Dr. El Sayed Mahmoud
Professor of Applied Computing at Sheridan  

 

Research

El Sayed’s research focus is to develop robust machine learning systems that provide innovative solutions for health and business applications including mobile computing, cloud computing and social networks. His experience in Engineering, Computer Science, Artificial Intelligence and Business for 25 years significantly support the breadth and depth of his multidisciplinary research projects.

Research project statistics

AI Fields

#projects

Project Titles

Grants

#students

Notes

Human-centered Approaches to AI

2

· Face-based identity Search without Racial Bias

· FaceMatch

Mitacs/Accelerate

 

OCE-CVTA

1

 

 

2

Completed

 

 

Completed

Robotics

3

· Web-based Process tracker 1

· Web-based Process tracker 2

· Automatic duct assembly

1-year Capstone

 

OCE-CVTA

 

NSEC/ARD

4

3

Completed

Completed

 

in process

Healthcare

2

· Arrhythmia Diagnostics Service Optimization

· Immersive Environments for People Living with Dementia

NSERC/EI

SSHRC/ New Frontiers in Research Fund: Exploration

13

Active / conducting a pilot study

Application in process/ LOI has been submitted

Equipment Recognition

2

· Smart Equipment detector 1

· Smart Equipment detector2

NSERC/Engage

OCE-CVTA

6

2



Food and health

2

· Exploring Identifiers of Research Articles Related to Food and Disease Using Artificial Intelligence

· AUTOMATIC EXTRACTION OF USEFUL INFORMATION FROM FOOD-HEALTH ARTICLES RELATED TO DIABETES, CARDIOVASCULAR DISEASE AND CANCER

Research Platforms and Portals grant/ Compute Canada


Research Platforms and Portals grant/ Compute Canada

1



1







CS Thesis of Marco Ross- done



CS Thesis of Ken Suong- done








Self-driving cars

1

DETECTION OF DISTRACTED PEDESTRIANS USING CONVOLUTIONAL NEURAL NETWORKS

NA

1

CS Thesis of Igor- in progress

Business

4

Smart Pattern Detector for Clothing

Customer intent detector that learns

Automatic Driver Background Check

Automatic Safety-Certification Tracker

NSERC/Engage

NSERC/Engage

NSERC/Engage

NSERC/Engage

4

4

4

3

Completed

Completed

Completed

Active

Readship distribution of El Sayed Mahmoud’s publications

Education

·     BSc. in Computer Engineering, Ain Shams University

·     MSc. In Computer Science, University of Guelph

·     PhD in Computer Science, University of Guelph

El Sayed’s Research Areas and applications

·     Artificial Intelligence, Machine Learning, Data Analytics, Big data, Cloud computing

·     Evolutionary Algorithms (Genetic Algorithms and Multi-Objective Genetic Algorithms)

·     Health (examples from current research: food article analytics for health promotion)

·     Business (examples from my current research: identify customer intents, equipment recognition, fashion trends and design, profitability)

·     Student engagement (examples: coding fluency, learning through play)

·     All meaningful combinations of the above areas and applications

El Sayed on social media

An interview with El Sayed Mahmoud

An article distinguishes El Sayed's PhD research

LinkedIn: www.linkedin.com/in/el-sayed-m-84676822

Twitter: El Sayed mahmoud @elsayed939

Email: elsayed.mahmoud@sheridancollege.ca


Primary research interests

My primary research interests are in the area of machine learning. The robustness and scalability of machine learning algorithms are limited by the amount of data available and the performance measures used to select the best model (or a combination of models) built. I aim to alleviate these limitations by investigating analytic models for machine-learning methods. This will allow me to derive theoretical guidelines to promote the robustness of these methods while reducing their computational cost. I have derived a novel analytic model for a multiple classifier methodology that models the classification problem. This model breaks down the fitness of this methodology into three measurable components. The three components are measured during the construction of the multi-classifier based system to optimize its fitness and the computational cost.  

Current research projects

My current research focuses on developing robust and scalable machine learning methodologies that provide innovative solutions for health and business applications (including mobile computing, cloud computing, social network, wearables, and augmented reality). Examples of current projects are:

 

1.    Automatic Safety-Certification Tracker ( May 2019 – Dec 2019) funded by NSERC (industry partner : Coolbreeze)

Safety-related certification tracking ensures workplace safety while fostering professional

development which, in turn, improves production quality and ultimately supports the Canadian

economy. However, currently the certification tracking is conducted manually,

which is not efficient. This project aims to develop algorithms to automatically collect, track and analyze technicians’ certifications and licenses for improving the process efficiency, accessibility and reliability while reducing costs associated with the manual tracking process.

2.    Arrhythmia Diagnostics Service Optimization with Cloud Data Analytics and Wearable Devices (2017-2020)- funded by NSERC

This project aims to benefit patients and physicians while improving health outcomes and reducing health care costs. It involves: (1) creating data analytics and visualization tools; (2) designing and implementing algorithms to provide actionable insights for patients, cardiologists and service providers such as nurses; and (3) exploring how wearable devices may be used to collect medical information and provide more accurate insights into treatments and advice. We explore advances in wearable devices for healthcare to select next-generation arrhythmia detection technologies and explore their integration into other in-home patient care systems.

