Doctoral Meeting: 'Explaining black-box AI models'

Machine learning is recently becoming a widespread standard tool for many different data-driven tasks. Unfortunately, many of the existing approaches employed have a completely obscure inner reasoning. This is an issue for all those applications where trust and liability are a must like medical decision-assistant tools, credit scoring models in finance and so forth. In this doctoral meeting, we will explore how the field of eXplainable AI (XAI) tries to address this problem by providing models that can explain their behaviour to a final user.