Pub Date : 2021-09-13DOI: 10.1109/IICAIET51634.2021.9573784
M. K. Tan, Mohd. Riezman Ladillah, H. S. Chuo, Kit Guan Lim, R. Chin, K. Teo
Traffic lights are the signaling devices located at a road intersection for granting right-of-way movement to road users. Thus, optimization of traffic signalization is essential to improve road service as it is the cost-effective way. Commonly, the signal optimization aims to minimize the average travel delay by manipulating the green signal timing. Besides to optimize the signal timing, the local intersection controller needs to collaborate with neighboring intersection controllers for minimizing the average delay for whole network as the congestion will be propagated to the downstream intersection. However, the current fixed-time signal controller is inadequate to manage the high growing demands of traffic as it is tuned offline using the nominal traffic flow data. Thus, this work aims to explore the potential of using Particle Swarm Optimization (PSO) to optimize the traffic signal timing for the traffic network. The proposed algorithm is texted using a benchmarked 1x2 traffic model and its performances are compared with the classical Genetic Algorithm (GA). The simulated results show the proposed PSO has improved the performances in minimizing average travel delay by 3.94 %.
{"title":"Optimization of Signalized Traffic Network using Swarm Intelligence","authors":"M. K. Tan, Mohd. Riezman Ladillah, H. S. Chuo, Kit Guan Lim, R. Chin, K. Teo","doi":"10.1109/IICAIET51634.2021.9573784","DOIUrl":"https://doi.org/10.1109/IICAIET51634.2021.9573784","url":null,"abstract":"Traffic lights are the signaling devices located at a road intersection for granting right-of-way movement to road users. Thus, optimization of traffic signalization is essential to improve road service as it is the cost-effective way. Commonly, the signal optimization aims to minimize the average travel delay by manipulating the green signal timing. Besides to optimize the signal timing, the local intersection controller needs to collaborate with neighboring intersection controllers for minimizing the average delay for whole network as the congestion will be propagated to the downstream intersection. However, the current fixed-time signal controller is inadequate to manage the high growing demands of traffic as it is tuned offline using the nominal traffic flow data. Thus, this work aims to explore the potential of using Particle Swarm Optimization (PSO) to optimize the traffic signal timing for the traffic network. The proposed algorithm is texted using a benchmarked 1x2 traffic model and its performances are compared with the classical Genetic Algorithm (GA). The simulated results show the proposed PSO has improved the performances in minimizing average travel delay by 3.94 %.","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"335 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123394974","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-09-13DOI: 10.1109/iicaiet51634.2021.9573577
{"title":"[Copyright notice]","authors":"","doi":"10.1109/iicaiet51634.2021.9573577","DOIUrl":"https://doi.org/10.1109/iicaiet51634.2021.9573577","url":null,"abstract":"","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121342217","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-09-13DOI: 10.1109/IICAIET51634.2021.9573655
Tsuyoshi Inoue, Kosuke Uehata, Chihiro Tomoda
We have developed a sit-to-stand assist system that predicts the movement speed and drives at that speed. The assistance system predicts the speed of sit-to-stand movement based on multiple regression analysis. The measurement of trunk angle and lower limb electromyogram (EMG) were used as the explanatory variables for the multiple regression analysis. To verify the effectiveness of the developed system, we conducted evaluation experiments on two participants. The evaluation was performed based on the difference of amount of system support between the conventional constant speed control and the proposed predictive speed control. The evaluation results show that the predictive speed control resulted in more support, confirming the effectiveness of the system control that predicted the sit-to-stand speed.
