Pub Date : 2023-03-15DOI: 10.1109/ISADS56919.2023.10092170
Gaku Sato, H. Yokoi, D. Toratani, T. Koga
With the increase in the number of small unmanned aircraft systems (sUAS) flights, the risk of collisions between sUAS and manned aircrafts such as helicopters flying at relatively low altitudes is also increasing. To improve safety at the low altitude airspace, we have been developing collision avoidance methods for sUAS. In our previous study, we developed a collision avoidance method that takes advantage of the performance of a multirotor sUAS in a assumed collision between a multirotor sUAS and a helicopter. However, for the sake of simplicity, this previous study did not consider the method of acquiring location information. Therefore, this study conducted collision avoidance simulations in a multilateration (MLAT) environment, which is assumed in helicopter surveillance, to examine how MLAT surveillance characteristics affect avoidance behavior.
{"title":"Collision Avoidance Method for Multirotor Small Unmanned Aircraft Systems in Multilateration Environments","authors":"Gaku Sato, H. Yokoi, D. Toratani, T. Koga","doi":"10.1109/ISADS56919.2023.10092170","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10092170","url":null,"abstract":"With the increase in the number of small unmanned aircraft systems (sUAS) flights, the risk of collisions between sUAS and manned aircrafts such as helicopters flying at relatively low altitudes is also increasing. To improve safety at the low altitude airspace, we have been developing collision avoidance methods for sUAS. In our previous study, we developed a collision avoidance method that takes advantage of the performance of a multirotor sUAS in a assumed collision between a multirotor sUAS and a helicopter. However, for the sake of simplicity, this previous study did not consider the method of acquiring location information. Therefore, this study conducted collision avoidance simulations in a multilateration (MLAT) environment, which is assumed in helicopter surveillance, to examine how MLAT surveillance characteristics affect avoidance behavior.","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124809586","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 : 2023-03-15DOI: 10.1109/ISADS56919.2023.10092171
G. Luca, Yinong Chen
The field of quantum computing is rapidly growing, with near term applications immediately available for use. The application of quantum computing to machine learning (i.e., quantum machine learning) is similarly growing rapidly. The presence of noisy intermediate-scale quantum (NISQ) era computers is further enabling research in the area. Historically, the barrier to entry of quantum computing has been nearly insurmountable for computer science students, or any other students who lack a strong physics background. However, quantum computing and quantum machine learning are becoming increasingly accessible, regardless of background. The goal of this paper is to present and demonstrate that the field is accessible to computer science students and to provide a sample curriculum. This curriculum can be used in a standalone class or as part of another machine learning class, as the authors have done.
{"title":"Teaching Quantum Machine Learning in Computer Science","authors":"G. Luca, Yinong Chen","doi":"10.1109/ISADS56919.2023.10092171","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10092171","url":null,"abstract":"The field of quantum computing is rapidly growing, with near term applications immediately available for use. The application of quantum computing to machine learning (i.e., quantum machine learning) is similarly growing rapidly. The presence of noisy intermediate-scale quantum (NISQ) era computers is further enabling research in the area. Historically, the barrier to entry of quantum computing has been nearly insurmountable for computer science students, or any other students who lack a strong physics background. However, quantum computing and quantum machine learning are becoming increasingly accessible, regardless of background. The goal of this paper is to present and demonstrate that the field is accessible to computer science students and to provide a sample curriculum. This curriculum can be used in a standalone class or as part of another machine learning class, as the authors have done.","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129027524","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 : 2023-03-15DOI: 10.1109/ISADS56919.2023.10092030
Gabriel Souza, Mickael Figueredo, Daniel Sabino, N. Cacho
The government has worked to improve technolo-gies to advance criminal investigations. It is very common for Brazilian public institutions to spend resources on systems to improve population security or investigations through artificial intelligence. A central point in this context is the data used by the institutions classified as highly sensitive. This sensitiveness creates a complex barrier to cooperation between governmental institutions from different areas. In this context, this study proposes a federated learning pipeline to encourage artificial intelligence model sharing between government institutions, taking advantage of high-security networks and computational resources from governmental institutions. We leveraged consolidated frameworks such as Docker and TensorFlow to ease the model sharing and training process without working with sensitive data risks. In this work, the performance of 5 different Federated Learning algorithms was tested using three different AI algorithms. In our experiments, the use of Federated Learning in the context of Brazilian governmental institutions proved to create models with performance similar to the standard Centralized Learning techniques in three different federated learning algorithms.
