Course name
----------
Intelligent Agents 2

Lecturers
-------

X. Parent, L. van der Torre, A. Steen
A. Farjami (Teaching assistant) 

Motivation: 
---------

This course is for the students interested in artificial intelligence,
and more specifically normative reasoning and deontic logic. Like
humans, AI systems evolve in a highly regulated environment. They have
to abide by legal rules (e.g., privacy requirements), but also make
ethical decisions when needed, and reason about the consequences of the
choices being made. Deontic logic studies how norms (obligations and
permissions) can be formalized, and how normative reasoning can be
mechanized. It is a first step towards responsible AI.

Objective
--------

The objective of this course is to introduce students to formal
methods for normative reasoning for agents (deontic logic).


Description
----------

We will study three frameworks for normative reasoning that have
dominated the landscape: monadic deontic logic (MDL), dyadic deontic
logic (DDL), and input/output logic (I/O logic). Basic issues from the
deontic logic literature will be revisited using these tools
(conflicts, reasoning about violation). This year we will also look at
the implementation in Isabelle/HOL.
 

Teaching material will be provided in the form of chapters from a
forthcoming textbook.

This course will extend the course on Knowledge Representation, and
complement the module "Selected topics in AI".

Language: English

Class hours: Wed 14:00-15:30

Grading system: final exam 60%; homework 40%.