{"title":"Evaluation of Assistance System to Predict Sit-to-stand Speed using Trunk Angle and Lower Limb EMG","authors":"Tsuyoshi Inoue, Kosuke Uehata, Chihiro Tomoda","doi":"10.1109/IICAIET51634.2021.9573655","DOIUrl":"https://doi.org/10.1109/IICAIET51634.2021.9573655","url":null,"abstract":"We have developed a sit-to-stand assist system that predicts the movement speed and drives at that speed. The assistance system predicts the speed of sit-to-stand movement based on multiple regression analysis. The measurement of trunk angle and lower limb electromyogram (EMG) were used as the explanatory variables for the multiple regression analysis. To verify the effectiveness of the developed system, we conducted evaluation experiments on two participants. The evaluation was performed based on the difference of amount of system support between the conventional constant speed control and the proposed predictive speed control. The evaluation results show that the predictive speed control resulted in more support, confirming the effectiveness of the system control that predicted the sit-to-stand speed.","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"221 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116098366","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-09-13DOI: 10.1109/IICAIET51634.2021.9573934
Kit Guan Lim, Chii Soon Huong, M. K. Tan, C. F. Liau, Min Yang, K. Teo
This paper presents the deep learning model by Convolution Neural Network (CNN) in training the crop disease classifier via image classification. A camera will be equipped and applied in artificial intelligent drone to operate as a crop monitoring system used for agriculture. Agriculture productivity is a key component of country economy. Crop diseases can lead to a drop in the quality and quantity of agricultural products. Famers are facing problems to detect the crop diseases accurately in huge region of crops. Therefore, CNN based method for crop disease detection is proposed. Dataset contains of 16,257 color images which has a total of categories have been fed into the model, out of which 10 categories are of diseased crop leaves. The CNN model contains 7 convolution layers with the number of filters 32, 64, two layers with 128 filters, three layers with 256 filters and filter size $3times 3$ is the proposed approach to perform crop disease classification, with the best testing accuracy of 99.02%. The crops are classified correctly using the suggested CNN design. The suggested CNN design is validated and evaluated which achieves accuracy of 99.02%, 0.98% error, 99% recall, 99% precision and 0.99 score of F-measure. In this paper, achievement of the proposed CNN model is reaching a promising result and simulated successfully in classifying the crop disease.
{"title":"Optimization of Crop Disease Classification using Convolution Neural Network","authors":"Kit Guan Lim, Chii Soon Huong, M. K. Tan, C. F. Liau, Min Yang, K. Teo","doi":"10.1109/IICAIET51634.2021.9573934","DOIUrl":"https://doi.org/10.1109/IICAIET51634.2021.9573934","url":null,"abstract":"This paper presents the deep learning model by Convolution Neural Network (CNN) in training the crop disease classifier via image classification. A camera will be equipped and applied in artificial intelligent drone to operate as a crop monitoring system used for agriculture. Agriculture productivity is a key component of country economy. Crop diseases can lead to a drop in the quality and quantity of agricultural products. Famers are facing problems to detect the crop diseases accurately in huge region of crops. Therefore, CNN based method for crop disease detection is proposed. Dataset contains of 16,257 color images which has a total of categories have been fed into the model, out of which 10 categories are of diseased crop leaves. The CNN model contains 7 convolution layers with the number of filters 32, 64, two layers with 128 filters, three layers with 256 filters and filter size $3times 3$ is the proposed approach to perform crop disease classification, with the best testing accuracy of 99.02%. The crops are classified correctly using the suggested CNN design. The suggested CNN design is validated and evaluated which achieves accuracy of 99.02%, 0.98% error, 99% recall, 99% precision and 0.99 score of F-measure. In this paper, achievement of the proposed CNN model is reaching a promising result and simulated successfully in classifying the crop disease.","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121191862","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-09-13DOI: 10.1109/IICAIET51634.2021.9573657
M. Teow
This study explores a convolutional autoencoder for image denoising with a proposed compositional subspace method. This modeling approach presents a structural and rigorous mathematical abstraction to understand a convolutional autoencoder's functional computation layers. The theoretical basis is that the best way to model a complex learning function is by using a composition of simple functions to form a multilayer successive cascaded function for complex representation. The proposed method has experimented with the Fashion-MNIST dataset. Experimental results are discussed and were consistent with the theoretical expectation.
