CA | ES | EN

EQUITY AND INTERPRETABILITY OF ARTIFICIAL INTELLIGENCE ALGORITHMS

EQUITY AND INTERPRETABILITY OF ARTIFICIAL INTELLIGENCE ALGORITHMS

It is well known that algorithmic solutions do not always behave “fairly”. There are many potential reasons for this behaviour, from subjective design to being based on training data that presents a bias, while still faithfully representing reality. This session will gather experiences, methodologies and new approaches in order to minimise not only this problem, but also its interpretation for cases in which the algorithm is of the “black box” type.

Presented by:

Ludovico Boratto
Research Scientist
Eurecat
www.ludovicoboratto.com
www.eurecat.org

Speakers:

Carlos Castillo
Distinguished Research Professor
Universitat Pompeu Fabra
chato.cl/research

Jose A. Rodriguez-Serrano
Data Science Program Manager
BBVA Data & Analytics
bbvadata.com

Oleguer Sagarra
Co-CEO
Dribia
www.dribia.com