Speaker: Emiliano Lorini
Date and Time: Thursday, April 6th 2023, 16:30-18:00
Venue: online only. (i.e. not hybrid!)
Title: Non-Classical Logics for Explanations in AI Systems
Abstract. I will show how non-classical logics with special emphasis on modal logic, epistemic logic, conditional logic and deontic logic can be used to represent and compare a rich variety of explanations in classifier systems widely used in the field machine learning including abductive, constrastive, counterfactual, objective vs subjective, interactive as well as concepts of bias, discrimination and protected feature. I will present proof-theoretic and complexity results for these logics and illustrate their expressiveness through some concrete examples.