Edge devices need intelligent data processing which requires both quick response times and reduced energy usage for smartphones and IoT sensors and wearable systems and drones and embedded boards. Deep learning models deliver accurate results, yet they require extensive computational power and memory resources, which makes them impractical for use on edge devices with limited computing capabilities. The research paper investigates lightweight deep learning frameworks which researchers designed to function effectively in edge computing environments. The research examines model compression methods together with pruning, quantization, knowledge distillation, and the analysis of MobileNet and SqueezeNet and ShuffleNet and TinyML-based networks. The experimental results show that lightweight models can achieve competitive accuracy while using less memory and running time and processing power. The results show that optimized architectures enable real-time AI applications on edge platforms while extending battery life and achieving reasonable prediction performance. The research develops AI solutions which are efficient and scalable and can be deployed in intelligent environments. The current demand for intelligent data processing at the device level has increased because of the expanding use of edge devices which include smartphones and IoT sensors and wearable systems and drones and embedded boards. Edge environments need their systems to make decisions instantly while maintaining low response times and protecting user data and using less energy for operations. The conventional deep learning models achieve high accuracy and strong performance but their massive computational needs and their requirement for extensive memory and their substantial energy demands create difficulties for using them on edge hardware with limited resources. Researchers established a relationship between computational efficiency and real-world edge applications which requires both components to be balanced. The research shows that baseline and optimized architectures can achieve model size reductions and floating-point operation decreases while maintaining high accuracy levels. The optimization framework develops through which faster inference occurs because models need less memory space and their energy consumption becomes lower thus making them suitable for use on devices with limited resources. The research shows that MobileNet, SqueezeNet and ShuffleNet architectures have become more essential because of their efficient design requirements. The combination of these networks with pruning quantization and knowledge distillation techniques creates a robust system which enables intelligent processing to occur on edge hardware. The experimental results demonstrate that lightweight models enable real-time analytics across multiple applications including smart surveillance and wearable health monitoring and autonomous navigation and industrial automation.
The collection includes Edge Computing, Edge Artificial Intelligence (Edge AI), Lightweight Deep Learning, Deep Neural Networks (DNN), Convolutional Neural Networks (CNN), Model Compression Techniques
International Journal of Trend in Scientific Research and Development - IJTSRD having
online ISSN 2456-6470. IJTSRD is a leading Open Access, Peer-Reviewed International
Journal which provides rapid publication of your research articles and aims to promote
the theory and practice along with knowledge sharing between researchers, developers,
engineers, students, and practitioners working in and around the world in many areas
like Sciences, Technology, Innovation, Engineering, Agriculture, Management and
many more and it is recommended by all Universities, review articles and short communications
in all subjects. IJTSRD running an International Journal who are proving quality
publication of peer reviewed and refereed international journals from diverse fields
that emphasizes new research, development and their applications. IJTSRD provides
an online access to exchange your research work, technical notes & surveying results
among professionals throughout the world in e-journals. IJTSRD is a fastest growing
and dynamic professional organization. The aim of this organization is to provide
access not only to world class research resources, but through its professionals
aim to bring in a significant transformation in the real of open access journals
and online publishing.