IEEE Consumer Electronics Magazine - September/October 2020 - 78

Special Section on Machine Learning for End Consumers

by the three articles selected for this special section from both application and implementation
perspectives.
In "Juice Recipe Recommendation System
Using Machine Learning in Mobile Edge Computing Environment," Juyong Lee and Jihoon Lee
propose a smart recommendation system for providing user-specific recipe information based on
mobile edge computing (MEC) and machine intelligence to facilitate secure user-specific health
services. MEC is a new type of platform that provides computing capabilities within a wireless
access network in close proximity to mobile users.
While existing systems transfer users' health
information to a remote cloud server for processing, the proposed system utilizes critical user
information only within the local MEC server. It
then transfers basic recipe information, without
any user information, to the external cloud server.
At the same time, their smart recipe recommendation system delegates the high computational burden to a MEC server.
Continuing the theme of preserving privacy in
a personal healthcare setting is B. S. Vidyalakshmi
and colleagues' article titled "Health Mentions on
Twitter: A Case Study to Identify Privacy Leaks."
The article provides an in-depth analysis of the
health condition mentions in tweets using ML
classifiers. Since health mentions can be used in
different contexts, e.g., as a joke, in a news article
link, or a genuine disclosure of health condition
suffered by the user, it is important to understand
the contexts of the tweets before analyzing
them. The authors address this issue by categorizing the tweets based on context. The article
describes experimentation involving several
ML classifiers. The experiments show that a combined classifier is better than using the classifiers
individually. A key finding is that each health mention has differing disclosure rates, affecting

privacy leaks differently and peaking in disclosure
rate at different time of the day. Another
key finding is that personal privacy leaks and secondary privacy leaks are affected differently by
each health mention.
The article titled "Optimizing Image Compression with Deep Super-Resolution Techniques" by
bastien Hamis and colleagues introduces a low
Se
and very low bitrate image compression scheme.
Their method combines a better portable graphics
(BPG) codec with deep superresolution techniques, namely SRGAN (superresolution generative
adversarial networks) and ESRGAN (enhanced
superresolution generative adversarial networks).
BPG format is well-suited for critical CE fields, such
as IoT or consumer healthcare. When coupled
with deep superresolution techniques, the scientific contribution to CE is far-reaching. Indeed, the
authors' experimental results demonstrate the
possibility of storing and transmitting images at
low bitrate and resolution, while preserving the
quality of the visual content. This means that there
is further potential of this work to influence CE
much more broadly. For example, the authors
have identified an extension to adapt their method
to perceptually driven models that can be applied
to achieve very good results for automatic object
recognition.
We would like to thank all the contributing
authors for their excellent contributions, as well
as the reviewers for their help and efforts.
Alvis Fong is currently a Computer Scientist with
Western Michigan University, Kalamazoo, MI, USA.
Contact him at alvis.fong@wmich.edu.
Muhammad Usman is currently the Head of Computer Science Department, Shaheed Zulfikar Ali Bhutto
Institute of Science and Technology, Islamabad,
Pakistan. Contact him at dr.usman@szabist-isb.edu.pk

IEEE Consumer Electronics Magazine

IEEE Consumer Electronics Magazine - September/October 2020

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - September/October 2020