DAUIN - GR-16 - SOFTWARE ENGINEERING GROUP - SOFTENG
Development of a data visual annotation system based on Bayesian Inference
External reference persons Elena Beretta (Vrije Universiteit Amsterdam)
Thesis type DATA ANALYSIS, DATA VISUALIZATION
Description Automated decision-making (ADM) systems may affect multiple aspects of our lives, both in a positive and in a negative way. A recurring problem is the systematic discrimination of specific population groups, in violation of the EU Charter of Fundamental Rights. One of the potential causes of discriminative behavior lies in the quality of the data used to train such ADM systems.
The purpose of the thesis is to detect and visualize the potential race discriminatory risk for future machine learning system by providing a data annotation system based on Bayesian Inference. The notation will serve as a diagnostic framework to immediately visualize data appropriateness and potential bias occurring when sampling thetraining set from an available dataset. The measures have been already defined in the following publication " Beretta, E., Vetr˛, A., Lepri, B., & Martin, J. C. D. (2021, March). Detecting discriminatory risk through data annotation based on Bayesian inferences. In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (pp. 794-804)" available at https://arxiv.org/abs/2101.11358 . Other publications, metrics, and reference datasets will be made available by the supervisors.
Required skills Basic data analysis concepts, programming in R (preferred) or python, basic concepts of data visualization
Deadline 18/06/2022 PROPONI LA TUA CANDIDATURA