Pub Date : 2021-07-01DOI: 10.1109/CCICT53244.2021.00047
P. Vishwakarma, Randeep Singh
The new IoT architecture helps us to build small devices that can sense, process, and communicate so that sensors, embedded, and other ‘services can be developed to enable us to understand the surroundings. For continuous cardiovascular health surveillance, the IoT-assisted electrocardiogram (ECG) system for safe data transmission has been suggested. The modern paradigm of the Internet of Things enables the creation of small devices with sensing, processing, and communication capabilities that make sensors, embedded devices, and other ‘stuff’ capable of understanding the environment The Internet of Things (IoT) and intelligent medical devices have transformed healthcare systems, allowing patient health conditions to be monitored and screened anywhere and every time. Due to the sudden and enormous increase in patients during a pandemic of corona-virus, it is essential that patients are constantly monitored until any serious illness or infection takes place. According to the transfer of the enormous amount of confidential health information produced by patients who do not wish to disclose their personal medical information, Concerns about IoT data protection are still an extremely serious issue. The advances made in IoT technology in recent years have supported interactions between smart objects – things through the Internet in a transparent fashion. One of the applications in IoT is healthcare and sensors, the processor aggregator, and the data storage platform.
{"title":"A – Review on IoT-Assisted ECG Monitoring Framework for Health Care Applications","authors":"P. Vishwakarma, Randeep Singh","doi":"10.1109/CCICT53244.2021.00047","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00047","url":null,"abstract":"The new IoT architecture helps us to build small devices that can sense, process, and communicate so that sensors, embedded, and other ‘services can be developed to enable us to understand the surroundings. For continuous cardiovascular health surveillance, the IoT-assisted electrocardiogram (ECG) system for safe data transmission has been suggested. The modern paradigm of the Internet of Things enables the creation of small devices with sensing, processing, and communication capabilities that make sensors, embedded devices, and other ‘stuff’ capable of understanding the environment The Internet of Things (IoT) and intelligent medical devices have transformed healthcare systems, allowing patient health conditions to be monitored and screened anywhere and every time. Due to the sudden and enormous increase in patients during a pandemic of corona-virus, it is essential that patients are constantly monitored until any serious illness or infection takes place. According to the transfer of the enormous amount of confidential health information produced by patients who do not wish to disclose their personal medical information, Concerns about IoT data protection are still an extremely serious issue. The advances made in IoT technology in recent years have supported interactions between smart objects – things through the Internet in a transparent fashion. One of the applications in IoT is healthcare and sensors, the processor aggregator, and the data storage platform.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123560644","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 : 2021-07-01DOI: 10.1109/CCICT53244.2021.00068
Rajesh P. Deshmukh, Mugdha Avadhut, Tanvi Vairagade, Shreya B. Kolhe, S. Dehankar
Humans cannot operate the loads concurrently in the industries. So, with the help of rapid technologies, they can resolve the problems. This particular paper discusses a way to operate load at a time in industries. The system strives to make use of ZigBee along with the microcontroller to upgrade the industrial monitoring standard. To perform the existing regular monitoring purpose efficiently, this method employs the ZigBee wireless technology for remote monitoring. Several sensors are deployed in our project to monitor industrial parameters like temperature, current, voltage, etc. If there is any problem with the load, they will be cut off and the necessary information will be conveyed through the ZigBee to the server. The application of the ZigBee is combined with a microcontroller and hence the industrial measurements constitute an efficient innovative technology. Performing wireless and wired computing measurement and monitoring, this technique can do correct and methodical monitoring operations. Parameters were carefully selected based on the potential hazards which can help in the normal working of the industrial machines.
