Search results for: evolutionary neural network

Authors: Jia Li, Xing Wei, Yuchen Hong, Yang Lu

Abstract:

Detecting vehicle behavior has always been the focus of intelligent transportation, but with the explosive growth of the number of vehicles and the complexity of the road environment, the vehicle behavior videos captured by traditional surveillance have been unable to satisfy the study of vehicle behavior. The traditional method of manually labeling vehicle behavior is too time-consuming and labor-intensive, but the existing object detection and tracking algorithms have poor practicability and low behavioral location detection rate. This paper proposes a vehicle behavior detection algorithm based on the dual-stream convolution network and the multi-dimensional video dynamic detection network. In the videos, the straight-line behavior of the vehicle will default to the background behavior. The Changing lanes, turning and turning around are set as target behaviors. The purpose of this model is to automatically mark the target behavior of the vehicle from the untrimmed videos. First, the target behavior proposals in the long video are extracted through the dual-stream convolution network. The model uses a dual-stream convolutional network to generate a one-dimensional action score waveform, and then extract segments with scores above a given threshold M into preliminary vehicle behavior proposals. Second, the preliminary proposals are pruned and identified using the multi-dimensional video dynamic detection network. Referring to the hierarchical reinforcement learning, the multi-dimensional network includes a Timer module and a Spacer module, where the Timer module mines time information in the video stream and the Spacer module extracts spatial information in the video frame. The Timer and Spacer module are implemented by Long Short-Term Memory (LSTM) and start from an all-zero hidden state. The Timer module uses the Transformer mechanism to extract timing information from the video stream and extract features by linear mapping and other methods. Finally, the model fuses time information and spatial information and obtains the location and category of the behavior through the softmax layer. This paper uses recall and precision to measure the performance of the model. Extensive experiments show that based on the dataset of this paper, the proposed model has obvious advantages compared with the existing state-of-the-art behavior detection algorithms. When the Time Intersection over Union (TIoU) threshold is 0.5, the Average-Precision (MP) reaches 36.3% (the MP of baselines is 21.5%). In summary, this paper proposes a vehicle behavior detection model based on multi-dimensional dynamic detection network. This paper introduces spatial information and temporal information to extract vehicle behaviors in long videos. Experiments show that the proposed algorithm is advanced and accurate in-vehicle timing behavior detection. In the future, the focus will be on simultaneously detecting the timing behavior of multiple vehicles in complex traffic scenes (such as a busy street) while ensuring accuracy.

Keywords: vehicle behavior detection, convolutional neural network, long short-term memory, deep learning

Procedia PDF Downloads 117

Authors: Mario Isaza-Ruget, Claudia C. Colmenares-Mejia, Nancy Yomayusa, Camilo A. González, Andres Cely, Jossie Murcia

Abstract:

Quantification of risks associated with complications development from chronic kidney disease (CKD) through accurate survival models can help with patient management. A retrospective cohort that included patients diagnosed with CKD from a primary care program and followed up between 2013 and 2018 was carried out. Time-dependent and static covariates associated with demographic, clinical, and laboratory factors were included. Deep Learning (DL) survival analyzes were developed for three CKD outcomes: CKD stage progression, >25% decrease in Estimated Glomerular Filtration Rate (eGFR), and Renal Replacement Therapy (RRT). Models were evaluated and compared with Random Survival Forest (RSF) based on concordance index (C-index) metric. 2.143 patients were included. Two models were developed for each outcome, Deep Neural Network (DNN) model reported C-index=0.9867 for CKD stage progression; C-index=0.9905 for reduction in eGFR; C-index=0.9867 for RRT. Regarding the RSF model, C-index=0.6650 was reached for CKD stage progression; decreased eGFR C-index=0.6759; RRT C-index=0.8926. DNN models applied in survival analysis context with considerations of longitudinal covariates at the start of follow-up can predict renal stage progression, a significant decrease in eGFR and RRT. The success of these survival models lies in the appropriate definition of survival times and the analysis of covariates, especially those that vary over time.

