important tropical crop. To improve its productivity and land management, it is imperative to have accurate census over large areas. Since, manual census is costly and prone to approximations, a methodology for automated census using panchromatic images from Cartosat-2, SkySat and World View-3 satellites is demonstrated. It is selected two different study sites in Indonesia. The customized set of training data and ground-truth data are created for this study from Cartosat-2 images. The pre-trained model of Single Shot MultiBox Detector (SSD) Lite MobileNet V2 Convolutional Neural Network (CNN) from the TensorFlow Object Detection API is subjected to transfer learning on this customized dataset. The SSD model is able to generate the bounding boxes for each oil palm and also do the counting of palms with good accuracy on the panchromatic images. The detection yielded an F-Score of 83.16 % on seven different images. The detections are buffered and dissolved to generate polygons demarcating the boundaries of the oil palm plantations. This provided the area under the plantations and also gave maps of their location, thereby completing the automated census, with a fairly high accuracy (≈100%). The trained CNN was found competent enough to detect oil palm crowns from images obtained from multiple satellite sensors and of varying temporal vintage. It helped to estimate the increase in oil palm plantations from 2014 to 2021 in the study area. The study proved that high-resolution panchromatic satellite image can successfully be used to undertake census of oil palm plantations using CNNs.

Keywords: object detection, oil palm tree census, panchromatic images, single shot multibox detector

Procedia PDF Downloads 160

Authors: Alexander N. Pisarchik, Parth Chholak

Abstract:

Intrinsic brain noise was estimated via magneto-encephalograms (MEG) recorded during perception of flickering visual stimuli with frequencies of 6.67 and 8.57 Hz. First, we measured the mean phase difference between the flicker signal and steady-state event-related field (SSERF) in the occipital area where the brain response at the flicker frequencies and their harmonics appeared in the power spectrum. Then, we calculated the probability distribution of the phase fluctuations in the regions of frequency locking and computed its kurtosis. Since kurtosis is a measure of the distribution’s sharpness, we suppose that inverse kurtosis is related to intrinsic brain noise. In our experiments, the kurtosis value varied among subjects from K = 3 to K = 5 for 6.67 Hz and from 2.6 to 4 for 8.57 Hz. The majority of subjects demonstrated leptokurtic kurtosis (K < 3), i.e., the distribution tails approached zero more slowly than Gaussian. In addition, we found a strong correlation between kurtosis and brain complexity measured as the correlation dimension, so that the MEGs of subjects with higher kurtosis exhibited lower complexity. The obtained results are discussed in the framework of nonlinear dynamics and complex network theories. Specifically, in a network of coupled oscillators, phase synchronization is mainly determined by two antagonistic factors, noise, and the coupling strength. While noise worsens phase synchronization, the coupling improves it. If we assume that each neuron and each synapse contribute to brain noise, the larger neuronal network should have stronger noise, and therefore phase synchronization should be worse, that results in smaller kurtosis. The described method for brain noise estimation can be useful for diagnostics of some brain pathologies associated with abnormal brain noise.

Keywords: brain, flickering, magnetoencephalography, MEG, visual perception, perception time

Procedia PDF Downloads 148