{"title":"Industry Monitoring System","authors":"Rajesh P. Deshmukh, Mugdha Avadhut, Tanvi Vairagade, Shreya B. Kolhe, S. Dehankar","doi":"10.1109/CCICT53244.2021.00068","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00068","url":null,"abstract":"Humans cannot operate the loads concurrently in the industries. So, with the help of rapid technologies, they can resolve the problems. This particular paper discusses a way to operate load at a time in industries. The system strives to make use of ZigBee along with the microcontroller to upgrade the industrial monitoring standard. To perform the existing regular monitoring purpose efficiently, this method employs the ZigBee wireless technology for remote monitoring. Several sensors are deployed in our project to monitor industrial parameters like temperature, current, voltage, etc. If there is any problem with the load, they will be cut off and the necessary information will be conveyed through the ZigBee to the server. The application of the ZigBee is combined with a microcontroller and hence the industrial measurements constitute an efficient innovative technology. Performing wireless and wired computing measurement and monitoring, this technique can do correct and methodical monitoring operations. Parameters were carefully selected based on the potential hazards which can help in the normal working of the industrial machines.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116399848","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 : 2021-07-01DOI: 10.1109/CCICT53244.2021.00027
Ankit Sharma, Manisha J Nen
Quantum is an emerging technology with constant research efforts in its various fields. One of the distinct advantage with the quantum technology is the use of qubits which gives inherent parallelism and security. However, at the same time qubits poses unique challenges due to their quantum properties. Quantum computation can further be enhanced with the use of quantum memory. In this paper we propose a technique to access stored quantum state multiple times or by multiple users at same instance.
{"title":"Quantum Memory Multiple Access","authors":"Ankit Sharma, Manisha J Nen","doi":"10.1109/CCICT53244.2021.00027","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00027","url":null,"abstract":"Quantum is an emerging technology with constant research efforts in its various fields. One of the distinct advantage with the quantum technology is the use of qubits which gives inherent parallelism and security. However, at the same time qubits poses unique challenges due to their quantum properties. Quantum computation can further be enhanced with the use of quantum memory. In this paper we propose a technique to access stored quantum state multiple times or by multiple users at same instance.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126720245","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 : 2021-07-01DOI: 10.1109/CCICT53244.2021.00019
Ankita Ankita, Shalli Rani
Present technology 5G deals with a variety of fields such as healthcare, industry, transportation making every domain smart. But security and privacy are important factors from previous to upcoming network 6G therefore an eye on the security and privacy domain is a must. In this paper, the main focus is on Malware and Ransomware attacks, and based on this, dealing with both of the attacks through Machine Learning and Deep Learning models and also by comparing both of them based on their accuracy. The collaboration of Machine and Deep Learning with the upcoming network 6G achieves the betterment in the security domain as it is the major field to work upon.
{"title":"Machine Learning and Deep Learning for Malware and Ransomware Attacks in 6G Network","authors":"Ankita Ankita, Shalli Rani","doi":"10.1109/CCICT53244.2021.00019","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00019","url":null,"abstract":"Present technology 5G deals with a variety of fields such as healthcare, industry, transportation making every domain smart. But security and privacy are important factors from previous to upcoming network 6G therefore an eye on the security and privacy domain is a must. In this paper, the main focus is on Malware and Ransomware attacks, and based on this, dealing with both of the attacks through Machine Learning and Deep Learning models and also by comparing both of them based on their accuracy. The collaboration of Machine and Deep Learning with the upcoming network 6G achieves the betterment in the security domain as it is the major field to work upon.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127438338","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 : 2021-07-01DOI: 10.1109/CCICT53244.2021.00036
Vedang Ratan Vatsa, P. Chhaparwal
With the rising digitization and efficiency in digital service delivery, Estonia is a good example of how technology has revolutionized the traditional approach of governance and delivery of services by the government. This paper has presented a review of the current ecosystem of various e-governance initiatives in Estonia by reviewing major service delivery modules and initiatives. This paper also intends to deliver a policy framework and technological model for developing economies while focusing on government bodies and non-profits to embrace digitization mechanisms.