Keywords: artificial intelligence, chronic kidney disease, deep neural networks, survival analysis

Procedia PDF Downloads 123

Authors: Yitao Lei, Xingxiang Zhai, Burra Venkata Durga Kumar

Abstract:

As the quantity and complexity of computing in large-scale software systems increase, distributed system computing becomes increasingly important. The distributed system realizes high-performance computing by collaboration between different computing resources. If there are no efficient resource scheduling resources, the abuse of distributed computing may cause resource waste and high costs. However, resource scheduling is usually an NP-hard problem, so we cannot find a general solution. However, some optimization algorithms exist like genetic algorithm, ant colony optimization, etc. The large scale of distributed systems makes this traditional optimization algorithm challenging to work with. Heuristic and machine learning algorithms are usually applied in this situation to ease the computing load. As a result, we do a review of traditional resource scheduling optimization algorithms and try to introduce a deep reinforcement learning method that utilizes the perceptual ability of neural networks and the decision-making ability of reinforcement learning. Using the machine learning method, we try to find important factors that influence the performance of distributed system computing and help the distributed system do an efficient computing resource scheduling. This paper surveys the application of deep reinforcement learning on distributed system computing resource scheduling proposes a deep reinforcement learning method that uses a recurrent neural network to optimize the resource scheduling, and proposes the challenges and improvement directions for DRL-based resource scheduling algorithms.

Keywords: resource scheduling, deep reinforcement learning, distributed system, artificial intelligence

Procedia PDF Downloads 96

Authors: Juzhong Tan, William Kerr

Abstract:

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

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

Procedia PDF Downloads 221

Authors: Tianyu Cao, KIRA (Ruizhi Zhao)

Abstract:

The purpose of our research is to study the possibility of people with physical disabilities manipulating mechanical prostheses through brain-computer interface (BCI) technology. The brain-machine interface (BCI) of the neural prosthesis records signals from neurons and uses mathematical modeling to decode them, converting desired movements into body movements. In order to improve the patient's neural control, the prosthesis is given a natural feeling. It records data from sensitive areas from the body to the prosthetic limb and encodes signals in the form of electrical stimulation to the brain. In our research, the brain-computer interface (BCI) is a bridge connecting patients’ cognition and the real world, allowing information to interact with each other. The efficient work between the two is achieved through external devices. The flow of information is controlled by BCI’s ability to record neuronal signals and decode signals, which are converted into device control. In this way, we could encode information and then send it to the brain through electrical stimulation, which has significant medical application.

Keywords: biomedical engineering, brain-computer interface, prosthesis, neural control

Procedia PDF Downloads 164

Authors: Abdelhamid Louliej, Younes Jabrane

Abstract:

Satellites are becoming indispensable assets to fifth-generation (5G) new radio architecture, complementing wireless and terrestrial communication links. The combination of satellites and 5G architecture allows consumers to access all next-generation services anytime, anywhere, including scenarios, like traveling to remote areas (without coverage). Nevertheless, this solution faces several challenges, such as a significant propagation delay, Doppler frequency shift, and high Peak-to-Average Power Ratio (PAPR), causing signal distortion due to the non-linear saturation of the High-Power Amplifier (HPA). To compensate for HPA non-linearity in 5G satellite transmission, an efficient pre-distorter scheme using Neural Networks (NN) is proposed. To assess the proposed NN pre-distorter, two types of HPA were investigated: Travelling Wave Tube Amplifier (TWTA) and Solid-State Power Amplifier (SSPA). The results show that the NN pre-distorter design presents EVM improvement by 95.26%. NMSE and ACPR were reduced by -43,66 dB and 24.56 dBm, respectively. Moreover, the system suffers no degradation of the Bit Error Rate (BER) for TWTA and SSPA amplifiers.

Keywords: satellites, 5G, neural networks, HPA, TWTA, SSPA, EVM, NMSE, ACPR

Procedia PDF Downloads 79

Authors: Rowanda Daoud Ahmed, Mansoor Abdulhak, Muhammad Azeem Afzal, Sezer Filiz, Usama Ahmad Mughal

Abstract:

This paper discusses important aspects of AI in the healthcare domain. The increase of data in healthcare both in size and complexity, opens more room for artificial intelligence applications. Our focus is to review the main AI methods within the scope of the health care domain. The results of the review show that recommendations for diagnosis and recommendations for treatment, patent engagement, and administrative tasks are the key applications of AI in healthcare. Understanding the potential of AI methods in the domain of healthcare would benefit healthcare practitioners and will improve patient outcomes.

Keywords: AI in healthcare, technologies of AI, neural network, future of AI in healthcare

Procedia PDF Downloads 103

Authors: Henok Mezemr Besfat, Haftom Gebreslassie Gebregwergs

Abstract:

wireless network is one of the most important mediums of transmission of information from one device to another devices. Wireless communication has a broad range of applications, including mobile communications through cell phones and satellites, Internet of Things (IoT) connecting several devices, wireless sensor networks for traffic management and environmental monitoring, satellite communication for weather forecasting and TV without requiring any cable or wire or other electronic conductors, by using electromagnetic waves like IR, RF, satellite, etc. This paper summarizes different wireless network technologies, applications of different wireless technologies and different types of wireless networks. Generally, wireless technology will further enhance operations and experiences across sectors with continued innovation. This paper suggests different strategies that can improve wireless networks and technologies.

