Open On-line Course - AI Planning

Course Programme, Schedule, Additional Resources and Exercises
Quicklinks to Course Materials on: Media Hopper, YouTube, AIAI Media Server

About the Course

This course is a free open-access learning experience which introduces artificial intelligence planning techniques and their applications. It comprises a 5 week course with up to 10 hours of lecture material. It is accessible at several levels of engagement from an "Awareness Level", through the core "Foundation Level", to a more involved "Performance Level" requiring programming and other assignments. It was presented to over 100,000 participants on the Coursera platform in 2013, 2014 and 2015. The course materials continue to be available via the University of Edinburgh Media Hopper, YouTube and the AIAI Media Server.

The course aims to provide a foundation in artificial intelligence techniques for planning, with an overview of the wide spectrum of different problems and approaches, including their underlying theory and their applications.

What is AI Planning?

planning: explicit deliberation process that chooses and organizes actions by anticipating their outcomes.
planning in AI: computational study of this deliberation process.

Summary of intended learning outcomes

Understand different planning problems.
Have the basic know how to design and implement AI planning systems.
Know how to use AI planning technology for projects in different application domains.
Ability to make use of AI planning literature.

The course may also be useful to provide a structured introduction to AI planning techniques for on-campus students taking related AI and knowledge-based systems courses, or undertaking projects which utilise AI planning concepts.

Recommended Background

The course can be studied at different levels, but you will need a basic understanding of logic and mathematical reasoning. Optional programming assignments require programming skills.
The course follows a text book, but this is not required for the course:
Automated Planning: Theory and Practice by M. Ghallab, D. Nau, and P. Traverso (Elsevier, ISBN 1-55860-856-7) 2004. See also the authors' later "Automated Planning and Acting" [PDF Copy]

Instructors

Dr. Gerhard Wickler: Obtained his Ph.D. in 1999 at Edinburgh in the area of AI Planning. He went on to hold research positions in Italy, Belgium, and Germany, working in several areas of AI. From 2004 to 2015 he was a senior researcher at the Artificial Intelligence Application Institute (AIAI) within the School of Informatics at the University of Edinburgh, where he taught the AI Planning course. Dr. Wickler regularly publishes in AI-related conferences and journals, reporting on his research in AI Planning and Intelligent Agents applied to emergency response. He is an active reviewer for a number of conferences and journals. He has been a member of the programme committees for various workshops and conferences, including the Intelligent Systems track at ISCRAM. He has been a member of the board of directors of the ISCRAM Association and has received the ISCRAM Distinguished Service Award.
Prof. Austin Tate: Director of the Artificial Intelligence Applications Institute (AIAI) and holds the Personal Chair of Knowledge-Based Systems at the University of Edinburgh. He is a Fellow of the Royal Academy of Engineering, Fellow of the Royal Society of Edinburgh (Scotland's National Academy), Fellow of the Association for the Advancement of AI, Fellow of the British Computer Society, Senior Visiting Research Scientist at the Institute of Human & Machine Cognition (IHMC) in Florida, and on the editorial board of a number of AI journals. His research background involves advanced knowledge and planning technologies, with a focus on their use in emergency response and collaborative systems especially using virtual worlds. He is the Coordinator for the Virtual University of Edinburgh (Vue) and Coordinator for Distance Education in the School of Informatics at the University of Edinburgh.

Modules

The course will adopt a framework describing:

formal definition of the (basic) planning problem and simplification to make it tractable
different approaches used to solve planning problems (algorithms)
extensions to the basic problem, reintroduction of realism, and advanced techniques
techniques for plan execution and example applications

Week-by-Week

Week 1: Introduction and Planning in Context

Week 2: State-Space Search: Heuristic Search and STRIPS

Week 3: Plan-Space Search and HTN Planning

Week 4: Graphplan and Advanced Heuristics

Week 5: Plan Execution and Applications

Frequently Asked Questions

Will I get a certificate after completing this class?
Students who completed the Coursera classes in 2013, 2014 and 2015 were offered a Statement of Accomplishment signed by the instructors.
Do I earn University of Edinburgh credits upon completion of this class?
The Statement of Accomplishment was not part of a formal qualification from the University of Edinburgh. However, it may be useful to demonstrate prior learning and interest in your subject to a higher education institution or potential employer.
What resources will I need for this class?
Nothing is required, but if you want to try out implementing some of the algorithms described in the lectures you'll need access to a programming environment. No specific programming language is required. Also, you may want to download existing planners and try those out. This may require you to compile them first.
What is the coolest thing I'll learn if I take this class?
In this class you will learn the basic algorithms that are used in robots to deliberate over a course of actions to take. Simpler, reactive robots don't need this, but if a robot is to act intelligently, this type of reasoning about actions is vital.
What Twitter hash tag should I use?
Use the hash tag #aiplan for tweets about the course. The course team Twitter username is @aiplanner.
How come this is free?
We are passionate about open on-line collaboration and education. Our taught AI planning course at Edinburgh has always published its course materials, readings and resources on-line for anyone to view. Our own on-campus students can access these materials at times when the course is not available if it is relevant to their interests and projects. We want to make the materials available in a more accessible form that can reach a broader audience who might be interested in AI planning technology. This achieves our primary objective of getting such technology into productive use. Another benefit for us is that more people get to know about courses in AI in the School of Informatics at the University of Edinburgh, or get interested in studying or collaborating with us.