At IE University, we are proud of the wealth of ambition and knowledge our students bring, and we work to provide the resources and space for students to realize their visions. Ryan Daher is one such student. His cutting-edge project on machine learning to earn his Bachelor in Data & Business Analytics exemplifies just how far you can go with your learning here at IE University.
Originally from Lebanon, Ryan has become a true citizen of the world by working and studying in numerous countries such as Spain, Canada and the USA. He is currently working in Dubai as a consultant for the Boston Consulting Group. His story showcases the innovation, curiosity and entrepreneurship that the IE Venture Lab seeks to nourish.
THE BIG IDEA
Ryan originally started out as a Computer Engineering student in Canada, but his first encounter with machine learning ignited a passion that led him to transfer to IE University’s Bachelor in Data & Business Analytics program. Intrigued by the power of “training an algorithm to perform tasks at levels that exceed human capabilities,” he set his sights on learning as much as he could about this complex technology to harness “its potential to really impact the world.”
To qualify his learning, Ryan settled on an ambitious but meaningful project: measuring transparent and fair machine learning. Carrying out the project required careful planning and intense research. As Ryan explains, “It is always best practice to break down large tasks of work into manageable steps that can be executed. It makes seemingly impossible objectives attainable.”
Tackling such an ambitious task might seem daunting. However, Ryan shared the approach that helped him to achieve success: “Start with the last step and back-pedal to your starting point. Understanding what exactly you want to deliver is the most important step in any project. By clearly defining your end goal, you can piece together the components you need to tackle to deliver that objective.” Having a plan—as well as having the support of his advisor Professor Manoel Gadi—meant that Ryan could dedicate himself to exploring his passion for this topic in an effective way.
STEP BY STEP TOWARD SUCCESS
To begin, Ryan first dove into the legal questions and gray areas related to machine learning and decisions. Given that this is such an avant-garde technology that continues to evolve, this was a challenging but necessary research stage. Ryan also became versed in “the psychology and philosophy of fairness, understanding the previous works conducted in the field, the definitions that were crafted by researchers, and their interpretations.”
Once Ryan had hammered out these foundational elements regarding how to understand fairness and bias in machine decisions, he returned to his scientific background by studying statistical measures of fairness. Based on past statistical assessments and formulas, he developed a library coded using Python. “By importing my newly created library, users could conduct complete and comprehensive assessments to report the bias and unfairness present within their datasets and their model’s outputs.”
To make this library as comprehensive as possible, Ryan tackled the enigmatic subject of black-box models. Decisions made by these models can be difficult to evaluate for bias because the computations made by the algorithm may not be accessible to users. “Utilizing state-of-the-art coding techniques and available python libraries, I compiled an additional component of the library that estimated decision factors in black-box models. This final step was the most complicated to crack; however, it was also the most rewarding as it really made my work feel complete.”
Although Ryan’s journey at IE University is complete, this project allowed him to engage meaningfully with a topic that he is passionate about. To other students, Ryan offers the following advice for a great IE University experience:
1) Work hard and smart on a subject you love.
2) Have something else going on besides studies; whether that be sports, hobbies or a side business.
3) Make strong and long-term friends.