You are on page 1of 14

Machine Learning

for Multimedia and


Edge Information
Processing

BY: MANAL ALI ALGADABY


Introduction
The advancements and progress in artificial intelligence (AI) and machine learning, and the
numerous availabilities of mobile devices and Internet technologies together with the growing
focus on multimedia data sources and information processing have led to the emergence of
new paradigms for multimedia and edge AI information processing, particularly for urban and
smart city environments. Compared to cloud information processing approaches where the
data are collected and sent to a centralized server for information processing, the edge
information processing paradigm distributes the tasks to multiple devices which are close to
the data source. Edge information processing techniques and approaches are well suited to
match current technologies for Internet of Things (IoT) and autonomous systems, although
there are many challenges which remain to be addressed.
Introduction
Edge Computing (EC) is a new architecture that
extends Cloud Computing (CC) services closer to
data sources. EC combined with Deep Learning
(DL) is a promising technology and is widely used
in several applications.
Problem
This research addresses the problem of efficient scheduling
of real-time multimedia tasks in edge devices. Edge devices,
such as smartphones, IoT devices, and edge servers, have
limited computational resources and energy constraints.
Moreover, real-time multimedia tasks, such as video
streaming and image processing, require fast and high-
quality execution. Therefore, there is a need for a smart
scheduling mechanism that can distribute resources
effectively and ensure real-time performance for multimedia
tasks in edge devices.
Solution
The proposed solution is based on
federated learning, which leverages
the computational capabilities of
edge devices while preserving data
privacy.
The importance
and objectives.
The importance
Research and study on this topic has
contributed to the development of new
and innovative technologies to improve
computing performance at the edges and
achieve more efficient and secure
applications in areas such as smart
industry, healthcare, smart cities, and
others.
objectives.
Edge Cloud Computing refers to the deployment of cloud services in strategic
locations at the edge of the network infrastructure. The main objectives and
goals of Edge Cloud Computing are to reduce data latency and maintain high
availability. By bringing cloud services closer to the data source, Edge Cloud
Computing aims to improve response time, handle high data volumes, ensure
privacy, cater to remote areas, optimize costs, and enable autonomous
operations
objectives.

Improved data security Reduced latency: Bandwidth optimization


Edge computing reduces the amount of data By processing data closer to the source, edge Edge computing reduces the amount of data that
transmitted and processed in the cloud, keeping computing minimizes the time it takes for data to needs to be transmitted over the network,
sensitive data on user devices and reducing the travel to the cloud and back, enabling real-time or optimizing bandwidth usage and reducing
risk of data compromise near-real-time applications congestion

Enhanced reliability Scalability Cost Reduction


Edge computing allows for the deployment of
With edge computing, applications can continue Edge Computing reduces data transmission costs
compute and storage resources at multiple edge
to function even if there is a disruption in the by processing data at the edge, closer to the
locations, enabling scalability and supporting the
network connection to the cloud, ensuring source, thereby minimizing bandwidth usage and
growing number of connected devices and data
uninterrupted service saving on communication costs
sources
newly
implemented
models
Here are three models implemented in the
field of edge computing:

OpenNESS Akraino MobiledgeX


• OpenNESS is an open source • Akraino is an open source project • MobiledgeX aims to provide an edge
framework that aims to facilitate the aimed at accelerating the development computing infrastructure that allows
development and deployment of edge and deployment of edge computing. edge computing applications to be
computing applications. easily developed and deployed.
• Akraino provides a set of models, tools
• OpenNESS provides application and projects based on edge computing. • MobiledgeX provides application
programming interfaces (APIs) and programming interfaces (APIs) and
tools to make it easier to develop and • Akraino aims to achieve compatibility tools to make it easier to develop and
manage applications on edge and integration between hardware, manage applications on edge
computing infrastructure. software and services used in edge computing infrastructure.
computing.
• OpenNESS supports many different use • MobiledgeX allows developers to
cases such as artificial intelligence, access edge computing resources
industrial automation, cloud gaming, deployed in the mobile
and augmented reality. communications network.
performance of models
01 02
Federated Learning
Distributed Machine Learning
The learning process is distributed across multiple local Machine learning models are trained on local data for each
machines in an edge network, where machine learning device in the edge network, and then important
models are trained on local data and results are shared information and collected results are shared between
between machines.. devices

03 04 05
Ensemble Learning Transfer Learning Collaborative Learning
A group of independent machine learning models is Machine learning models are used that are pre- where knowledge and results are shared between
used and their results are collected to obtain trained on big data in the remote cloud, devices to improve model performance and
better and more accurate predictions. and then modified and improved using enhance predictive ability.
local data in the edge network
challenges

Complexity Dealing with


Lack of resources
of the topic big data
The field of machine learning for difficulty in obtaining the resources With the increasing volume of data
multimedia and edge information necessary for research and experiments related to multimedia and information
processing is a complex and diverse in this field. at the edges, it becomes difficult to
field. store and process this data.
May need access to large data sets and
It requires a deep understanding of powerful computing hardware to carry May need to develop techniques and
advanced techniques in machine out experiments and tests. tools to handle and analyze big
learning, multimedia processing, and data efficiently.
edge computing.
References

• Zhiyu Wang, Mohammad Goudarzi, Mingming Gong, Rajkumar Buyya,(2023) Deep Reinforcement Learning-based scheduling for optimizing system
load and response time in edge and fog computing environments,Future Generation Computer Systems,

• Cheng, X., et al. (2022). Providing Location Information at Edge Networks: A Federated Learning-Based Approach.

• Al-Rakhami, M., et al. (2023). Challenges and Techniques in Distributed Machine Learning for Edge Computing.

• Wang, Y., et al. (2023). Distributed Machine Learning Techniques for Edge Computing: A Comprehensive Survey.

• Verbraeken, R., et al. (2023). Challenges and Opportunities in Distributed Machine Learning for Edge Computing

• Zhou, Y., et al. (2023). Distributed Machine Learning in Edge Computing: Challenges and Techniques

• Dianlei Xu, et al. (2023). Distributed Machine Learning in Edge Computing: Challenges and Techniques

You might also like