Mitigating Bias in Decision-Making Systems: a Control Systems Perspective

Abstract : 

Prediction-based decision-making systems are becoming increasingly prevalent in various domains. Previous studies have demonstrated that such systems are vulnerable to runaway feedback loops which exacerbate existing biases. The automated decisions have dynamic feedback effects on the system itself. In this talk we will show how existence of feedback loops in the machine learning-based decision-making pipeline can perpetuate and reinforce machine learning biases and propose control strategies to counteract their undesired effects.

Biography :

Dr Giulia De Pasquale is currently a PostDoc at ETH Zürich, Switzerland.
She earned her PhD in Control and Systems Engineering at the University of Padova, Italy,  in 2023.
In 2021/2022, she was a Visiting Research Scholar at the University of California, Santa Barbara.
She holds both a Master Degree in Control Engineering (2019) and the Bachelor Degree in Information Engineering (2017) from the University of Padova. 
During her Master Degree, she participated in the Erasmus program with a research experience at the Luleå Tekniska Universitet, Sweden, in collaboration with RI.SE SICS Luleå and ABB Västerås. In 2018 she spent six months as a Visiting Student at ETH Zürich, Switzerland.
Starting January 2025, she will join the Control Systems group at Eindhoven University of Technology as an Assistant Professor.

Her current research interests include modeling, analysis and control of networked socio-technical systems.