{"title":"A pipeline to collaborative AI models creation between Brazilian governmental institutions","authors":"Gabriel Souza, Mickael Figueredo, Daniel Sabino, N. Cacho","doi":"10.1109/ISADS56919.2023.10092030","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10092030","url":null,"abstract":"The government has worked to improve technolo-gies to advance criminal investigations. It is very common for Brazilian public institutions to spend resources on systems to improve population security or investigations through artificial intelligence. A central point in this context is the data used by the institutions classified as highly sensitive. This sensitiveness creates a complex barrier to cooperation between governmental institutions from different areas. In this context, this study proposes a federated learning pipeline to encourage artificial intelligence model sharing between government institutions, taking advantage of high-security networks and computational resources from governmental institutions. We leveraged consolidated frameworks such as Docker and TensorFlow to ease the model sharing and training process without working with sensitive data risks. In this work, the performance of 5 different Federated Learning algorithms was tested using three different AI algorithms. In our experiments, the use of Federated Learning in the context of Brazilian governmental institutions proved to create models with performance similar to the standard Centralized Learning techniques in three different federated learning algorithms.","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114086639","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 : 2023-03-15DOI: 10.1109/ISADS56919.2023.10092045
E. Moya-Albor, Armin Schwartzman, J. Brieva, Mauricio Pardo, Hiram Ponce, Rodrigo Chávez-Domínguez
In this paper, an automatic segmentation approach of the Viedma glacier terminus is proposed. The method uses multi-spectral images from the Landsat-5 satellite to determine the area of the glacier through computer vision techniques. The area of the glacier is estimated, and a linear model is fitted, obtaining a correlation of 0.968 between the measured area and a fit linear regression model. On the other hand, a bio-inspired optical flow estimation approach is used to calculate and visualize the displacement of the glacier through time. In addition, an analysis is performed between the temperature variation in the Southern Cone and the decrease of the glacier in the function of time. A linear trend (r2=0.95) shows that the analyzed area of the glacier has decreased by about 1.9% annually in the observation season. It reveals an inverse relationship between the change in the size of the glacier and global warming, showing that if the same conditions remain, the glacier’s zone analyzed in this work would be close to its disappearance in around 70 years, the time lapse in which a global temperature increase of 1.24 oC would be reached.
{"title":"A Computer Vision Approach to Terminus Movement Analysis of Viedma Glacier","authors":"E. Moya-Albor, Armin Schwartzman, J. Brieva, Mauricio Pardo, Hiram Ponce, Rodrigo Chávez-Domínguez","doi":"10.1109/ISADS56919.2023.10092045","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10092045","url":null,"abstract":"In this paper, an automatic segmentation approach of the Viedma glacier terminus is proposed. The method uses multi-spectral images from the Landsat-5 satellite to determine the area of the glacier through computer vision techniques. The area of the glacier is estimated, and a linear model is fitted, obtaining a correlation of 0.968 between the measured area and a fit linear regression model. On the other hand, a bio-inspired optical flow estimation approach is used to calculate and visualize the displacement of the glacier through time. In addition, an analysis is performed between the temperature variation in the Southern Cone and the decrease of the glacier in the function of time. A linear trend (r2=0.95) shows that the analyzed area of the glacier has decreased by about 1.9% annually in the observation season. It reveals an inverse relationship between the change in the size of the glacier and global warming, showing that if the same conditions remain, the glacier’s zone analyzed in this work would be close to its disappearance in around 70 years, the time lapse in which a global temperature increase of 1.24 oC would be reached.","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"167 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115203700","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 : 2023-03-15DOI: 10.1109/ISADS56919.2023.10092070
Roberto García Cedillo, D. Martínez-López, Eduardo Martínez Quintana, Andrea Pérez Guerra, J. P. S. H. Moreno, José Manuel Vega Hernández, E. Moya-Albor, Hiram Ponce, J. Brieva
In this paper, we present the development of an intelligent bicycle which will be able to help the user achieve a more efficient exercise routine via the control of a DC motor. This project was developed in several stages, from the approach of the system’s functions to the components that would conform to it in order to achieve a detailed concept that can meet the requirements correctly. The sector of the population that motivated the realization of this project and to whom it is mainly directed are all those who cycle in Mexico City and find their routines inefficient. Through the use of this bicycle, which has a heart rate measurement system, it is possible to monitor it to regulate the intensity of the exercise. It will be made possible by incorporating a motor that is activated as soon as it detects an elevated heart rate, which may mean that the user requires assistance or has to stop the exercise altogether. The results provide evidence that assisting the user does indeed help reduce overexertion.