{"title":"Convolutional Autoencoder for Image Denoising: A Compositional Subspace Representation Perspective","authors":"M. Teow","doi":"10.1109/IICAIET51634.2021.9573657","DOIUrl":"https://doi.org/10.1109/IICAIET51634.2021.9573657","url":null,"abstract":"This study explores a convolutional autoencoder for image denoising with a proposed compositional subspace method. This modeling approach presents a structural and rigorous mathematical abstraction to understand a convolutional autoencoder's functional computation layers. The theoretical basis is that the best way to model a complex learning function is by using a composition of simple functions to form a multilayer successive cascaded function for complex representation. The proposed method has experimented with the Fashion-MNIST dataset. Experimental results are discussed and were consistent with the theoretical expectation.","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124927624","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-09-13DOI: 10.1109/IICAIET51634.2021.9573645
S. Tan, Ka-Man Chirs Lo, Y. Leau, G. Chung, F. Ahmedy
Mobile healthcare (mHealth) application and technologies have promised their cost-effectiveness to enhance healthcare quality, particularly in rural areas. However, the increased security incidents and leakage of patient data raise the concerns to address security risks and privacy issues of mhealth applications urgently. While recent mobile health applications that rely on password-based authentication cannot withstand password guessing and cracking attacks, several countermeasures such as One-Time Password (OTP), grid-based password, and biometric authentication have recently been implemented to protect mobile health applications. These countermeasures, however, can be thwarted by brute force attacks, man-in-the-middle attacks and persistent malware attacks. This paper proposed grid-based honey encryption by hybridising honey encryption with grid-based authentication. Compared to recent honey encryption limited in the hardening password attacks process, the proposed grid-based honey encryption can be further employed against shoulder surfing, smudge and replay attacks. Instead of rejecting access as a recent security defence mechanism in mobile healthcare applications, the proposed Grid-based Honey Encryption creates an indistinct counterfeit patient's record closely resembling the real patients' records in light of each off-base speculation legitimate password.
{"title":"Securing mHealth Applications with Grid-Based Honey Encryption","authors":"S. Tan, Ka-Man Chirs Lo, Y. Leau, G. Chung, F. Ahmedy","doi":"10.1109/IICAIET51634.2021.9573645","DOIUrl":"https://doi.org/10.1109/IICAIET51634.2021.9573645","url":null,"abstract":"Mobile healthcare (mHealth) application and technologies have promised their cost-effectiveness to enhance healthcare quality, particularly in rural areas. However, the increased security incidents and leakage of patient data raise the concerns to address security risks and privacy issues of mhealth applications urgently. While recent mobile health applications that rely on password-based authentication cannot withstand password guessing and cracking attacks, several countermeasures such as One-Time Password (OTP), grid-based password, and biometric authentication have recently been implemented to protect mobile health applications. These countermeasures, however, can be thwarted by brute force attacks, man-in-the-middle attacks and persistent malware attacks. This paper proposed grid-based honey encryption by hybridising honey encryption with grid-based authentication. Compared to recent honey encryption limited in the hardening password attacks process, the proposed grid-based honey encryption can be further employed against shoulder surfing, smudge and replay attacks. Instead of rejecting access as a recent security defence mechanism in mobile healthcare applications, the proposed Grid-based Honey Encryption creates an indistinct counterfeit patient's record closely resembling the real patients' records in light of each off-base speculation legitimate password.","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132369533","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-09-13DOI: 10.1109/IICAIET51634.2021.9573602
J. Khoo, Solomon Haw, Nicholas Su, Shakeeb Mulaafer
The outlook of maintaining higher quality perishable food sparks a lot of interest in the agriculture business. Food security is an important aspect to meet the demand of the growing population. For instance, postharvest losses amount to 1.3 billion tons a year which amounts to 33 percent of production as stated by the Food and Agriculture Department of United States. Real time monitoring of the supply chain can provide insight on perishable food to better handle pricing and allow respective stakeholders to act accordingly to maintain quality standards. Shelf life is described as the duration of a product to be safely consumed by the microbiological standards and retaining a desired sensory, physico-chemical and nutritional quality. Arrhenius equation is commonly used in the assessment of food quality albeit time consuming. The proposed approach with multiple linear regression (MLR) model is designed to estimate the lowest possible shelf life outcome given the monitoring condition during transportation.