Keywords: wireless senser, wireless technology, wireless network, internet of things

Procedia PDF Downloads 36

Authors: Amit R. Bhende, G. K. Awari

Abstract:

Remaining useful life (RUL) prediction is one of key technologies to realize prognostics and health management that is being widely applied in many industrial systems to ensure high system availability over their life cycles. The present work proposes a data-driven method of RUL prediction based on multiple health state assessment for rolling element bearings. Bearing degradation data at three different conditions from run to failure is used. A RUL prediction model is separately built in each condition. Feed forward back propagation neural network models are developed for prediction modeling.

Keywords: bearing degradation data, remaining useful life (RUL), back propagation, prognosis

Procedia PDF Downloads 426

Authors: Zafer Yüce, Paşa Yayla, Alev Taşkın

Abstract:

There are heaps of analytical methods to estimate the crack growth life of a component. Soft computing methods have an increasing trend in predicting fatigue life. Their ability to build complex relationships and capability to handle huge amounts of data are motivating researchers and industry professionals to employ them for challenging problems. This study focuses on soft computing methods, especially random forest regressors and artificial neural networks with hyperparameter optimization algorithms such as grid search and random grid search, to estimate the crack growth life of an aircraft wing splice joint under variable amplitude loading. TensorFlow and Scikit-learn libraries of Python are used to build the machine learning models for this study. The material considered in this work is 7050-T7451 aluminum, which is commonly preferred as a structural element in the aerospace industry, and regarding the crack type; corner crack is used. A finite element model is built for the joint to calculate fastener loads and stresses on the structure. Since finite element model results are validated with analytical calculations, findings of the finite element model are fed to AFGROW software to calculate analytical crack growth lives. Based on Fighter Aircraft Loading Standard for Fatigue (FALSTAFF), 90 unique fatigue loading spectra are developed for various load levels, and then, these spectrums are utilized as inputs to the artificial neural network and random forest regression models for predicting crack growth life. Finally, the crack growth life predictions of the machine learning models are compared with analytical calculations. According to the findings, a good correlation is observed between analytical and predicted crack growth lives.

Keywords: aircraft, fatigue, joint, life, optimization, prediction.

Procedia PDF Downloads 157

Authors: Safak Baykal

Abstract:

The purpose of this study is analyzing the relationship between self leadership and social capital of people with using Social Network Analysis. In this study, two aspects of social capital will be focused: bonding, homophilous social capital (BoSC) which implies better, strong, dense or closed network ties, and bridging, heterophilous social capital (BrSC) which implies weak ties, bridging the structural holes. The other concept of the study is Trust (Tr), namely interpersonal trust, willingness to ascribe good intentions to and have confidence in the words and actions of other people. In this study, the sample group, 61 people, was selected from a private firm from the defense industry. The relation between BoSC/BrSC and Tr is shown by using Social Network Analysis (SNA) and statistical analysis with Likert type-questionnaire. The results of the analysis show the Cronbach’s alpha value is 0.73 and social capital values (BoSC/BrSC) is highly correlated with Tr values of the people.

Keywords: bonding social capital, bridging social capital, trust, social network analysis (SNA)

Procedia PDF Downloads 516

Authors: Mary Anne Roa

Abstract:

Congestion in network occurs due to exceed in aggregate demand as compared to the accessible capacity of the resources. Network congestion will increase as network speed increases and new effective congestion control methods are needed, especially for today’s very high speed networks. To address this undeniably global issue, the study focuses on the development of a fuzzy-based congestion control model concerned with allocating the resources of a computer network such that the system can operate at an adequate performance level when the demand exceeds or is near the capacity of the resources. Fuzzy logic based models have proven capable of accurately representing a wide variety of processes. The model built is based on bandwidth, the aggregate incoming traffic and the waiting time. The theoretical analysis and simulation results show that the proposed algorithm provides not only good utilization but also low packet loss.

Keywords: congestion control, queue management, computer networks, fuzzy logic

Procedia PDF Downloads 382

Authors: Alexander Hempfing

Abstract:

The desire to understand business cycle fluctuations, trade interdependencies and co-movement has a long tradition in economic thinking. From input-output economics to business cycle theory, researchers aimed to find appropriate answers from an empirical as well as a theoretical perspective. This paper empirically analyses how the production structure of the global economy and several states developed over time, what their distributional properties are and if there are network specific metrics that allow identifying structurally important nodes, on a global, national and sectoral scale. For this, the World Input-Output Database was used, and different statistical methods were applied. Empirical evidence is provided that the importance of the Eastern hemisphere in the global production network has increased significantly between 2000 and 2014. Moreover, it was possible to show that the sectoral eigenvector centrality indices on a global level are power-law distributed, providing evidence that specific national sectors exist which are more critical to the world economy than others while serving as a hub within the global production network. However, further findings suggest, that global production cannot be characterized as a scale-free network.