{"title":"Development of an Electric Powered Assisted Cycle with a Heart Rate Sensor Control System","authors":"Roberto García Cedillo, D. Martínez-López, Eduardo Martínez Quintana, Andrea Pérez Guerra, J. P. S. H. Moreno, José Manuel Vega Hernández, E. Moya-Albor, Hiram Ponce, J. Brieva","doi":"10.1109/ISADS56919.2023.10092070","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10092070","url":null,"abstract":"In this paper, we present the development of an intelligent bicycle which will be able to help the user achieve a more efficient exercise routine via the control of a DC motor. This project was developed in several stages, from the approach of the system’s functions to the components that would conform to it in order to achieve a detailed concept that can meet the requirements correctly. The sector of the population that motivated the realization of this project and to whom it is mainly directed are all those who cycle in Mexico City and find their routines inefficient. Through the use of this bicycle, which has a heart rate measurement system, it is possible to monitor it to regulate the intensity of the exercise. It will be made possible by incorporating a motor that is activated as soon as it detects an elevated heart rate, which may mean that the user requires assistance or has to stop the exercise altogether. The results provide evidence that assisting the user does indeed help reduce overexertion.","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127904557","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 : 2023-03-15DOI: 10.1109/ISADS56919.2023.10092039
Jianyong Zhu, Bin Lu, Xiaoqiang Yu, Jie Xu, Tianyu Wo
Finding performance bottlenecks through bench-marking is one of the driving forces to improve the resource provision efficiency of cloud computing. Although existing benchmarks have been designed to improve the effectiveness in system performance evaluation, the following problems still exist in these benchmarks due to insufficient consideration of the characteristics of jobs in the production environment: (i) lacking of understanding for the details of workloads composition in the production environment, which reduces the authenticity of the job. (ii) the design of workloads submission patterns lacks quantization and reproducibility, which often relies on a random setting. In our benchmarking, multiple workloads are generated by analyzing and fine-grained matching the composition of workloads in the real production, and a design of workloads submission pattern based on LSTM time series prediction is proposed to simulate the real submission behavior. We finally demonstrate the effectiveness of our work by evaluating the impact of different workloads submission patterns on system performance evaluation.
{"title":"An Approach to Workload Generation for Cloud Benchmarking: a View from Alibaba Trace","authors":"Jianyong Zhu, Bin Lu, Xiaoqiang Yu, Jie Xu, Tianyu Wo","doi":"10.1109/ISADS56919.2023.10092039","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10092039","url":null,"abstract":"Finding performance bottlenecks through bench-marking is one of the driving forces to improve the resource provision efficiency of cloud computing. Although existing benchmarks have been designed to improve the effectiveness in system performance evaluation, the following problems still exist in these benchmarks due to insufficient consideration of the characteristics of jobs in the production environment: (i) lacking of understanding for the details of workloads composition in the production environment, which reduces the authenticity of the job. (ii) the design of workloads submission patterns lacks quantization and reproducibility, which often relies on a random setting. In our benchmarking, multiple workloads are generated by analyzing and fine-grained matching the composition of workloads in the real production, and a design of workloads submission pattern based on LSTM time series prediction is proposed to simulate the real submission behavior. We finally demonstrate the effectiveness of our work by evaluating the impact of different workloads submission patterns on system performance evaluation.","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116753370","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 : 2023-03-15DOI: 10.1109/ISADS56919.2023.10092108
D. Grimm, Marc Schindewolf, E. Sax
Due to the increasing automation and connectivity of future, software-defined vehicles, the interest in safety and security in this field is growing. In contrast to automated driving functions, which today are only tested and approved for specific operating conditions, the vehicle must be safe and secure in any situation, even if there is a malfunction or intentional manipulation. Resilient software and hardware architectures for vehicles are, therefore, a research topic of growing importance. However, due to the holistic nature of this approach, researching and testing these systems is fraught with challenges. For example, once Machine Learning-based approaches come into play, large amounts of data are required. At the same time, tests of the systems need to take into account the whole vehicle and its ecosystem and be scalable and user-friendly. This work, therefore, presents a new simulation-based method for testing and developing functions for software-defined connected vehicles. The focus is especially on safety and security in combination with cloud services and the consideration of the vehicle fleet. The technologies are presented in detail, relying mainly on the simulator CARLA, the virtualization with Proxmox, and ROS2-based vehicle functions. What differentiates this approach from others is the purely virtual and open-source approach, which increases the availability for others. Results are shown based on early quantitative measures and on outlining two exemplary use cases.
