Topics in Big Data Science in collaboration with Riaktr. How can eXplainable artificial intelligence tools be used to obtain information and human comprehension on black-box predictive models ? Towards explainable prediction of data-related revenue of telecommunication antennas in Benin

Details

Supervisors

Absil, Pierre-Antoine

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil en mathématiques appliquées, à finalité spécialisée

Abstract

The issue of interpreting any predictive model has been a research topic of great interest in recent years. Interpreting a model can provide additional insights into the data by understanding how the model makes its decisions. Interpreting and being able to explain a model have also become crucial when integrating the model into users' lives, such as in recommendation systems or performance analysis in business settings, for example. In this master's thesis in collaboration with Riaktr, our goal is to determine how recently developed interpretation tools in a research field called "Explainable Artificial Intelligence" can provide additional information on black-box predictive models. In this master's thesis in collaboration with Riaktr, our goal is to determine how recently developed interpretation tools in a research field called "Explainable Artificial Intelligence" can provide additional information on black-box predictive models.