Keywords: economic integration, industrial organization, input-output economics, network economics, production networks

Procedia PDF Downloads 263

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

Abstract:

Typically, virtual communities exhibit the well-known phenomenon of participation inequality, which means that only a small percentage of users is responsible of the majority of contributions. However, the sustainability of the community requires that the group of active users must be continuously nurtured with new users that gain expertise through a participation process. This paper analyzes the time evolution of Open Source Software (OSS) communities, considering users that join/abandon the community over time and several topological properties of the network when modeled as a social network. More specifically, the paper analyzes the role of those users rejoining the community and their influence in the global characteristics of the network.

Keywords: open source communities, social network Analysis, time series, virtual communities

Procedia PDF Downloads 515

Authors: Younghyun Jeon, Seungjoo Maeng

Abstract:

In the Ad-hoc network, the great interests regarding MIMO scheme leads to their combination, which is also utilized into its applicable network. We manage the field of the problem into Reverse-link MIMO Ad-hoc Network (RMAN) and propose the methodology to maximize the data rate with its power consumption using Node-Cooperative beamforming technique. Based on the result of mathematical optimization formulation, we design the algorithm to construct optimal orthogonal weight vector according to channel feedback and control its transmission power according to QoS-pricing value level. In simulation results, we show the validity of the proposed mathematical optimization result and algorithm which mean that the sum-rate of each link is converged into some point.

Keywords: ad-hoc network, MIMO, cooperative beamforming, transmit power

Procedia PDF Downloads 386

Authors: Benoît Bucher, Hika Kuroshima, Kazuo Fujita

Abstract:

Although altruism is commonly observed in humans and is considered one of the most important factors in the survival of our species, its cognitive mechanisms and evolutionary roots are yet to be explained. This study is based on the previous findings that bonobos (Pan Paniscus) preferred to share a limited amount of food with others regardless of their relationships with the others. Findings such as this suggest that humans’ propensity for altruistic food-sharing may be shared among apes and may have evolved much longer ago than previously considered. We thus adapted the previous experimental design using tufted capuchins (Cebus apella), New World monkeys separating from humans about 40 million years ago. In order to achieve this, 12 pairs of capuchins (consisting of a benefactors and a partner) were tested in a row of two adjacent cages separated by a swinging door locked by a key (Fig.1). We observed whether the monkeys in possession of food (the benefactors) would allow their partner to enter their cage by unlocking the door between them. Results showed that the monkeys clearly preferred to monopolize the food for themselves, even though they in a few cases unlocked the door after eating the preferred food. This suggests that this species, which has been shown to be sensitive to the others’ welfare, would not actively share food at a cost of their own. Although further studies are needed, our results suggest the existence of significant differences in the evolutionary development of the pro social tendencies between bonobos and capuchin monkeys.

Keywords: altruism, capuchin monkeys, food sharing, pro social behaviors

Procedia PDF Downloads 440

Authors: Eu Tteum Ha, Seyoung Kim, Jeongmin Kim, Kwang Ryel Ryu

Abstract:

The smart home environment backed up by IoT (internet of things) technologies enables intelligent services based on the awareness of the situation a user is currently in. One of the convenient sensors for recognizing the situations within a home is the smart meter that can monitor the status of each electrical appliance in real time. This paper aims at learning a Bayesian network that models the causal relationship between the user situations and the status of the electrical appliances. Using such a network, we can infer the current situation based on the observed status of the appliances. However, learning the conditional probability tables (CPTs) of the network requires many training examples that cannot be obtained unless the user situations are closely monitored by any means. This paper proposes a method for learning the CPT entries of the network relying only on the user feedbacks generated occasionally. In our case study with a robot vacuum cleaner, the feedback comes in whenever the user gives an order to the robot adversely from its preprogrammed setting. Given a network with randomly initialized CPT entries, our proposed method uses this feedback information to adjust relevant CPT entries in the direction of increasing the probability of recognizing the desired situations. Simulation experiments show that our method can rapidly improve the recognition performance of the Bayesian network using a relatively small number of feedbacks.

