Skip to content
  • Nsirc
  • News
  • 2023
  • Explainable Machine Learning a Reality for Industry

Explainable Machine Learning a Reality for Industry

Tue, 14 March, 2023

"Hesham’s PhD work demonstrated his ability to come to a detailed understanding of explainable AI."

Said Dr Kai Yang, a technology expert from TWI Ltd, and one of Hesham's PhD Industrial Supervisors. "He was responsible for proposing and developing the explanation frameworks for industrial inspection datasets, and his work provided a proof-of-concept implementation with valuable insight into the decision-making process."

According to Hesham Yusuf, getting his first personal computer at the age of six was the moment that his love of technology began. A gift from his father, Hesham would later be taught networking by him. His father who had worked at Batelco (Bahrain Telecommunication Company), and set up the first intranet website.

As he progressed, Hesham started learning FrontPage and IIS, and used them to create a games repository for his younger brother.

His true inspiration, however, came in secondary school, where with encouragement from his father, he worked on a project to create a prototype for a traffic light system.

Hesham started his career journey with Bahrain Petroleum Company as an Instrumentation Engineer Intern, later becoming Control Systems Engineer Intern, then progressing to Control Systems Maintenance Engineer.

During his leisure time, Hesham worked on his passion for tech and spent time improving his skills. This lead himt work as a Full Stack Developer at Derma One Medical Centre and Web Designer/Co-founder at Mama's House in Bahrain.

“These explorations fostered my passion for creativity and led me to pursue engineering as my career path", said Hesham. "Seeing the benefit of my work is the most rewarding part of being an engineer. That is why I chose to do an industrial PhD, because it would allow me to contribute knowledge that could translate into products and services."

Before pursuing his doctorate, Hesham completed a Bachelor's Degree in Systems and Control Engineering and Master's Degree in Advanced Control and Systems Engineering with distinction from The University of Sheffield. Shorty after this was completed, he joined the NSIRC PhD Programme. His doctoral degree was awarded by The University of Sheffield and supported by Lloyd's Register Foundation (LRF) and TWI Ltd.

Today, Hesham is a Co-Founder of MyBubble Ltd in the United Kingdom, where he is responsible for managing the technology stack for the company. MyBubble uses mobile technology, artificial intelligence and blockchain technology, to provide free and tailored mental health guidance.

Hesham (left) and fellow NSIRC student. Photo: TWI / NSIRC
Hesham (left) and fellow NSIRC student. Photo: TWI / NSIRC

Explainability in Advanced Manufacturing: Leveraging the Interpretability of Multi-Criteria Decision-Making (MCDM) and Neutrosophic Logic

Hesham's PhD project investigated a new class of machine learning (ML) models called Interpretable ML, for use in pipe inspection automation.

Interpretable ML is also commonly referred to as explainable ML or explainable artificial intelligence (AI). The Interpretability of a model refers to the human understandability of the various components. Explainability refers to whether interpretable information can be then converted into an explanation.

The study explored how an Explainable ML model will function in an industrial setting when used for classification, and used the non-destructive testing of high-density polyethylene (HDPE) butt-fusion welds.

Results from Hesham's study have been used as a stepping stone for developing technology-ready explainable frameworks for practical application within industry.

Like all NSIRC PhD Students, Hesham received industrial and academic supervision during his research, comprising of Dr Kai Yang, Senior Project Leader at TWI Ltd, and Prof George Panoutsos and Dr Hua-Liang Wei from the University of Sheffield.

Impact of the Research Project

The research has made a significant contribution to academia by; creating a framework for extracting graphical and textual explanations for the model result, developing a data-driven interpretable classifier based on MCDM and fuzzy logic, extending the classification framework with neutrosophic logic for enhanced explainability, and applying the frameworks to a real-world industrial case study; HDPE butt-fusion pipe weld inspection.

Highlights of Hesham’s PhD Journey

During his PhD, Hesham presented the research at many conferences that including NDT, NSIRC, IEEE WCCI, and UKCI. Alongside these, his research results are also published in IEEE International Conference on Fuzzy Systems and Springer.

Recalling his experience at NSIRC, Hesham said, “I have gained invaluable knowledge about industry and developed a variety of skills. I am deeply indebted to my supervisors, for their tremendous support throughout. I would like to dedicate my accomplishment to my family, without whom you would not be reading this right now."

Hesham's plans for the future include exploring new areas of AI and its potential applications. Just like his NSIRC friends, he will continue to collaborate with other experts in their respective fields, tackling complex problems and build solutions for the future of industries.

Read the latest PhD Opportunities at NSIRC

Professional Development Through our New Course for Senior Leaders