{"title":"Kiwi Fruit IoT Shelf Life Estimation During Transportation with Cloud Computing","authors":"J. Khoo, Solomon Haw, Nicholas Su, Shakeeb Mulaafer","doi":"10.1109/IICAIET51634.2021.9573602","DOIUrl":"https://doi.org/10.1109/IICAIET51634.2021.9573602","url":null,"abstract":"The outlook of maintaining higher quality perishable food sparks a lot of interest in the agriculture business. Food security is an important aspect to meet the demand of the growing population. For instance, postharvest losses amount to 1.3 billion tons a year which amounts to 33 percent of production as stated by the Food and Agriculture Department of United States. Real time monitoring of the supply chain can provide insight on perishable food to better handle pricing and allow respective stakeholders to act accordingly to maintain quality standards. Shelf life is described as the duration of a product to be safely consumed by the microbiological standards and retaining a desired sensory, physico-chemical and nutritional quality. Arrhenius equation is commonly used in the assessment of food quality albeit time consuming. The proposed approach with multiple linear regression (MLR) model is designed to estimate the lowest possible shelf life outcome given the monitoring condition during transportation.","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122484673","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-09-13DOI: 10.1109/IICAIET51634.2021.9573739
Cassandra Sze Jin Yong, Soon Yee Chong, Chelvam Dasaratha Raman, R. Chin, Sainarayanan Gopalakrishnan, K. Teo
Electrical Impedance Tomography (EIT) estimates the electrical impedance distribution within a medium and produces cross-sectional images of an admittivity distribution inside an electrically conducting object. EIT in biomedicine application was first applied in lung due to it being large organs that allow large conductivity changes and is a promising technique since it allows continuous monitoring of the ventilation distribution. This study aims to explore the potential EIT technique in medical applications, with strategies to enhance the image reconstruction of the lung images. Performance of the enhanced image reconstruction is analyzed through simulation on the thorax Finite Element Model (FEM) based on a thorax CT image generated using NETGEN Mesher. To integrate and simulate EIT image of the thorax model, data are obtained from the forward and inverse model. Graz consensus Reconstruction algorithm for EIT (GREIT) technique is then applied as the consensus linear reconstruction algorithm for lung EIT images. Subsequently, the involvement of 3D imaging opens the opportunity to explore more electrode placement strategies for enhancement in image reconstruction. Performance of the reconstructed images based on electrode numbers and placement strategies are analyzed using the five figures of merits and classified into poor, average and good using Fuzzy Logic (FL). From the analysis, planar-offset configuration with 16-electrodes outperforms all others while planar configuration with 16-electrodes followed closely.
{"title":"Electrical Impedance Tomography with Fuzzy Logic Classification in Lung Image Reconstruction","authors":"Cassandra Sze Jin Yong, Soon Yee Chong, Chelvam Dasaratha Raman, R. Chin, Sainarayanan Gopalakrishnan, K. Teo","doi":"10.1109/IICAIET51634.2021.9573739","DOIUrl":"https://doi.org/10.1109/IICAIET51634.2021.9573739","url":null,"abstract":"Electrical Impedance Tomography (EIT) estimates the electrical impedance distribution within a medium and produces cross-sectional images of an admittivity distribution inside an electrically conducting object. EIT in biomedicine application was first applied in lung due to it being large organs that allow large conductivity changes and is a promising technique since it allows continuous monitoring of the ventilation distribution. This study aims to explore the potential EIT technique in medical applications, with strategies to enhance the image reconstruction of the lung images. Performance of the enhanced image reconstruction is analyzed through simulation on the thorax Finite Element Model (FEM) based on a thorax CT image generated using NETGEN Mesher. To integrate and simulate EIT image of the thorax model, data are obtained from the forward and inverse model. Graz consensus Reconstruction algorithm for EIT (GREIT) technique is then applied as the consensus linear reconstruction algorithm for lung EIT images. Subsequently, the involvement of 3D imaging opens the opportunity to explore more electrode placement strategies for enhancement in image reconstruction. Performance of the reconstructed images based on electrode numbers and placement strategies are analyzed using the five figures of merits and classified into poor, average and good using Fuzzy Logic (FL). From the analysis, planar-offset configuration with 16-electrodes outperforms all others while planar configuration with 16-electrodes followed closely.","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122896281","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-09-13DOI: 10.1109/IICAIET51634.2021.9573654
Sheng Zhu, M. K. Tan, R. Chin, B. Chua, Xiaoxi Hao, K. Teo
Engine failure is one of the major factors caused vehicle breakdown. In the current practice, the engine faults are diagnosed manually by mechanics and the accuracy is highly relied on their experience. Therefore, this study would like to explore the feasibility of implementing auto fault diagnosis using Probabilistic Neural Network (PNN). A benchmarked engine fault model is developed and simulated in Maltab. The proposed algorithm is designed to detect 9 common engine faults based on the information extracted from exhaust gas, such as hydrocarbon (HC), carbon monoxide (CO), oxides of nitrogen (NOx), carbon dioxide (CO2) and dioxygen (O2). The proposed PNN is trained using the collected engine fault data from experiment and the probability density of PNN is determined based on the Parzen window estimation method. Bayes decision rule is implemented for classifying the types of the engine faults. The simulated results show that the proposed algorithm has faster diagnosis speed, higher accuracy and consistent. The algorithm takes 0.038 s in diagnosing the fault and the average accuracy is 98.3 %.