Keywords: Bayesian network, IoT, learning, situation -awareness, smart home

Procedia PDF Downloads 512

Authors: Ujjwall Sai Sunder Uppuluri

Abstract:

Group theory is a powerful tool that researchers can use to provide a structural foundation for their Agent Based Models. These Agent Based models are argued by this paper to be the future of the Social Science Disciplines. More specifically, researchers can use them to apply evolutionary theory to the study of complex social systems. This paper illustrates one such example of how theoretically an Agent Based Model can be formulated from the application of Group Theory, Systems Dynamics, and Evolutionary Biology to analyze the strategies pursued by states to mitigate risk and maximize usage of resources to achieve the objective of economic growth. This example can be applied to other social phenomena and this makes group theory so useful to the analysis of complex systems, because the theory provides the mathematical formulaic proof for validating the complex system models that researchers build and this will be discussed by the paper. The aim of this research, is to also provide researchers with a framework that can be used to model political entities such as states on a 3-dimensional plane. The x-axis representing resources (tangible and intangible) available to them, y the risks, and z the objective. There also exist other states with different constraints pursuing different strategies to climb the mountain. This mountain’s environment is made up of risks the state faces and resource endowments. This mountain is also layered in the sense that it has multiple peaks that must be overcome to reach the tallest peak. A state that sticks to a single strategy or pursues a strategy that is not conducive to the climbing of that specific peak it has reached is not able to continue advancement. To overcome the obstacle in the state’s path, it must innovate. Based on the definition of a group, we can categorize each state as being its own group. Each state is a closed system, one which is made up of micro level agents who have their own vectors and pursue strategies (actions) to achieve some sub objectives. The state also has an identity, the inverse being anarchy and/or inaction. Finally, the agents making up a state interact with each other through competition and collaboration to mitigate risks and achieve sub objectives that fall within the primary objective. Thus, researchers can categorize the state as an organism that reflects the sum of the output of the interactions pursued by agents at the micro level. When states compete, they employ a strategy and that state which has the better strategy (reflected by the strategies pursued by her parts) is able to out-compete her counterpart to acquire some resource, mitigate some risk or fulfil some objective. This paper will attempt to illustrate how group theory combined with evolutionary theory and systems dynamics can allow researchers to model the long run development, evolution, and growth of political entities through the use of a bottom up approach.

Keywords: complex systems, evolutionary theory, group theory, international political economy

Procedia PDF Downloads 121

Authors: Oleg Vlasovets, Fabian Schaipp, Christian L. Mueller

Abstract:

we conduct a network analysis and predict the sex of 1000 participants based on ”connectome” - pairwise Pearson’s correlation across 436 brain parcels. We solve the non-smooth convex optimization problem, known under the name of Graphical Lasso, where the solution includes a low-rank component. With this solution and machine learning model for a sex prediction, we explain the brain parcels-sex connectivity patterns.

Keywords: network analysis, neuroscience, machine learning, optimization

Procedia PDF Downloads 135

Authors: Yilin Liao, Hewen Li, Paula McConvey

Abstract:

Speed skaters often face a risk of concussion when they fall on the ice floor and impact crash mats during practices and competitive races. Several variables, including those related to the skater, the crash mat, and the impact position (body side/head/feet impact), are believed to influence the severity of the skater's concussion. While computer simulation modeling can be employed to analyze these accidents, the simulation process is time-consuming and does not provide rapid information for coaches and teams to assess the skater's injury risk in competitive events. This research paper promotes the exploration of the feasibility of using AI techniques for evaluating skater’s potential concussion severity, and to develop a fast concussion prediction tool using artificial neural networks to reduce the risk of treatment delays for injured skaters. The primary data is collected through virtual tests and physical experiments designed to simulate skater-mat impact. It is then analyzed to identify patterns and correlations; finally, it is used to train and fine-tune the artificial neural networks for accurate prediction. The development of the prediction tool by employing machine learning strategies contributes to the application of AI methods in sports science and has theoretical involvements for using AI techniques in predicting and preventing sports-related injuries.

Keywords: artificial neural networks, concussion, machine learning, impact, speed skater

Procedia PDF Downloads 85

Authors: Sifat Rahman Ahona, Rifat Tasnim, Naima Hassan

Abstract:

Data aggregation is a helpful technique for reducing the data communication overhead in wireless sensor network. One of the important tasks of data aggregation is positioning of the aggregator points. There are a lot of works done on data aggregation. But, efficient positioning of the aggregators points is not focused so much. In this paper, authors are focusing on the positioning or the placement of the aggregation points in wireless sensor network. Authors proposed an algorithm to select the aggregators positions for a scenario where aggregator nodes are more powerful than sensor nodes.