{"title":"Fleet in the Loop: An Open Source approach for design and test of resilient vehicle architectures","authors":"D. Grimm, Marc Schindewolf, E. Sax","doi":"10.1109/ISADS56919.2023.10092108","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10092108","url":null,"abstract":"Due to the increasing automation and connectivity of future, software-defined vehicles, the interest in safety and security in this field is growing. In contrast to automated driving functions, which today are only tested and approved for specific operating conditions, the vehicle must be safe and secure in any situation, even if there is a malfunction or intentional manipulation. Resilient software and hardware architectures for vehicles are, therefore, a research topic of growing importance. However, due to the holistic nature of this approach, researching and testing these systems is fraught with challenges. For example, once Machine Learning-based approaches come into play, large amounts of data are required. At the same time, tests of the systems need to take into account the whole vehicle and its ecosystem and be scalable and user-friendly. This work, therefore, presents a new simulation-based method for testing and developing functions for software-defined connected vehicles. The focus is especially on safety and security in combination with cloud services and the consideration of the vehicle fleet. The technologies are presented in detail, relying mainly on the simulator CARLA, the virtualization with Proxmox, and ROS2-based vehicle functions. What differentiates this approach from others is the purely virtual and open-source approach, which increases the availability for others. Results are shown based on early quantitative measures and on outlining two exemplary use cases.","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"238 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124253457","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 : 2023-03-15DOI: 10.1109/ISADS56919.2023.10092024
X. Zhang, Y. Lai
Industrial Control System(ICS) security is one of the lifebloods of national development. Fully understanding of its vulnerabilities plays an important role in the actual application scenarios. Meanwhile, an attacker may also exploit multiple vulnerabilities to achieve the final malicious purpose, such as the Stuxnet worm. In order to solve the above problems, we construct a Knowledge Graph(KG) of heterogeneous ICSs, and propose a potential relationship mining method (R-HetGNN) based on this graph. The method solves the multi-modality problem in KG aggregation and KG-heterogeneity problem. Besides, we use random walk algorithm to solve the ulti-level neighbor problem. Experimental results on a real-world dataset show that R-HetGNN achieved 83.0% on the F1 score, superior to other knowledge reasoning modules, such as GAT and TransE.
{"title":"Mining of Potential Relationships based on the Knowledge Graph of Industrial Control Systems","authors":"X. Zhang, Y. Lai","doi":"10.1109/ISADS56919.2023.10092024","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10092024","url":null,"abstract":"Industrial Control System(ICS) security is one of the lifebloods of national development. Fully understanding of its vulnerabilities plays an important role in the actual application scenarios. Meanwhile, an attacker may also exploit multiple vulnerabilities to achieve the final malicious purpose, such as the Stuxnet worm. In order to solve the above problems, we construct a Knowledge Graph(KG) of heterogeneous ICSs, and propose a potential relationship mining method (R-HetGNN) based on this graph. The method solves the multi-modality problem in KG aggregation and KG-heterogeneity problem. Besides, we use random walk algorithm to solve the ulti-level neighbor problem. Experimental results on a real-world dataset show that R-HetGNN achieved 83.0% on the F1 score, superior to other knowledge reasoning modules, such as GAT and TransE.","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134008335","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 : 2023-03-15DOI: 10.1109/ISADS56919.2023.10092053
Abeer M. Almalky, Khaled R. Ahmed
Since the number of people worldwide is anticipated to reach 9 billion people by 2050, the agriculture production needs to be increased up to 70% to manage the anticipated increasing of human demand. However, weeds are one of the most harmful factors that negatively impact the crops production, quality, and cause economical loses. Accordingly, automating the weed detection, classification, and counting of weeds per their growth stages will help farmers to choose the appropriate weeds’ controlling techniques. In this paper, UAV was used for collecting a dataset, which consists of four weed (Consolida Regalis) growth stages. Additionally, a deep learning model (YOLOv5) was developed and trained for detecting weed, classifying weed’s growth stages, and counting the number of weeds occurrences in each part of the field. The results report that the best precision (82.7%) is generated by the Yolov5-Large model in detecting and classifying the weed’s growth stages. According to the best performance in terms of recall, Yolov5-sma11 model has the best recall of 79.4%. For counting the instances of weeds per the four growth stages in real-time, Yolov5-sma11 model showes counting time of 0.033 millisecond per frame.