{"title":"Estonia’s e-governance and digital public service delivery solutions","authors":"Vedang Ratan Vatsa, P. Chhaparwal","doi":"10.1109/CCICT53244.2021.00036","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00036","url":null,"abstract":"With the rising digitization and efficiency in digital service delivery, Estonia is a good example of how technology has revolutionized the traditional approach of governance and delivery of services by the government. This paper has presented a review of the current ecosystem of various e-governance initiatives in Estonia by reviewing major service delivery modules and initiatives. This paper also intends to deliver a policy framework and technological model for developing economies while focusing on government bodies and non-profits to embrace digitization mechanisms.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130663512","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 : 2021-07-01DOI: 10.1109/CCICT53244.2021.00046
Rohit Bansal, Ankur Gupta, Ram Singh, V. K. Nassa
Research has focused on the implementation of E-Learning amidst the COVID-19 pandemic. During COVID-19, the school and educational institutions were closed due to lockdown. During this period the classes of students are taken online. The digital technology used for e-learning during the COVID-19 pandemic has gained popularity in a very short period. Online classes are taken using Microsoft team, any desk, Zoom, WhatsApp applications. Educational contents are transferred frequently over the internet. Research is considering the impact of e-learning amidst COVID-19 and considering issues such as performance and security during transmission of digital content. The education is provided over cloud environment in a more secure manner with better performance. It has been observed that there have been several kinds of research in the area of cloud computing to provide online education. Issues in such research are performance and security of data. There is a need for a high-speed network to transfer educational content from one place to another. The educational contents needed to be secured and compressed at the time of data transfer. Cloud computing applications and the role of the cloud in e-learning are considered during research. The proposed work is supposed to integrate the proposed mechanism in the educational module. The proposed system is supposed to be secure and fast because data is compressed first then data is encrypted on the sender side. On receiving end the data is decrypted and decompressed. Delay in transmission issue is resolved because the size of data is less during transmission. Moreover, the packet dropping ratio gets reduced. The probability of cracking encrypted files also gets reduced as the data is encrypted after compression.
{"title":"Role and Impact of Digital Technologies in E-Learning amidst COVID-19 Pandemic","authors":"Rohit Bansal, Ankur Gupta, Ram Singh, V. K. Nassa","doi":"10.1109/CCICT53244.2021.00046","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00046","url":null,"abstract":"Research has focused on the implementation of E-Learning amidst the COVID-19 pandemic. During COVID-19, the school and educational institutions were closed due to lockdown. During this period the classes of students are taken online. The digital technology used for e-learning during the COVID-19 pandemic has gained popularity in a very short period. Online classes are taken using Microsoft team, any desk, Zoom, WhatsApp applications. Educational contents are transferred frequently over the internet. Research is considering the impact of e-learning amidst COVID-19 and considering issues such as performance and security during transmission of digital content. The education is provided over cloud environment in a more secure manner with better performance. It has been observed that there have been several kinds of research in the area of cloud computing to provide online education. Issues in such research are performance and security of data. There is a need for a high-speed network to transfer educational content from one place to another. The educational contents needed to be secured and compressed at the time of data transfer. Cloud computing applications and the role of the cloud in e-learning are considered during research. The proposed work is supposed to integrate the proposed mechanism in the educational module. The proposed system is supposed to be secure and fast because data is compressed first then data is encrypted on the sender side. On receiving end the data is decrypted and decompressed. Delay in transmission issue is resolved because the size of data is less during transmission. Moreover, the packet dropping ratio gets reduced. The probability of cracking encrypted files also gets reduced as the data is encrypted after compression.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128830755","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 : 2021-07-01DOI: 10.1109/CCICT53244.2021.00054
S. Zakariya, I. Khan
The method of identifying related images in an image database is known as image retrieval. Text-based and content-based data processing are the two main forms of image retrieval processes. The visual characteristics of an image, such as color, texture, shape, and spatial design, are used in the content-based approach of image retrieval. In unsupervised mode, the images are retrieved using a cluster-based graph partitioning algorithm. The efficiency of different content based image retrieval systems is contrasted in this paper by fusing multiple image characteristics. The creation of four versions resulted from the integration of several features. Compute the union of all four models by normalizing the value between 0 and 1. The data comes from the COREL image database, which includes 1000 images of the same resolution. According to this article, images can be best recovered using three model-based features rather than two features. The accuracy of the union is thought to be superior.