 

 

Previous research projects

3.    Face-based identity Search without Racial Bias (Feb 2019 – Oct 2019)

This project aims to develop new bio-metric face recognition model without racial bias. More than 3 million of photos for ninety thousand identities from seven different races have been annotated for building this new unbiased model for identity search based on face photo. This work is a continuation for the FaceMatch project funded by the OCE.

4.    Automatic Driver Background Check (Oct 2018 – April 2019) Funded by NSERC

The Uber business model facilitates providing various services online. However, checking the car drivers’ background is significant, time consuming and expensive process of this model. The automation of this process would improve the Uber model’s reliability while reducing its cost. It would also increase the safety of the clients. This research project investigates the usage of the available machine learning and data analytics platforms, user generated data, and public information on the internet to check the background of potential drivers automatically. The project will focus on building an efficient system that identified opportunities rather than exclude people.

 

5.    Food for Health Promotion (2016 -2019) funded by Compute Canada
This project aims to understand how variations in writing styles and the flexibility of text-mining methods control the methods’ performance in extracting useful information from articles about food and health. The project deliverables are a database and query tools that will change the way we search and collect information from scientific publications and the way we analyze this information for further applications.

 

6.    FaceMatch (Sep – Dec 2018) Funded by OCE

The current face recognition solutions are racially biased. This bias affects the accuracy of recognizing white faces negatively. Additionally, the current solutions don’t meet the security and privacy requirements of the local police. This project aims to develop new bio-metric face recognition model that is able to fix this biased facial recognition technology and to allow implementing the security and privacy requirements.

 

7.    Web-based process tracker (May – Sep 2018) funded by OCE

The aim of this project is to develop a robust Web Application for Lynch Fluid Controls Inc. The application automatically tracks and evaluates the profitability of standard products on a real time basis. The application provides all services online and maintains a secure database for process information collected. Additionally, it evaluates the profitability of Standard Products on a real time basis and generates reports/notifications that support management and planning teams

8.    Smart Pattern Detector for Clothing (Feb-Aug 2018) funded by NSERC
Keeping track of clothing trends by professional designers requires significant time and effort, which makes clothing design process expensive and slow. However, automatic identification of these trends can assist the clothing design companies to make the process more efficient and cheaper. This research investigated the performance of the available various pattern recognition algorithms such as neural networks and support vector machine in recognizing clothing patterns. This investigation produced guidelines to identify the unique characteristics of clothing. These guidelines are useful for various clothing applications such as automatic identification of articles and their brands and helping the clothing designer to create new styles based on the common patterns used in the successful styles.

9.    Smart Equipment detector (Oct 2017-April 2018) funded by NSERC
Mobile apps enable companies to place maintenance/repair service requests instantly from anywhere and track their progress with real-time updates. However, automatic identification of equipment with an issue in their facilities is a much more valuable piece of information for service efficiency. This project build two deep learning models to identify equipment and its category based on its photo. This identification facilitates filling out the maintenance/repair requests.    

10. Customer intent detector that learns (Sep- April 2017) funded by NSERC
Root-cause analysis of user-generated content in social media helps to answer why customers dislike a product or service. However, identifying customer intent after the root-cause analysis is a much more valuable piece of information in terms of marketing and customer service. This project developed an add-on service to the root-cause analysis platform of Kaypok Inc. The service identifies customer intent in user-generated content. Natural language processing and support vector machines have bee used to predict customer intent based on analyzing their user-generated content in social media. For example, the message: “Hi, my smartphone is broken, guess I need a new one…” indicates an intent to buy. The algorithm targeted various intents and determines the fingerprints (signs) of each one. The applied research undertaken during this project added useful features to Kaypok’s product that helped the company to attract new customers in Canada and Australia. In addition, it has engage four students in addressing real-world problems.

 

11. A smart web application that empowers women (2016-2017) (industry partner
The deliverable of this capstone project was a robust Web Application that models a web-based club for Zonta club of Mississauga. The application provides all services online, tracks the website traffic and maintains a secure database for club transactions including internal documents. Additionally, it analyzes the archived transactions for admin and statistical purposes that support future resource planning.

 

Future research

My future research plan includes deriving new analytic models for new multimodal methodologies, and investigating the parallelism of implementing these methodologies to improve the efficiency of the methodologies. I will use these analytic models as guidelines to develop robust and scalable techniques that model classification and regression problems and, in turn, investigate the possibility of applying the robust and scalable techniques for health and business applications. These applications may include extracting useful information from health-food scientific papers, combining wearables, augmented reality, mobile computing and cloud computing for online clinical care, and industry classification.

 


Scholarly talks

1.    I was invited for a scholar interview as a computer Scientist on the international Future Technology Conference 2017 held at Vancouver.  Click HERE to watch the interview

2.    I was invited as a speaker to give a talk about using the Artificial Intelligence for Food Safety on the food safety seminars, April 2019.

