I am glad to share that I have successfully defended my thesis entitled “Towards Holistic Assessment of Machine Learning Trustworthiness.” For the past year, I have been working hard on developing a unified framework to evaluate machine learning models across different dimensions, namely utility, privacy, robustness and fairness. Throughout my research, I had to develop custom training algorithms, model architectures and evaluation metrics. Doing independent research has proven to be a difficult yet rewarding journey, where I was constantly reading state of the art research, learning new concepts and applying novel techniques. I am looking forward to having my work published on the Deep Blue archive of the University of Michigan.

I would like to acknowledge the The Fulbright Program for sponsoring my graduate education and stay here in the United States. I would also like to acknowledge my advisor Birhanu Eshete for providing me with valuable feedback throughout the process and the computing resources needed to run my experiments. Finally, a special thanks goes to Di M., Srijita Das and Ang Li for serving as committee members for my defense.

Check my published thesis at https://dx.doi.org/10.7302/22653.

Go Blue!