Final Exam:
- Most programming questions have feedback; some have hidden test cases
- Covers content from the entire course: slightly less emphasis on content from lab test such that lab test (20%) + exam give even grade coverage
- 2019 exam will be available: note that some exam questions have been used in this year’s quizzes
- Past exams had some different content (e.g. Clojure, computer vision)
- Wrong selections in multi-choice questions take away from the mark for the grade (higher weight than correct selections)
- No access to lecture notes; there will be info on how to run prolog etc.
Topics covered:
- Planning (graph algorithms, path finding)
- Problem solving (constraint satisfaction problems)
- Logic
- Reasoning (classical logic, probabilistic reasoning)
- Working with uncertainty (probabilistic methods)
- Learning
- Competing with adversaries (games)
A recurring pattern across these topics is that the problems they solve can be reduced to search problems.
- The industrial revolution fundamentally changed how physical goods were produced, redistributing wealth and job opportunities. Will AI lead to a fundamental shift in how decisions are made?
- Fairness: AI in justice systems, recruiting etc.
- Singularity: recursive self-improvement leading to runaway systems
- AI probably won’t destroy humanity; we’re doing a good enough job already
Related topics:
- Probability theory, multivariate statistics, Bayesian inference
- Linear algebra, optimization, functional analysis
CSSE courses:
- COSC401 machine learning (more fundamental and theoretical)
- COSC440 deep learning (new!): deep neural networks
- COSC420 intelligent tutoring systems (unavailable - Tanja away next year): applied AI
- COSC428 computer vision
- DATA430: medical data informatics: applied AI