Pub Date : 2024-01-10DOI: 10.2174/0126662558286875231215054324
Salma Fayaz, Syed Zubair Ahmad Shah, Nusrat Mohi ud din, Naillah Gul, Assif Assad
Deep Learning (DL) models have demonstrated remarkable proficiency in image classification and recognition tasks, surpassing human capabilities. The observed enhancement in performance can be attributed to the utilization of extensive datasets. Nevertheless, DL models have huge data requirements. Widening the learning capability of such models from limited samples even today remains a challenge, given the intrinsic constraints of small da-tasets. The trifecta of challenges, encompassing limited labeled datasets, privacy, poor general-ization performance, and the costliness of annotations, further compounds the difficulty in achieving robust model performance. Overcoming the challenge of expanding the learning ca-pabilities of Deep Learning models with limited sample sizes remains a pressing concern even today. To address this critical issue, our study conducts a meticulous examination of estab-lished methodologies, such as Data Augmentation and Transfer Learning, which offer promis-ing solutions to data scarcity dilemmas. Data Augmentation, a powerful technique, amplifies the size of small datasets through a diverse array of strategies. These encompass geometric transformations, kernel filter manipulations, neural style transfer amalgamation, random eras-ing, Generative Adversarial Networks, augmentations in feature space, and adversarial and me-ta-learning training paradigms. Furthermore, Transfer Learning emerges as a crucial tool, leveraging pre-trained models to fa-cilitate knowledge transfer between models or enabling the retraining of models on analogous datasets. Through our comprehensive investigation, we provide profound insights into how the synergistic application of these two techniques can significantly enhance the performance of classification tasks, effectively magnifying scarce datasets. This augmentation in data availa-bility not only addresses the immediate challenges posed by limited datasets but also unlocks the full potential of working with Big Data in a new era of possibilities in DL applications.
{"title":"Advancements in Data Augmentation and Transfer Learning: A Comprehensive Survey to Address Data Scarcity Challenges","authors":"Salma Fayaz, Syed Zubair Ahmad Shah, Nusrat Mohi ud din, Naillah Gul, Assif Assad","doi":"10.2174/0126662558286875231215054324","DOIUrl":"https://doi.org/10.2174/0126662558286875231215054324","url":null,"abstract":"\u0000\u0000Deep Learning (DL) models have demonstrated remarkable proficiency in image classification and recognition tasks, surpassing human capabilities. The observed enhancement in performance can be attributed to the utilization of extensive datasets. Nevertheless, DL models have huge data requirements. Widening the learning capability of such models from limited samples even today remains a challenge, given the intrinsic constraints of small da-tasets. The trifecta of challenges, encompassing limited labeled datasets, privacy, poor general-ization performance, and the costliness of annotations, further compounds the difficulty in achieving robust model performance. Overcoming the challenge of expanding the learning ca-pabilities of Deep Learning models with limited sample sizes remains a pressing concern even today. To address this critical issue, our study conducts a meticulous examination of estab-lished methodologies, such as Data Augmentation and Transfer Learning, which offer promis-ing solutions to data scarcity dilemmas. Data Augmentation, a powerful technique, amplifies the size of small datasets through a diverse array of strategies. These encompass geometric transformations, kernel filter manipulations, neural style transfer amalgamation, random eras-ing, Generative Adversarial Networks, augmentations in feature space, and adversarial and me-ta-learning training paradigms.\u0000Furthermore, Transfer Learning emerges as a crucial tool, leveraging pre-trained models to fa-cilitate knowledge transfer between models or enabling the retraining of models on analogous datasets. Through our comprehensive investigation, we provide profound insights into how the synergistic application of these two techniques can significantly enhance the performance of classification tasks, effectively magnifying scarce datasets. This augmentation in data availa-bility not only addresses the immediate challenges posed by limited datasets but also unlocks the full potential of working with Big Data in a new era of possibilities in DL applications.\u0000","PeriodicalId":506582,"journal":{"name":"Recent Advances in Computer Science and Communications","volume":"89 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139440441","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}
The pervasive adoption of edge computing is reshaping real-time big data analysis, smart city management, intelligent transportation, and various other domains. Its appeal lies in its distributed nature, decentralization, low latency, mobile support, and spatial awareness. However, this ubiquity exposes data to security threats, jeopardizing privacy and integrity. Consequently, access control assumes paramount importance in securing data sharing within edge computing and blockchain technologies. This research addresses this critical issue by conducting a comprehensive study on access control techniques within the context of edge computing and blockchain for secure data sharing. Our methodology commences with an exhaustive review of relevant articles, aiming to identify and expound upon gaps in existing research. Subsequently, we perform a meticulous analysis of access control methods, mechanisms, and performance metrics, seeking to establish a holistic understanding of the landscape. The culmination of this research effort is a multifaceted contribution. We distill insights from a diverse range of access control schemes, shedding light on their nuances and effectiveness. Our analysis extends to evaluating the performance metrics vital for ensuring robust access control. Through this research, we also pinpoint critical research gaps within traditional access control methods, creating a foundation for innovative approaches to address the evolving challenges within edge computing and blockchain environments. In conclusion, this research venture paves the way for secure data sharing in edge computing and blockchain by offering a thorough examination of access control. The findings from this study are anticipated to guide future developments in access control techniques and facilitate the evolution of secure, privacy-conscious, and efficient data sharing practices in the dynamic landscape of digital technology.
{"title":"Study of Access Control Techniques on the Blockchain-Enabled Secure\u0000Data Sharing Scheme in Edge Computing","authors":"Neha Mathur, Shweta Sinha, Rajesh Kumar Tyagi, Nishtha Jatana","doi":"10.2174/0126662558276547231213075235","DOIUrl":"https://doi.org/10.2174/0126662558276547231213075235","url":null,"abstract":"\u0000\u0000The pervasive adoption of edge computing is reshaping real-time big\u0000data analysis, smart city management, intelligent transportation, and various other domains. Its\u0000appeal lies in its distributed nature, decentralization, low latency, mobile support, and spatial\u0000awareness. However, this ubiquity exposes data to security threats, jeopardizing privacy and\u0000integrity. Consequently, access control assumes paramount importance in securing data sharing\u0000within edge computing and blockchain technologies.\u0000\u0000\u0000\u0000This research addresses this critical issue by conducting a comprehensive study on\u0000access control techniques within the context of edge computing and blockchain for secure data\u0000sharing. Our methodology commences with an exhaustive review of relevant articles, aiming to\u0000identify and expound upon gaps in existing research. Subsequently, we perform a meticulous\u0000analysis of access control methods, mechanisms, and performance metrics, seeking to establish\u0000a holistic understanding of the landscape.\u0000\u0000\u0000\u0000The culmination of this research effort is a multifaceted contribution. We distill insights from a diverse range of access control schemes, shedding light on their nuances and effectiveness. Our analysis extends to evaluating the performance metrics vital for ensuring robust access control. Through this research, we also pinpoint critical research gaps within traditional access control methods, creating a foundation for innovative approaches to address the\u0000evolving challenges within edge computing and blockchain environments.\u0000\u0000\u0000\u0000In conclusion, this research venture paves the way for secure data sharing in edge\u0000computing and blockchain by offering a thorough examination of access control. The findings\u0000from this study are anticipated to guide future developments in access control techniques and\u0000facilitate the evolution of secure, privacy-conscious, and efficient data sharing practices in the\u0000dynamic landscape of digital technology.\u0000","PeriodicalId":506582,"journal":{"name":"Recent Advances in Computer Science and Communications","volume":"32 10","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139444860","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：