{"title":"Real Time Deep Learning Algorithm for Counting Weed’s Growth Stages","authors":"Abeer M. Almalky, Khaled R. Ahmed","doi":"10.1109/ISADS56919.2023.10092053","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10092053","url":null,"abstract":"Since the number of people worldwide is anticipated to reach 9 billion people by 2050, the agriculture production needs to be increased up to 70% to manage the anticipated increasing of human demand. However, weeds are one of the most harmful factors that negatively impact the crops production, quality, and cause economical loses. Accordingly, automating the weed detection, classification, and counting of weeds per their growth stages will help farmers to choose the appropriate weeds’ controlling techniques. In this paper, UAV was used for collecting a dataset, which consists of four weed (Consolida Regalis) growth stages. Additionally, a deep learning model (YOLOv5) was developed and trained for detecting weed, classifying weed’s growth stages, and counting the number of weeds occurrences in each part of the field. The results report that the best precision (82.7%) is generated by the Yolov5-Large model in detecting and classifying the weed’s growth stages. According to the best performance in terms of recall, Yolov5-sma11 model has the best recall of 79.4%. For counting the instances of weeds per the four growth stages in real-time, Yolov5-sma11 model showes counting time of 0.033 millisecond per frame.","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"335 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122917736","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 : 2023-03-15DOI: 10.1109/ISADS56919.2023.10091996
Steven A. Wright
Deep learning, big data, IoT and blockchain are individually very important research topics of today’s technology, and their combination has the potential to generate additional synergy. Such synergy could enable decentralized and intelligent automated applications to achieve safety, security and optimize performance and economy. Deep learning, big data, IoT and blockchain all rely on infrastructure capabilities in computing and communications that are increasingly decentralized. Edge computing deployments and architectures are commencing with 5G and expected to accelerate in 6G. Existing application domains like healthcare and finance are starting to explore the integration of these technologies. Newly emerging application areas such as the metaverse may well require native support of decentralized deep learning to achieve their potential. But the path of new technology development is never smooth. New challenges have been identified and additional architectural frameworks have been developed to overcome some of these issues. Decentralizing deep learning enables increased scale for AI implementations, but also enables improvements in privacy and trustworthiness. The plethora of literature emerging on decentralized deep learning prompts the need for rationale criteria to support design decisions for implementation to utilize decentralized deep learning
{"title":"Why Decentralize Deep Learning?","authors":"Steven A. Wright","doi":"10.1109/ISADS56919.2023.10091996","DOIUrl":"https://doi.org/10.1109/ISADS56919.2023.10091996","url":null,"abstract":"Deep learning, big data, IoT and blockchain are individually very important research topics of today’s technology, and their combination has the potential to generate additional synergy. Such synergy could enable decentralized and intelligent automated applications to achieve safety, security and optimize performance and economy. Deep learning, big data, IoT and blockchain all rely on infrastructure capabilities in computing and communications that are increasingly decentralized. Edge computing deployments and architectures are commencing with 5G and expected to accelerate in 6G. Existing application domains like healthcare and finance are starting to explore the integration of these technologies. Newly emerging application areas such as the metaverse may well require native support of decentralized deep learning to achieve their potential. But the path of new technology development is never smooth. New challenges have been identified and additional architectural frameworks have been developed to overcome some of these issues. Decentralizing deep learning enables increased scale for AI implementations, but also enables improvements in privacy and trustworthiness. The plethora of literature emerging on decentralized deep learning prompts the need for rationale criteria to support design decisions for implementation to utilize decentralized deep learning","PeriodicalId":412453,"journal":{"name":"2023 IEEE 15th International Symposium on Autonomous Decentralized System (ISADS)","volume":"3 1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117319309","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：