Agoritsa Polyzou

Assistant Professor KFSCIS

Lecture Information:
  • February 4, 2022
  • 2:00 PM
Agoritsa Polyzou

Speaker Bio

Dr. Agoritsa Polyzou is an Assistant Professor of Computer Science in the Knight Foundation
School of Computing and Information Sciences at Florida International University (FIU). Before
joining FIU, she was a postdoctoral Fritz family fellow in the Massive Data Institute (MDI) of
McCourt School of Public Policy at Georgetown University. She received her Ph.D. in Computer
Science and Engineering from the University of Minnesota in 2020, and her bachelor in
Computer Engineering and Informatics from the University of Patras, Greece. She is engaging in
projects at the intersection of big data, machine learning, ethics, and fairness. Her research
interests include data mining, recommender systems, the application of machine learning
techniques within educational contexts, and the fairness concerns that arise from their use. Her
goal is to help students succeed using data and machine learning models. At the same time, she
is interested in ensuring that such models will be fair and responsible towards all for their users.


There has been a lot of research work that applies computational models to explore patterns
and data from various sources and applications. While these advancements highlight the value
of data science, they also demonstrate their power over people's lives and decision making.
However, there is not much discussion about their core values. Most of existing discussion is
about general principles that traditional data analytics approaches need to follow and, in
particular, in areas of research that directly involve human subjects (e.g., biomedical domain),
autonomous systems, or human-computer interaction). It is rare to come across specific
guidelines that cover the ethical aspects of the research questions and processes and provide
an ethical standard for researchers to follow. More specific existing research on AI ethics is
mainly focused on narrow topics, like battling unfairness and bias present in models' outcomes,
or securing the privacy of the users' information. However, there are other factors that may
make a fair (based on some existing definition of fairness) and/or privacy-preserving model
‘unethical’ for a researcher to develop. In this talk, we consider the researchers as the main
stakeholders for whom we want to propose some clear and easy-to-follow ideas that promote
ethical thinking. The goal is to offer a more holistic view of what it means for a model to be
ethical and present a number of practical ethical guidelines that researchers need to consider
during any phase of their work. We want to encourage a conversation about researchers'
responsibilities, particularly when the data they use are collected from a social media platform,
and not from a well-defined research-oriented process.