3.    Presented the lessons learned from supervising a thesis student to the faculty of Applied Science and technology, Sheridan on the professional development conference, 2019   

Publications (Note: the asterisk (*) indicates the presenter of the paper)

Journal articles

1.    M. Ross, E. Mahmoud, E. Abdaal. Exploring Identifiers of Research Articles Related to Food and Disease Using Artificial Intelligence, the International Journal of Advanced Computer Science and Applications (IJACSA), Volume 9 Issue 11, 2018

2.    B. Eram, J. Patel, A. Satvedi, R. Sneyd, E. Mahmoud. Video-call Platforms for Online Healthcare, International Journal of Advances in Science Engineering and Technology, Vol-6, Iss-3, Spl. Issue-2, pp 65-64,2018.

3.    E. Mahmoud, An Engagement Strategy for Teaching Computing Concepts, the International Journal for Digital Society (IJDS), Volume 8, Issue 3, ISSN: 2040-2570, 2017

Conference abstracts

4.    E. Mahmoud*, R. Hurtadomolero. Smart pattern detector for clothing, Sheridan Creates Conference, 2018.

5.    E. Mahmoud*, R. Sneyd. Smart equipment detector, Sheridan Creates Conference, 2018

6.    E. Mahmoud*, M. Ross. Food-health Article Detector: A way to reduce health risks. Sheridan Creates conference, 2018

7.    E. Mahmoud*. Experience-based Analogical Model for Learning Computing Concepts: A Kinesthetic Learning Strategy to Help Tactile Learners, EduTeach International Conference, Toronto, 2016

Conference Papers

8.    B. Eram, J. Patel, A. Satvedi, R. Sneyd, E. Mahmoud*. Video-call Platforms for Online Healthcare, Proceedings of 121st ISERD International Conference, Toronto, Canada

9.    S. Akulick, E. Mahmoud*. Intent Detection through Text Mining and Analysis, Future Technologies Conference (FTC), Vancouver, Canada, 2017

10. E. Mahmoud*. Learning Catalyst: A Play-based Strategy for Teaching Applied Computing Concepts, The Canada International Conference on Education (CICE), Toronto, 2017.

11. E. Dancy*, E. Mahmoud. Practice and Refactoring Log: A Reflection Based Learning Strategy to Improve The Fluency of Computing Students in Writing Computer Programs, EduTeach International Conference, Toronto, 2016.

12. E. Mahmoud*, D. Calvert. A robust system for distributed data mining and preserving-privacy, SmartData, pp 129-138, 2013.

13. E. Mahmoud* and D. Calvert. "Regression of Representative Keys: A simple Learning Approach," Intelligent Engineering Systems through Artificial Neural Networks, vol. 19, pp. 463-470, 2009.

14. E. Mahmoud* and D. Calvert, "Auto-calibration of support vector machines for detecting disease outbreaks." in 2009 IEEE Toronto International Conference - TIC-STH'09, pp. 112-117, 2009.

15. E. Mahmoud. "Computers and TAs", Teaching Assistant Advisory Council (TAAC) Newsletter, the University of Guelph, 2009.

16. E. Mahmoud* and D. Calvert, "Comparing Performance of Back Propagation Networks and Support Vector Machines in Detecting Disease Outbreaks," Proceeding of Intelligent Engineering Systems through Artificial Neural Networks, vol. 18, pp. 245-252, 2008.

17. E. Mahmoud and D. Calvert*, "Comparing Syndromic Surveillances using Two Aspects: Emergency and Telehealth Data Sources," Advances in Disease Surveillance, vol. 4, pp. 102, 2007.

18. E. Mahmoud and D. Stacey*, "Identifying Syndromic Fingerprints in Reason Fields in Emergency Department or Telehealth Records using N-grams for Similarity Analysis,"  Advances in Disease Surveillance, vol. 4, pp. 55, 2007.


Teaching

·     Successfully taught introductory and upper-level undergraduate courses in software engineering, web development, data structures and algorithms, distributed mobility (parallel programming and cloud computing), Business Intelligence and data mining including Big data.

·     Led the development of a new computer science degree with five specilizations since Feb 2018

·     Designed and implemented fourteen in-class activities that help students to learn crucial concepts. 

·     Developed five courses in Parallel Programming, Internet of things ,Visualization, Distributed mobility and Business Intelligenec and Data Mining

·     Carried out an additional responsibility as a course leader of two mobile computing degree courses titled distributed mobility (cloud computing) and Business intelligence and data mining

 

Service

·     A member of the Sheridan Senate representing Faculty of Applied Science and Technology (FAST)

·     A member of the local academic counsel and Scholarship, Research and Creativity Committees

·     A member of the program quality assurance committee of FAST

·     A reviewer of NSERC and International Journal of Advanced Computer Science and Application

·     Reviewed scientific papers for the Artificial Neural Networks in engineering conference

·     Served as a session chair in the FTC 2017 international conference held in Vancouver and the CICE international conference held In Toronto in 2017.

·     Participated in the walkathon for pediatric autism spectrum disorders

·     Participated in Heart & Stroke campaign for the years 2018 and 2019