{"title":"Engine Fault Diagnosis using Probabilistic Neural Network","authors":"Sheng Zhu, M. K. Tan, R. Chin, B. Chua, Xiaoxi Hao, K. Teo","doi":"10.1109/IICAIET51634.2021.9573654","DOIUrl":"https://doi.org/10.1109/IICAIET51634.2021.9573654","url":null,"abstract":"Engine failure is one of the major factors caused vehicle breakdown. In the current practice, the engine faults are diagnosed manually by mechanics and the accuracy is highly relied on their experience. Therefore, this study would like to explore the feasibility of implementing auto fault diagnosis using Probabilistic Neural Network (PNN). A benchmarked engine fault model is developed and simulated in Maltab. The proposed algorithm is designed to detect 9 common engine faults based on the information extracted from exhaust gas, such as hydrocarbon (HC), carbon monoxide (CO), oxides of nitrogen (NOx), carbon dioxide (CO2) and dioxygen (O2). The proposed PNN is trained using the collected engine fault data from experiment and the probability density of PNN is determined based on the Parzen window estimation method. Bayes decision rule is implemented for classifying the types of the engine faults. The simulated results show that the proposed algorithm has faster diagnosis speed, higher accuracy and consistent. The algorithm takes 0.038 s in diagnosing the fault and the average accuracy is 98.3 %.","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"178 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132188440","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-09-13DOI: 10.1109/IICAIET51634.2021.9573567
Aishwarya Asesh
A universal server coordinates the training of a single model on a largely distributed network of computers in federated learning. This setting can easily be expanded to a multi-task learning system in order to manage real-world federated datasets with high statistical heterogeneity across devices. Federated learning is very useful as a framework for real-world data and federated multi-task learning has been applied to convex models. This research work discusses and evaluates possibility of sparser gradient changes to outperform the existing state-of-the-art for federated learning on real-world federated datasets as well as imputed data values. The experiments investigate the effect of rolling data or data randomization and adaptive global frequency update scheduling on the convergence of the federated learning algorithm. The results show that convergence speed and gradient curve are considerably affected by number of contact rounds between worker and aggregator and is unaffected by data heterogeneity or client sampling. The research is the core part of an extended experimental setup that will follow to better understand the behavior of distributed learning, by developing a simulation to track weights and loss function gradients during the training.
{"title":"Federated Learning: Optimizing Objective Function","authors":"Aishwarya Asesh","doi":"10.1109/IICAIET51634.2021.9573567","DOIUrl":"https://doi.org/10.1109/IICAIET51634.2021.9573567","url":null,"abstract":"A universal server coordinates the training of a single model on a largely distributed network of computers in federated learning. This setting can easily be expanded to a multi-task learning system in order to manage real-world federated datasets with high statistical heterogeneity across devices. Federated learning is very useful as a framework for real-world data and federated multi-task learning has been applied to convex models. This research work discusses and evaluates possibility of sparser gradient changes to outperform the existing state-of-the-art for federated learning on real-world federated datasets as well as imputed data values. The experiments investigate the effect of rolling data or data randomization and adaptive global frequency update scheduling on the convergence of the federated learning algorithm. The results show that convergence speed and gradient curve are considerably affected by number of contact rounds between worker and aggregator and is unaffected by data heterogeneity or client sampling. The research is the core part of an extended experimental setup that will follow to better understand the behavior of distributed learning, by developing a simulation to track weights and loss function gradients during the training.","PeriodicalId":234229,"journal":{"name":"2021 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116257036","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：