Advanced Topics on Scalable Deployment of Machine Learning and Drone-Based Search and Rescue

On Thursday, July 23, 2020 at 1:00 p.m., Dalia Hanna and Mujahid Sultan will be presenting “Advanced Topics on Scalable Deployment of Machine Learning and Drone-Based Search and Rescue”.

Day & Time: Thursday, July 23, 2020
1:00 p.m. – 4:00 p.m.

Speakers: Dalia Hanna, Mujahid Sultan

Organizers: IEEE Toronto WIE, IEEE IM/RA, Ryerson CS Graduate Student Council, IEEE Ryerson Computational Intelligence Chapter, Ryerson CSCU

Location: Virtual – Zoom

Contact: Ayda Naserialiabadi

Title: Factors affecting the Automation of the Search and Rescue Operations: An Algorithm on Finding Missing Lost Persons Living with Dementia

Abstract: Unmanned Aerial Vehicles (UAV) are now used in many applications. The focus in this presentation is on their use in public safety, specifically in search and rescue (SAR) operations involving lost persons living with dementia (LPLWD). When it comes to saving lives, there are many human factors associated with UAV operations that impact the performance of expert human SAR teams that could be improved through forms of automation. These include familiarity with the search location, tasks associated with piloting and search/flight management during SAR operations.  A LPLWD may not be interested in assisting in their own rescue as they may not know they are lost. As such, it has been observed that they tend to keep walking until they are faced with an obstacle that bars their further progress. The approach presented in this research work focuses on developing a people finding algorithm to identify higher probability locations where an LPLWD might be found, through informed, behavior-based analysis of the search location; then, developing an algorithm to fly a UAV to the vicinity of these higher probability locations.  The algorithm was tested and validated through field testing. The results from both the data collection process and the field tests indicated that there are efficiencies in using the drone, which enhances the probability of finding the lost person alive.  An informed cleaning process involving both manual and ‘R’-automated approaches to scrub and augment the data–adding any missing values in the dataset, helped in understanding the behaviour of the lost person and in determining what significant variables enhanced their survivability. Linear regression was utilized to acquire the correlation among the numeric values in the database. The analysis indicated that there was no significant correlation among the independent variables; however, the data indicated that the wanderer tended to be found closer to where they left or were last seen. Logistic regression was used to investigate the survivability using three classification models. Finally, a framework is presented considering all the factors form the field tests and data analysis.

Title: How to build and deploy machine learning models in the scalable cloud 

Abstract: Machine learning model development is a skill taught at schools and is a good skill to have but where most of the student’s lake is how to serve these models to the clients. How to scale. Make sure that the server does not die if it gets a million hits in a second. How to build security around it.

Agenda: Interested students who want to build along with me, can bring their laptop with MobaXterm installed and we can do the following together.

  1. login to a cloud environment (I will provide the cloud login credentials during the presentation)
  2. create a virtual environment for development
  3. build a semantic search engineby pulling libraries from the net
  4. pick a visualization and presentation method from D3JS
  5. develop an application using MVC pattern like the flask
  6. wrap the application in a docker container
  7. install scalable web engine like NGINX
  8. host it to the cloud (azure)
  9. provide secure access with a username and password to anyone on the internet

This presentation will expose the tools required to build scalable machine learning applications in the cloud.

Registration: Please visit to register.


Dalia Hanna
Factors affecting the Automation of the Search and Rescue Operations: An Algorithm on Finding Missing Lost Persons Living with Dementia

Dalia Hanna is a PhD Candidate in the Department of Computer Science, Ryerson University. She is a member of Ryerson’s Network-Centric Applied Research Lab, a multidisciplinary Computational Public Safety-focused research lab. She has a B.Sc. in Electronics and Communication Engineering and M.Sc. in Instructional Design and Technology with a specialization in Online Learning. Dalia is also a certified project management professional (PMP ® ) and a certified facilitator. Her research interest in utilizing technology tools for public safety, search and rescue, and emergency management operations. . Dalia authored several research papers and presented in national and international conferences.

Mujahid Sultan
Topic: Factors affecting the Automation of the Search and Rescue Operations: An Algorithm on Finding Missing Lost Persons Living with Dementia

Mujahid Sultan is a senior computer scientist and enterprise architect with vast experience in machine learning, pattern recognition, deep learning, NLP, text synthesis, transcription, time-series forecasting and cloud-native developments (Python, microservices, APIs, Docker, Kubernetes). His current research focus: a) working to develop a robust clustering method with mathematical proofs b) improving learning from imbalanced data on graph-based deep learning backends (TensorFlow, Torch and CNTK), and c) building Machine Learning based dynamic SDN controllers.

He has authored in high impact journals in fields of Machine Learning, Artificial Intelligence, Data Visualization, Genetics and Drug Discovery for Cancer, Requirements Engineering and Enterprise Architecture. His publications can be found at

Areas of Expertise include: Regression, Clustering, Classification, Deep Learning, Convolutional and Recurrent Neural Networks (LSTMs), Natural Language Processing (NLP), Self-Organizing Maps (SOM), Topic Modeling and Parallel Processing. Expert in info visualization using matlab, matplotlib, D3js and plotly.

Skills: Full-stack development: (Angular+Flask+Docker); Python: (Scikit-Learn, Keras, TensorFlow, NLTK, Spacy, NumPy, Matplotlib, SpaCy to name a few); MATLAB: (toolboxes: statistics, microeconomics, parallel processing, bioinformatics to name a few).

Platform experience: Docker Containers and Kubernetes on AWS, Azure/Azure Stack and Google Cloud Platform. PaaS/IaaS: (AWS: (Elastic Beanstalk, Lambda, Poly, Sage-Maker), Azure ML, and Heroku).

IEEE Virtual Presentation Poster