COMP4620/8620 - Advanced Topics in AI
2011: Reasoning about Discrete Event Systems
This year (2011) the Advanced AI course will focus on reasoning about discrete event systems, including topics such as automated planning, diagnosis, and runtime verification and model checking. Note that the course content varies from year to year: material from past years can be accessed through the links in the menu to the left.
- No news is good news.
Offered By: The AI Group @
School of Computer Science @
Australian National University
Offered In: Second Semester, 2011.
Coordinator/lecturer: Patrik Haslum
Target: Undergraduate (COMP4620) and Graduate (COMP8620) students. Others welcome.
Enrollment: Undergraduates: The usual way via ISIS. Honors&Graduates&Others: Contact lecturer.
Admin: Julie Arnold
Course Subject: Computer Science
Unit Value: 6 units
Indicative Assessment: Assessment will be based on assignments.
Indicative Workload: TBA.
Prerequisite: COMP3620 (Intro2AI) or equivalent; a good grasp of basic theoretical computer science (propositional logic, basic complexity theory); some programming experience.
Prescribed texts: Reading material, consisting mostly of research papers, will be distributed.
Study@ANU page: PLEASE IGNORE THE STUDY@ PAGE. IT HAS NOT BEEN UPDATED. THE INFORMATION ON IT IS NOT VALID.
StudentDB page: https://cecs.anu.edu.au/studentdb/courses/students/csg_student.cgi?Course_Code=COMP4620
Wattle page: http://wattlecourses.anu.edu.au/course/view.php?id=10671
This page: http://cs.anu.edu.au/courses/COMP4620/2011.html
Discrete event models capture the dynamics of systems that evolve by discrete events. Examples of such systems include digital control systems, as found in embedded and autonomous systems (including robots), telecommunications, etc., but also protocols, workflows, games, and even models of evolution.
System models can be used for many purposes:
- Monitoring & diagnosis: Observing (as far as it can be observed) the behaviour of the system over time, the model is used to infer the state of the system, check if it is functioning correctly, and if not, to determine what faults may have occurred.
- Planning & control: Using the model of system behaviour, devise a plan of control actions to drive/guide the system to a desired goal state or keep it in desired states.
- Verification: Analyse the model, either in advance or at runtime, to determine if it meets desired criteria.
The course will cover some material related to all these uses, as well as the art of modelling practical systems in discrete-event formalisms.
Assessment will be based on assignments. The exact contents of assignments is yet to be determined, but in general, each one will require students to read and understand some material (probably research papers), describing, e.g., methods of solving a particular type of problem, or different positions in a debate, and then apply that understanding, for example by designing experiments to evaluate the capability of a problem-solving method, or the validity of arguments, and summarise their findings in a short written report.