Innovation for Society Program
During the 2021-2022 academic year, I joined the first edition of the Innovation for Society excellence program at Ghent University. This program brings together future engineers to explore their role in achieving positive and lasting impact on the world.
Students were divided into smaller groups, each receiving a specific case to work on. We came together regularly for guest lectures and to present our progress. I thoroughly enjoyed this program, as it brought together young engineers with diverse specializations ranging from architecture to computer science.
The Challenge
My team’s case involved detecting racial bias in facial recognition systems. With the increasing deployment of innovative technologies in justice and national security applications, ensuring fair judgment for everyone is crucial. While this might seem feasible in theory, reality presents a different picture. Biases in training data or unbalanced datasets can expose innocent individuals to significant risks.
Our Findings
Throughout the project, we confirmed the presence of racial bias in facial recognition systems and proposed several mitigation techniques, including the use of synthetic data to balance training datasets.
Acknowledgements
I would like to thank Prof. dr. ir. Hendrik de Bie for organizing this program and showing us that engineers can and should make a positive impact on society.