Keywords: aggregation point, data communication, data aggregation, wireless sensor network

Procedia PDF Downloads 146

Authors: Richard Draper

Abstract:

A supercomputer can be constructed from identical building blocks which are small parallel processors connected by a network referred to as the local network. The routers have unused ports which are used to interconnect the building blocks. These connections are referred to as the global network. The address space has a global and a local component (g, l). The conventional way to connect the building blocks is to connect (g, l) to (g’,l). If there are K blocks, this requires K global ports in each router. If a block is of size M, the result is a machine with KM routers having diameter two. To increase the size of the machine to 2K blocks, each router connects to only half of the other blocks. The result is a larger machine but also one with greater diameter. This is a crude description of how the network of the CRAY XC® is designed. In this paper, a family of interconnection networks using routers with K global and M local ports is defined. Coordinates are (c,d, p) and the global connections are (c,d,p)↔(c’,p,d) which swaps p and d. The network is denoted D3(K,M) and is called a Swapped Dragonfly. D3(K,M) has KM2 routers and has diameter three, regardless of the size of K. To produce a network of size KM2 conventionally, diameter would be an increasing function of K. The family of Swapped Dragonflies has other desirable properties: 1) D3(K,M) scales linearly in K and quadratically in M. 2) If L < K, D3(K,M) contains many copies of D3(L,M). 3) If L < M, D3(K,M) contains many copies of D3(K,L). 4) D3(K,M) can perform an all-to-all exchange in KM2+KM time which is only slightly more than the time to do a one-to-all. This paper makes several contributions. It is the first time that a swap has been used to define a linearly scalable family of networks. Structural properties of this new family of networks are thoroughly examined. A synchronizing packet header is introduced. It specifies the path to be followed and it makes it possible to define highly parallel communication algorithm on the network. Among these is an all-to-all exchange in time KM2+KM. To demonstrate the effectiveness of the swap properties of the network of the CRAY XC® and D3(K,16) are compared.

Keywords: all-to-all exchange, CRAY XC®, Dragonfly, interconnection network, packet switching, swapped network, topology

Procedia PDF Downloads 111

Authors: Fei Gao, Xinyang Dong, Yisu Ge, Shufang Lu, Libo Weng

Abstract:

Vehicle color recognition plays an important role in traffic accident investigation. However, due to the influence of illumination, weather, and noise, vehicle color recognition still faces challenges. In this paper, a vehicle color classification network based on spatial cluster loss and channel attention mechanism (SCNet) is proposed for vehicle color recognition. A channel attention module is applied to extract the features of vehicle color representative regions and reduce the weight of nonrepresentative color regions in the channel. The proposed loss function, called spatial clustering loss (SC-loss), consists of two channel-specific components, such as a concentration component and a diversity component. The concentration component forces all feature channels belonging to the same class to be concentrated through the channel cluster. The diversity components impose additional constraints on the channels through the mean distance coefficient, making them mutually exclusive in spatial dimensions. In the comparison experiments, the proposed method can achieve state-of-the-art performance on the public datasets, VCD, and VeRi, which are 96.1% and 96.2%, respectively. In addition, the ablation experiment further proves that SC-loss can effectively improve the accuracy of vehicle color recognition.

Keywords: feature extraction, convolutional neural networks, intelligent transportation, vehicle color recognition

Procedia PDF Downloads 163

Authors: Xiao-Yin Liu, Liang-Xue Zhou

Abstract:

Traumatic brain injury (TBI), as a kind of nerve trauma caused by an external force, affects people all over the world and is a global public health problem. Although there are various clinical treatments for brain injury, including surgery, drug therapy, and rehabilitation therapy, the therapeutic effect is very limited. To improve the therapeutic effect of TBI, scaffolds combined with exosomes are a promising but challenging method for TBI repair. In this study, we examined whether a novel 3D-printed collagen/chitosan scaffold/exosomes derived from neural stem cells (NSCs) pretreated with insulin growth factor-1 (IGF-I) scaffolds (3D-CC-INExos) could be used to improve TBI repair and functional recovery after TBI. Our results showed that composite scaffolds of collagen-, chitosan- and exosomes derived from NSCs pretreated with IGF-I (INExos) could continuously release the exosomes for two weeks. In the rat TBI model, 3D-CC-INExos scaffold transplantation significantly improved motor and cognitive function after TBI, as assessed by the Morris water maze test and modified neurological severity scores. In addition, immunofluorescence staining and transmission electron microscopy showed that the recovery of damaged nerve tissue in the injured area was significantly improved by 3D-CC-INExos implantation. In conclusion, our data suggest that 3D-CC-INExos might provide a potential strategy for the treatment of TBI and lay a solid foundation for clinical translation.