{"title":"Analysis of Multi-Features Combination of Unsupervised Content Based Image Retrieval with Different Degrees of Accuracy","authors":"S. Zakariya, I. Khan","doi":"10.1109/CCICT53244.2021.00054","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00054","url":null,"abstract":"The method of identifying related images in an image database is known as image retrieval. Text-based and content-based data processing are the two main forms of image retrieval processes. The visual characteristics of an image, such as color, texture, shape, and spatial design, are used in the content-based approach of image retrieval. In unsupervised mode, the images are retrieved using a cluster-based graph partitioning algorithm. The efficiency of different content based image retrieval systems is contrasted in this paper by fusing multiple image characteristics. The creation of four versions resulted from the integration of several features. Compute the union of all four models by normalizing the value between 0 and 1. The data comes from the COREL image database, which includes 1000 images of the same resolution. According to this article, images can be best recovered using three model-based features rather than two features. The accuracy of the union is thought to be superior.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"142 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134055879","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}
Predicting the correlation between the events and price movement is quiet challenge in the dynamic environment. The refreshing rate of stock price results huge volume of data and therefore forecasting the financial series is chaotic and stochastic. The volume participation of share is determined by other facts including sectorial or individual share news and it leads to volume increased and price increased shares in the nonlinear market. Most of the times the price relies on mathematical model rather than news or information shared in the media. New investors find it is difficult to consider either news or mathematical models for prediction. Our proposed model recommends decision to novice traders to avoid such loses in their portfolio using massive data. Using this approach, an investor can see the impact of an event and its outcome instead of betting on the shares randomly and reduce the false effect on trading the news.
{"title":"A study on Strategies of Trading the News Using Massive Data Mining","authors":"Prabakaran Natarajan, Rajasekaran Palaniappan, Kannadasan Rajenderan, Nagarajan Pandian","doi":"10.1109/CCICT53244.2021.00029","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00029","url":null,"abstract":"Predicting the correlation between the events and price movement is quiet challenge in the dynamic environment. The refreshing rate of stock price results huge volume of data and therefore forecasting the financial series is chaotic and stochastic. The volume participation of share is determined by other facts including sectorial or individual share news and it leads to volume increased and price increased shares in the nonlinear market. Most of the times the price relies on mathematical model rather than news or information shared in the media. New investors find it is difficult to consider either news or mathematical models for prediction. Our proposed model recommends decision to novice traders to avoid such loses in their portfolio using massive data. Using this approach, an investor can see the impact of an event and its outcome instead of betting on the shares randomly and reduce the false effect on trading the news.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"143 9","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114102097","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 : 2021-07-01DOI: 10.1109/CCICT53244.2021.00038
Shubham Vyas, R. Tyagi, Charu Jain, Shashank Sahu
For data ware housing projects, there are multiple licensed ETL (Extraction Transformation & Load) tools available in the market. To process and pass the data between two applications industry is using ETL tools like IBM Info-sphere Data Stage, Informatica, Ab Initio etc. These tools are exceptionally costly and has recurring enterprise licensees, and processed data is not available in real time, data is getting processed in batches and is available during pre-defined time intervals or on demand. Industry has started adopting the Open Source technologies to avoid the huge licensing cost and that also includes the complete end to end IT infrastructure cost. Open Source technologies and frameworks enables users to run projects with best in class performance and within the budget.In this literature survey paper, all possible technologies have been studied and evaluated, available in the market capable of real/ “near-real-time” streaming. All licensed and open source products which are utilized and evaluated by various IT organizations and which are also evaluated by researchers have been included in this survey. There is a need of a distributed scalable technology that enables the users to ensure availability of data from one end point to another in real time with good throughput, performance and low latency. To study this, a detailed comparative survey of an open source technology Apache Kafka has been done and it compared with the other available technologies capable of doing real time streaming.
