Deep Graph neural network-based spammer detection under the perspective of heterogeneous cyberspace

The containers play a crucial role in the cloud environment during application deployment as it shares same OS kernel. It reduces resource requirements and start-up time for deploying applications by an individual organizations or users. Even though containers provide light-weight virtualization, it generates a security bottleneck for the number of dedicated resources, libraries, and applications since the container isolation is comparatively weak to the legacy VMs. In the general architec ture of container, attackers can perform privilege escalation by exploiting the kernel vulnerabilities to gain the root privilege and leaks the critical information of a system. To address the present security concerns in the container, a better security based solution is essential. In this work, an extensive analysis is performed to predict the various existing access control mechanisms used for security purposes and the challenges encountered during the architecture modeling. Some use cases are considered to ensure the fulfilment of security requirements such as container protection, inter-container protection, and host protection, and it needs to provide both software and hardware solutions. This work also includes the research problems, research gaps, and further research extensions to provide security to the containers.

Section: Discussionmentioning

confidence: 99%

Container Security: An Extensive Roadmap

Subramanian¹,

Honnavalli²,

Shylaja³

2021

“…Sachin Shetty and Y. S. Rao proposed an SVMbased machine learning approach to identify Parkinson's disease based on gait analysis [25]. The Support vector machine (SVM) classifier built on a Gaussian radial basis function kernel achieves an overall precision of 83.33 %, a strong identification rate for Parkinson's disease of 74.99 %, and poor false positive findings of 16.66 % [36][37][38][39][40].…”

Section: Techniques For Machine Learning In Parkinson's Diseasementioning

confidence: 99%

A Study on Brain Tumor and Parkinson’s Disease Diagnosis and Detection using Deep Learning

Warjurkar¹,

Ridhorkar²

2021

Consider the possibility that we live in an area far from a doctor, or that we may not have enough resources to pay the hospital cost, or that we may not have enough time to take off work. The use of advanced computers to diagnose diseases will be lifesaving in such situations. Scientists have developed a number of artificially intelligent diagnostic algorithms for illnesses such as cancer, lung disease and Parkinson's disease. Deep learning employs massive artificial neural network layers of interlinked nodes that can reorganize themselves in response to updated data. This approach enables machines to self-learn without the need for assistance from humans. The emphasis of this article is on current developments in machine learning that have had major effects on identification for the detection of a variety of illnesses, such as brain tumor segmentation. Human-assisted manual categorization may lead to erroneous prediction and diagnosis, thus one of the most important and a useful technique is brain tumor segmentation tasks in medical image processing that are difficult. Furthermore, it is a difficult challenge because there is a vast volume of data to assist. Since brain tumors have such a wide range of appearances and since tumor and normal tissues are so close, extracting tumor regions from photographs becomes difficult. The advancement of clinical decision systems of support necessitates the identification and recognition of the appropriate biomarkers in relation to specific health problems. It has been established that handwriting deficiency is proportionate to the severity of the situation of individuals' Parkinson's disease (PD).

“…As some systems provide service to only particular type of disability and some systems provide service to differently abled people but to help them to communicate with other but, the SAARD not only help to recognize activities to differently abled people by giving information about surrounding objects but also help them to know surrounding sounds so that they don't get into accidents while crossing or walking on roads. It also helps them to detect the obstacle on their way if any and alert them, so they won't fall on ground and hurt themselves and can go anywhere without depending on others [28][29][30].…”

Section: Figure 4 Saard With Approximately 70% Efficiencymentioning

confidence: 99%

Smart Assistive Activity Recognition Device for Differently Abled People Based on Machine Learning -SAARD

Agarkhed¹,

Tahreem²

2021

Science and Technology have made human life addictive to comfort, yet concurrently there exists an oppressed gathering of individuals who are battling for tracking down a creative way that can make their life easier for them. After concentrating and highlighting the problems faced by the differently abled people like blind and deaf, solving it through a device alone is a very hard task. A ton of exploration has been done on every issue and arrangements have been proposed independently. Objective of the smart assistive device SAARD (Smart Assistive Activity Recognition Device for Differently Abled People) is to recognize activity for differently abled people so; they feel confident and independent by helping them to know objects surrounding them. The Proposed device SAARD help the differently abled people by taking images and give the output in form of audio. Along with that it also detects obstacles and surrounding sound which alert them.