School of CSE - Details Needed for New Course Proposal Course Title: An Introduction to Intelligent Agent Architectures Proposer: Dr William Uther Rationale Why is the new course being proposed? - Students reaching Honours/PhD are under-prepared in this area. It should also make the current 4th year experimental robotics course more useful for the students by giving them more background. - Specifically it will prepare students for research using structured architectures on real agent systems (e.g RoboCup). What are the academic objectives? The students should leave the course: - Having been introduced to a wide range of agent architectures - with a knowledge of the strengths and weaknesses of each type of architecture - Having had practical experience with a particular agent architecture - The current agent architecture for the CSE Legged or Simulator League RoboCup entries Which programs/stage does it serve? We expect this to be a course in the second semester of 3rd year. This gives the students background thatis useful for the RoboCup project in their 4th year, for experimental robotics, and for any other Ugrad, masters or PhD research involving robots. The course should also be available for incoming PhD students who have a weak background in this area. Why can the same objectives not be achieved with existing courses? The experimental robotics course is too late to fit in before this course. The current AI course could be modified to include this material, but this would mean dropping other material out - I don't think this is worthwhile. How does the proposed course relate to other courses? This is a complementary course with Undergraduate AI - a student would do well to take both. This is a good course to have before experimental robotics. Wayne Wobcke's course is the closest course in subject matter. What overlap is there? There is very little overlap. Any overlap that exists is with: - AI - Experimental Robotics - Wayne Wobcke's agent's course If there is any overlap, why is this justified/not a problem? This course is significantly more focused that the AI course, and with a different focus than Wayne's course (although I do have one lecture on BDI agents). It is an introduction to the material for the experimental robotics course. Stakeholders and Consultation Who are the potential stakeholders, who was consulted about the proposal (inside the School as well as outside), what was the result of that consultation? Claude Sammut: is very interested in the course as a practical introduction to the AIBO platform. NICTA (Arun): is interested in the course as a more in depth introduction to agent architectures for PhD students. Enrolment Impacts Likely enrolment (with justification), and impact on enrolments of other courses. Justification of Prerequisites (or lack thereof) There are 2 pre-reqs: - Linear Algebra: Students will need this for the kinematics and localization - 2011: Students will need to have seen Stacks/Queues/PriQueues/Breadth First Search/Depth First Search. Having seen Dijkstra's algorithm would be good too, but not neccessary. It would be nice to recommend that students who take this course also take AI, but it shouldn't be a pre-requisite. Any Courses this is Replacing, and Why? Not replacing any courses. Delivery and Assessment Anything noteworthy about delivery mode, assessment (with justification). Each week: 2 hours lectures, 1 hour lab Assignments: 3 assignments spread over semester - all deal with AIBOs (1st half of lecture material) Exam: Deals with non-AIBO stuff (2nd half of lecture material) Exam and Assignments have equal weight Handbook Entry An introduction to a variety of artificially intelligent agent achitectures. In particular, an in-depth study of the rUNSWift Soccer Playing architecture. An introduction to a variety of other action planning systems, including A* search, STRIPs Style planning, Partial Order Planning and Markov Decision Processes. Textbooks/References Text-book: none Reference: Russell & Norvig, "Artificial Intelligence" (used by the AI course). Syllabus Indicative syllabus / overview of contents (at a level of detail well beyond that of the handbook entry) See other file. Effect on School Resources: 1. Who is proposed to teach the proposed new course, and what impact would this have on their planned/current allocation? Will Uther (with possible guest lectures) - No impact - NICTA employee 2. What sort of tutorial component is proposed, if any? 1 hour lab per week. No tutes. Course will be held to < 20 students. Will will give lab. 3. What is the likely impact on lab utilisation (this relates to assignment and project work as well as scheduled labs? 1st half of course - will be using RoboCup machines. 2nd half of course - Can be held in any lab, incl. MacOS X lab. 4. Any other resource needs? E.g. special print/disk quota, access to servers, access to special machines, special labs.