对于数据仓库项目，市场上有多种授权的ETL(提取转换和加载)工具。为了在两个应用程序之间处理和传递数据，业界正在使用诸如IBM Info-sphere data Stage、Informatica、Ab Initio等ETL工具。这些工具非常昂贵，并且需要反复获得企业许可，处理后的数据不是实时可用的，数据是分批处理的，可以在预定义的时间间隔内或按需使用。业界已经开始采用开源技术，以避免巨大的许可成本，其中还包括完整的端到端IT基础设施成本。开源技术和框架使用户能够在预算范围内以一流的性能运行项目。在这篇文献调查论文中，所有可能的技术都被研究和评估，在市场上能够实现真实/“近实时”流。所有被各种IT组织使用和评估的、也被研究人员评估的许可和开源产品都包含在本次调查中。需要一种分布式可扩展技术，使用户能够确保数据从一个端点到另一个端点的实时可用性，并且具有良好的吞吐量、性能和低延迟。为了研究这一点，我们对开源技术Apache Kafka进行了详细的比较调查，并将其与其他能够进行实时流的可用技术进行了比较。
{"title":"Literature Review : A Comparative Study of Real Time Streaming Technologies and Apache Kafka","authors":"Shubham Vyas, R. Tyagi, Charu Jain, Shashank Sahu","doi":"10.1109/CCICT53244.2021.00038","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00038","url":null,"abstract":"For data ware housing projects, there are multiple licensed ETL (Extraction Transformation & Load) tools available in the market. To process and pass the data between two applications industry is using ETL tools like IBM Info-sphere Data Stage, Informatica, Ab Initio etc. These tools are exceptionally costly and has recurring enterprise licensees, and processed data is not available in real time, data is getting processed in batches and is available during pre-defined time intervals or on demand. Industry has started adopting the Open Source technologies to avoid the huge licensing cost and that also includes the complete end to end IT infrastructure cost. Open Source technologies and frameworks enables users to run projects with best in class performance and within the budget.In this literature survey paper, all possible technologies have been studied and evaluated, available in the market capable of real/ “near-real-time” streaming. All licensed and open source products which are utilized and evaluated by various IT organizations and which are also evaluated by researchers have been included in this survey. There is a need of a distributed scalable technology that enables the users to ensure availability of data from one end point to another in real time with good throughput, performance and low latency. To study this, a detailed comparative survey of an open source technology Apache Kafka has been done and it compared with the other available technologies capable of doing real time streaming.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"49 6","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120882940","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 : 2021-07-01DOI: 10.1109/CCICT53244.2021.00012
S Vignesh, Bhawna Bhawna, Shubham Anand, Shailender Kumar
It is evident from recent developments that deep learning has the potential to be a fundamental part in almost every new technology that comes in the future. Deep learning models have performed exceedingly well on standard image classification problems, but their performance drops drastically when presented with adversarial inputs that are created by adding specific small perturbations to the original image. This paper will be divided into two sections. The first section will be a complete review of the existing research in this field. We will provide the reader with the basic concepts of adversarial learning and a broad classification of various adversarial attacks and defenses. In the second section, we propose an ensemble model with max voting and test the impact of adversarial attacks by converting the 10-class problem over MNIST images into 10 binary classification problems. Each weight in the middle layers of a multi-class neural network is shared across all the output classes of the model. We call them the shared weights of the network. Our proposed model consists of no shared weights, shows a slight improve in accuracy against adversarial samples and can detect out-of-domain inputs.
{"title":"A Comprehensive Review of Adversarial Learning and Impact of Unsharing Weights Across Classes","authors":"S Vignesh, Bhawna Bhawna, Shubham Anand, Shailender Kumar","doi":"10.1109/CCICT53244.2021.00012","DOIUrl":"https://doi.org/10.1109/CCICT53244.2021.00012","url":null,"abstract":"It is evident from recent developments that deep learning has the potential to be a fundamental part in almost every new technology that comes in the future. Deep learning models have performed exceedingly well on standard image classification problems, but their performance drops drastically when presented with adversarial inputs that are created by adding specific small perturbations to the original image. This paper will be divided into two sections. The first section will be a complete review of the existing research in this field. We will provide the reader with the basic concepts of adversarial learning and a broad classification of various adversarial attacks and defenses. In the second section, we propose an ensemble model with max voting and test the impact of adversarial attacks by converting the 10-class problem over MNIST images into 10 binary classification problems. Each weight in the middle layers of a multi-class neural network is shared across all the output classes of the model. We call them the shared weights of the network. Our proposed model consists of no shared weights, shows a slight improve in accuracy against adversarial samples and can detect out-of-domain inputs.","PeriodicalId":213095,"journal":{"name":"2021 Fourth International Conference on Computational Intelligence and Communication Technologies (CCICT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129701741","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：