Artificial Intelligence

6.034 is MIT's undergraduate artificial intelligence course, and Patrick Henry Winston's fall 2010 lectures are published on OpenCourseWare in full. The course introduces the basic knowledge representation, problem solving and learning methods of AI; on completion, students should be able to build intelligent systems by assembling solutions to concrete computational problems, understand the roles of representation, problem solving and learning in intelligent-system engineering, and appreciate how problem solving, vision and language relate to human intelligence from a computational perspective. The lectures begin with reasoning via goal trees and rule-based expert systems, then search, including depth-first, hill climbing, beam, branch and bound, A*, and games with minimax and alpha-beta. A block on constraints covers interpreting line drawings, domain reduction and visual object recognition.

The learning section spans nearest neighbours, identification trees, neural nets and deep neural nets, genetic algorithms, sparse spaces, near misses, support vector machines and boosting, followed by representations, cognitive architectures and two lectures on probabilistic inference. Mega-recitation videos, tutorials, programming assignments, exams and demonstrations accompany the lectures, together with instructor insights on teaching. Everything is free under a Creative Commons licence, with no registration and no certificate. Note that the course predates the modern deep learning era, so treat it as foundations rather than current practice.