“Intelligent Agents: Reasoning, Planning, Perception, and Cooperation”

PhD course, fall term 2021

Course coordinator: Rudolf Mester, IDI

This PhD course is designed with the intention to provide to PhD students, in particular those working on Artificial Intelligence topics or using AI methods, intermediate level and advanced methods for the analysis and design of Intelligent Agents. The term "Agent" is used here with a special focus on embodied AI, that is: either systems that are acting in a physical environment, or those that are interacting with such an environment.

In the present academic year, the course is held as a DIXIL course. PhD students that are interested in participation are requested to address the course coordinator by August 13, 2021. This is not an exclusive deadline, but helps to organize the course well. Use the coordinator email address as specified on his NTNU profile page https://www.ntnu.edu/employees/rudolf.mester

Further detail information on contents, format, exam modalities will be provided soon. If you are interested, please see at your earliest convenience that you are registered on the course mailing list.

The first course event is going to held in week 35, probably on Tuesday of that week. The kick-off will be held in digital format. Precise day and time will be provided to the registered participants a.s.a.p. Those who participate in that kickoff event may take the course. On that event, I will give a general introduction to the course and its organization. The first proper lecture will be in (early) September. The current plan is to have all course events on Tuesdays.

The course is based on the following textbooks:

• Michael Wooldridge: An Introduction to Multiagent Systems
• Steven M. LaValle: Planning algorithms. Available as printed book or online from http://lavalle.pl/planning/

Course format: lectures (60 %) plus seminar style workshops with student presentations (40%) Evaluation form: a) individual presentation in seminars, b) report, and c) oral exam (weighting: 1/1/1). Grades: pass/fail

Planned course contents:

Part I: Intelligent Autonomous Agents

• Introduction
• Software Agents and embodied Agents
• Autonomy
• Agents as Intentional Systems
• Deductive Reasoning Agents
• Practical Reasoning Agents
• Means--Ends Reasoning
• Procedural Reasoning System
• Reactive and Hybrid Agents
• Communication and Cooperation
• Ontologies
• Speech Acts
• Cooperative Distributed Problem Solving
• Coordination between agents

Part II Foundations of Agent Perception

• Visual Perception
• Auditory Perception
• Other Sensor Modalities
• From Sensor Data to Percepts and Interpretations

Part III Planning

• Algorithms, Planners, and Plans
• Discrete Planning
• Motion Planning for Mobile Agents
• Decision-Theoretic Planning
• Sequential Decision Theory
• Sensors and Information Spaces
• Planning Under Sensing Uncertainty

The above list of contents is a catalog: the weight that will be put on individual topics will be adapted to the existing prior knowledge of the participants; this means that certain fundamental topics will be dealt with extensively while others will be covered more in an overview mode. This allows us to focus on areas of particular interest for the participants and on recent research trends.