Intro to Artificial Intelligence Re-cap

Yesterday I got caught up on and finished the first Homework, at the last minute, just like when I was in college. Or so I thought it was last minute - apparently enough people waited longer than me that the servers were overloaded & they extended the deadline by a day.

Some of my thoughts and observations on the class:

Social Effect

The social networking phenomenon for this class is quite impressive. The class enrollment is HUGE. It's like this class is being echoed over & over again via social networks.

One way I've benefited from this is the written lecture notes on Google Docs. But beyond that, it seems like it's just the same thing over & over again, just rehashed in different forms. There are some good questions & clarifications posted, but the signal to noise ration seems quite high. The value per person starts very high, with Peter Norvig & Sebastian Thrun at the top, and the AI class team following. After that, each additional person seems to contribute rather little additional value. Or so it seems to me from my quick observations ...

It would be interesting to see some statistics & correlation in terms of the students that are most active with the social networking for the class, and their class performance. :-)

Granted, I haven't looked that hard, but it seems that most of the social discussion about the class is about the actual AI topics, and mundane class details. I think those discussing the class at a higher level are much fewer. Hopefully this post adds something useful to that less-explored area.

Quiz & Homework Questions

I am a visual learner, and I thrive on precise & specific written instructions. The fact that the quiz & homework questions are not fully written out presents an additional challenge, and is probably a big part of the reason for the errors I've made. I might need to resort to transcribing (or getting the transcriptions from the web) the questions in order to do better.

Video vs. Textbook

There are a lot of conflicts & errors in the material being presented in the videos. This leads to another problem when referring to an exact written transcript of the video lectures. This is a big difference between a lecture/video and a textbook, where the text is reviewed & corrected before publication. The number of Clarifications - especially on homework questions - that are posted are a good reminder of human error, even with the professors. It also shows how communication oversights, missing details, and a lack of mathematically defined problem descriptions has shown to be a bit of an issue for the course.


The homework & questions seem simple compared to the possibilities for the topics covered & presented. Maybe that will change as the course progresses. But so far, the class has been easy, compared to what it could be.

On the other hand, the course does a good job of covering the AI topics, without getting deeply in to testing how intelligent the student is by presenting problems that really challenge his or her problem solving ability. I'm guessing this is on purpose, considering the large & wide audience, and also the goal of bringing AI to a large audience without making it too difficult. While being able to solve hard AI problems is very important for those at Google, or working on advancing the field, it isn't necessary for this class.

My Results

I've scored 87% on Unit 1 quizzes, 66% on Unit 2 quizzes, and 96% on Homework 1. Good enough for a last minute cram session Sat night and Sunday afternoon :-)
Shared publicly