Invited Session Challenges towards Fairness and Transparency for Data Processes, Algorithms and Decision-Support Models Organizer/Chair: Claudia Tarantola (Università di Pavia) Discussant: Gloria Polinesi (Università Politecnica delle Marche) Room: T30 Floor: ground Short summary: In this session, we will explore various aspects of the ethics of data science. Nowadays, many decisions are made by taking predictive models based on observed data as suggestions. Despite being created through a fair and well-intentioned learning process, these models can still unintentionally discriminate against certain groups of people, leading to unfair outcomes. Algorithmic bias is a pressing problem in this regard, as machine learning models can perpetuate existing discrimination by taking into account factors such as race, gender, or age. In addition, the use of sensitive data to train and validate these models raises issues of privacy and data security.
#
Papers
1.
A new measure of discrimination in machine learning algorithms Author(s) Roberta Pappada' Francesco Pauli
2.
Challenges on Ethics, and Privacy in AI Applications to Fintech Author(s) Joana Matos Dias Bernardete Ribeiro Catarina Silva
3.
Uncertainty and fairness metrics Author(s) Anna Gottard
Dipartimento di Scienze Economiche e Sociali (Di.S.E.S.)