Keywords: traumatic brain injury, exosomes, insulin growth factor-1, neural stem cells, collagen, chitosan, 3D printing, neural regeneration, angiogenesis, functional recovery

Procedia PDF Downloads 66

Authors: Rao Fu, Jie Chen, Shutian Zeng, Yiping Zhuang, Agus Sudjianto

Abstract:

Generative Adversarial Net (GAN) has proved to be a powerful machine learning tool in image data analysis and generation. In this paper, we propose to use Conditional Generative Adversarial Net (CGAN) to learn and simulate time series data. The conditions include both categorical and continuous variables with different auxiliary information. Our simulation studies show that CGAN has the capability to learn different types of normal and heavy-tailed distributions, as well as dependent structures of different time series. It also has the capability to generate conditional predictive distributions consistent with training data distributions. We also provide an in-depth discussion on the rationale behind GAN and the neural networks as hierarchical splines to establish a clear connection with existing statistical methods of distribution generation. In practice, CGAN has a wide range of applications in market risk and counterparty risk analysis: it can be applied to learn historical data and generate scenarios for the calculation of Value-at-Risk (VaR) and Expected Shortfall (ES), and it can also predict the movement of the market risk factors. We present a real data analysis including a backtesting to demonstrate that CGAN can outperform Historical Simulation (HS), a popular method in market risk analysis to calculate VaR. CGAN can also be applied in economic time series modeling and forecasting. In this regard, we have included an example of hypothetical shock analysis for economic models and the generation of potential CCAR scenarios by CGAN at the end of the paper.

Keywords: conditional generative adversarial net, market and credit risk management, neural network, time series

Procedia PDF Downloads 128

Authors: Stefan Papastefanou

Abstract:

Artificial intelligence (AI) is an interdisciplinary field of computer science with the aim of creating intelligent machine behavior. Early approaches to AI have been configured to operate in very constrained environments where the behavior of the AI system was previously determined by formal rules. Knowledge was presented as a set of rules that allowed the AI system to determine the results for specific problems; as a structure of if-else rules that could be traversed to find a solution to a particular problem or question. However, such rule-based systems typically have not been able to generalize beyond the knowledge provided. All over the world and especially in IT-heavy industries such as the United States, the European Union, Singapore, and China, machine learning has developed to be an immense asset, and its applications are becoming more and more significant. It has to be examined how such products of machine learning models can and should be protected by IP law and for the purpose of this paper patent law specifically, since it is the IP law regime closest to technical inventions and computing methods in technical applications. Genetic breeding models are currently less popular than recursive neural network method and deep learning, but this approach can be more easily described by referring to the evolution of natural organisms, and with increasing computational power; the genetic breeding method as a subset of the evolutionary algorithms models is expected to be regaining popularity. The research method focuses on patentability (according to the world’s most significant patent law regimes such as China, Singapore, the European Union, and the United States) of AI inventions and machine learning. Questions of the technical nature of the problem to be solved, the inventive step as such, and the question of the state of the art and the associated obviousness of the solution arise in the current patenting processes. Most importantly, and the key focus of this paper is the problem of patenting inventions that themselves are developed through machine learning. The inventor of a patent application must be a natural person or a group of persons according to the current legal situation in most patent law regimes. In order to be considered an 'inventor', a person must actually have developed part of the inventive concept. The mere application of machine learning or an AI algorithm to a particular problem should not be construed as the algorithm that contributes to a part of the inventive concept. However, when machine learning or the AI algorithm has contributed to a part of the inventive concept, there is currently a lack of clarity regarding the ownership of artificially created inventions. Since not only all European patent law regimes but also the Chinese and Singaporean patent law approaches include identical terms, this paper ultimately offers a comparative analysis of the most relevant patent law regimes.

Keywords: algorithms, inventor, genetic breeding models, machine learning, patentability

Procedia PDF Downloads 101

Authors: S. Sabatino, V. Calderaro, V. Galdi, G. Graber, L. Ippolito

Abstract:

Climate change is a rapidly growing global threat caused mainly by increased emissions of carbon dioxide (CO₂) into the atmosphere. These emissions come from multiple sources, including industry, power generation, and the transport sector. The need to tackle climate change and reduce CO₂ emissions is indisputable. A crucial solution to achieving decarbonization in the transport sector is the adoption of electric vehicles (EVs). These vehicles use lithium (Li-Ion) batteries as an energy source, making them extremely efficient and with low direct emissions. However, Li-Ion batteries are not without problems, including the risk of overheating and performance degradation. To ensure its safety and longevity, it is essential to use a battery management system (BMS). The BMS constantly monitors battery status, adjusts temperature and cell balance, ensuring optimal performance and preventing dangerous situations. From the monitoring carried out, it is also able to optimally manage the battery to increase its life. Among the parameters monitored by the BMS, the main ones are State of Charge (SoC), State of Health (SoH), and State of Power (SoP). The evaluation of these parameters can be carried out in two ways: offline, using benchtop batteries tested in the laboratory, or online, using batteries installed in moving vehicles. Online estimation is the preferred approach, as it relies on capturing real-time data from batteries while operating in real-life situations, such as in everyday EV use. Actual battery usage conditions are highly variable. Moving vehicles are exposed to a wide range of factors, including temperature variations, different driving styles, and complex charge/discharge cycles. This variability is difficult to replicate in a controlled laboratory environment and can greatly affect performance and battery life. Online estimation captures this variety of conditions, providing a more accurate assessment of battery behavior in real-world situations. In this article, a hybrid approach based on a neural network and a statistical method for real-time estimation of SoC, SoH, and SoP parameters of interest is proposed. These parameters are estimated from the analysis of a one-day driving profile of an electric vehicle, assumed to be divided into the following four phases: (i) Partial discharge (SoC 100% - SoC 50%), (ii) Partial discharge (SoC 50% - SoC 80%), (iii) Deep Discharge (SoC 80% - SoC 30%) (iv) Full charge (SoC 30% - SoC 100%). The neural network predicts the values of ohmic resistance and incremental capacity, while the statistical method is used to estimate the parameters of interest. This reduces the complexity of the model and improves its prediction accuracy. The effectiveness of the proposed model is evaluated by analyzing its performance in terms of square mean error (RMSE) and percentage error (MAPE) and comparing it with the reference method found in the literature.

Keywords: electric vehicle, Li-Ion battery, BMS, state-of-charge, state-of-health, state-of-power, artificial neural networks

Procedia PDF Downloads 58

Authors: Monica Sharma, Harjit Pal Singh, Jasbinder Singh, Manju Bala

Abstract:

In Voice over Internet Protocol (VoIP) system, the speech signal is degraded when passed through the network layers. The speech signal is processed through the best effort policy based IP network, which leads to the network degradations including delay, packet loss and jitter. The packet loss is the major issue of the degradation in the VoIP signal quality; even a single lost packet may generate audible distortion in the decoded speech signal. In addition to these network degradations, the quality of the speech signal is also affected by the environmental noises and coder distortions. The signal quality of the VoIP system is improved through the interleaving technique. The performance of the system is evaluated for various types of noises at different network conditions. The performance of the enhanced VoIP signal is evaluated using perceptual evaluation of speech quality (PESQ) measurement for narrow band signal.

Keywords: VoIP, interleaving, packet loss, packet size, background noise

Procedia PDF Downloads 468

Authors: Benjamin Bobbia, Matthias Picard

Abstract:

In recent years active learning methodologies based on the representativity of the data seems more promising to limit overfitting. The presented query methodology for regression using the Wasserstein distance measuring the representativity of our labelled dataset compared to the global distribution. In this work a crucial use of GroupSort Neural Networks is made therewith to draw a double advantage. The Wasserstein distance can be exactly expressed in terms of such neural networks. Moreover, one can provide explicit bounds for their size and depth together with rates of convergence. However, heterogeneity of the dataset is also considered by weighting the Wasserstein distance with the error of approximation at the previous step of active learning. Such an approach leads to a reduction of overfitting and high prediction performance after few steps of query. After having detailed the methodology and algorithm, an empirical study is presented in order to investigate the range of our hyperparameters. The performances of this method are compared, in terms of numbers of query needed, with other classical and recent query methods on several UCI datasets.

Keywords: active learning, Lipschitz regularization, neural networks, optimal transport, regression

Procedia PDF Downloads 71

Authors: Silumpa Suboonsan, Suwat Pattaramalai

Abstract:

In this paper, the performance of internet usage by using Vertical Handover (VHO) between cellular network and wireless local area network (WLAN) on Silom line Bangkok Mass Transit System (BTS) is evaluated. In the evaluation model, there is the WLAN on every BTS station and there are cellular base stations along the BTS path. The maximum data rates for cellular network are 7.2, 14.4, 42, and 100Mbps and for WLAN are 54, 150, and 300Mbps. The simulation are based on users using internet, watching VDOs and browsing web pages, on the BTS train from first station to the last station (full time usage) and on the BTS train for traveling some number of stations (random time). The results shows that VHO system has throughput a lot more than using only cellular network when the data rate of WLAN is more than one of cellular network. Lastly, the number of watching HD VDO and Full HD VDO is higher on VHO system on both regular time and rush hour of BTS travelling.

Keywords: vertical handover, WLAN, cellular, silom line BTS

Procedia PDF Downloads 469