AAAI 2021 and AAMAS 2021 tutorial: Designing Agents' Preferences, Beliefs, and Identities


Instructions for AAMAS 2021 tutorial:
The tutorial is on May 3, 2pm-5pm BST. Please try to sign in a little early if this is your first event this AAMAS, and contact me if you have trouble. The way to find the tutorial (I think): after signing in with underline, find your way to workshops and tutorials, click on "all sessions", and find this tutorial. Then click on "Live Session" which I think opens 20-30 minutes before the official start time. ("Website" takes you to this page.) During the tutorial we will run some (anonymous, optional) zoom polls so if you want to participate in those, make sure you're on a device that allows that. The slides are below, which also include lots of commented-out slides in the pptx version, and links to other materials.

Brief description:
We often assume that each agent has a well-defined identity, well-defined preferences over outcomes, and well-defined beliefs about the world. However, when designing agents, we in fact need to specify where the boundaries between one agent and another in the system lie, what objective functions these agents aim to maximize, and to some extent even what belief formation processes they use. What is the right way to do so? As more and more AI systems are deployed in the world, this question becomes increasingly important. In this tutorial, I will show how it can be approached from the perspectives of decision theory, game theory, social choice theory, and the algorithmic and computational aspects of these fields. (No previous background required.)

A AAAI'19 blue sky writeup on a subset of these ideas can be found here: https://users.cs.duke.edu/~conitzer/designingAAAI19.pdf

Outline: Slides: pptx, pdf.