CSC2542 – Topics in Knowledge Representation and Reasoning: KR and Modern AI (Fall 2023)
I will be teaching CSC2542 in Fall 2023, with Sheila McIlraith. Questions about the course should be directed to
. For enrolled students the course website will be on Quercus. The first class will be on Friday, September 15th from noon-2pm. The course is in-person and the room is UC161 (note that this is changed from the originally booked room), which is in the southeast corner of University College.
Knowledge representation (KR) is a long-standing subfield of AI which studies symbolic representations of knowledge and associated mechanisms for automated reasoning. In this advanced course, we will overview topics in KR and consider the role they may play in modern AI research and systems, such as large language models and reinforcement learning systems. The course will comprise a mix of lectures and student presentations. Students will read research papers, participate in seminar-style presentations, and complete assignment(s) and a course project. Topics include: commonsense reasoning, logics for KR, probabilistic programming, representing goals and preferences, and the relationship between large language models and knowledge bases.
Prerequisites:
an introductory course in formal logic or an introductory course in symbolic AI comparable to CSC384, an introductory course in machine learning, and basic familiarity with reinforcement learning.
Meeting times:
Fridays noon-2pm (first class on September 15th)
Location:
UC161 (note that this is changed from the originally booked room)