This week the focus of the course was on representation through logic and planning. These are the tools with which our AI agent represents and solves problems or plans a course of action.
If I’m honest (and I’m trying to be) week four was the toughest for me so far. So far the course has introduced a concept, demonstrated its operation or application and then reinforced it with quizzes. The explanations behind the quiz answers have been some of the most useful aspects of the course.
For me, I like to be able to understand the underlying principles of how an answer was arrived at, the quizzes assist with this. This week there quiz element was quite sparse.
Norvig and Thrun have quite distinct teaching styles and I think I’m more attuned to Thrun. It does sound like an excuse but again, I think it was the absence of quizzes which was the main issue. Again time has been a factor here but hoping to get back on track before the appearance of next week’s mid-term.
The other highlight of the week was the introduction of virtual office hours using Google + hangouts, potentially I think this is a really good idea to provide interaction with students. Again there were a few technical issues (I’m waiting to see the session on YouTube) but as in any experiment of this scale there were bound to be a few slight hiccups.
This week it’s planning and uncertainty and reinforcement learning. Uncertainty I can deliver in spades and anything that reinforces my learning can only be a good thing!