Okay, time's up.

I'm very happy that 192 of you actually took part in this.

You've already seen from the development of the bar, if you looked up, that this was much

too long, which of course is explainable because it was extremely easy, I think.

So you were just being able to top it off, which is perfectly fine with me.

You did well on average. Okay?

So what I'm trying to do, just to reiterate with the quizzes, is actually keep you working,

keep you prepared, and that seems to have worked.

Okay, let's see how that develops over the course of the semester.

I can already tell you that being prepared is very well correlated with good grades and

by all accounts having more fun. So keep it up.

Following quizzes, which are much more, the problems are much more, well, I consider the

ones in this quiz to be kind of propaganda questions, much more than did you really understand

the fine detail. The other questions will be much more tricky.

So prepare for them. It's an excellent idea to surprise your friends with your own quiz

questions. When you're waiting in the men's IQ or so,

just dream up a quick quiz question and challenge one of your friends.

If you come up with quiz questions you're proud of, feel free to actually email them

to me because I'm always on the lookout for good quiz questions.

Okay, any questions? Yes?

I hope. Wow. This is something I have absolutely no control over.

I was told I have to press hybrid mode here and I did that.

Let me send an email.

Let's hope for the best then. There is an advantage to being here after all, right?

Thank you. Okay, good. Livestream is working. Excellent.

So the last thing we did on Tuesday was we looked at the topics we're covering and we

saw that we do symbolic AI which includes problem solving by search, gameplay, constraint

satisfaction problems, then a big block of knowledge and reasoning, and then something

called planning which is really all of the before plus time.

And then in the semester after we kind of complete the picture by adding uncertainty

to it which makes everything much more interesting let's say.

The algorithm is more complex but of course more realistic as well because as agents,

intelligent agents in the world we don't know everything unfortunately.

And then we do machine learning a little bit and communication which is applying machine

learning mostly to the problem of understanding natural language.

Good. We are basically keeping things on a relatively

theoretical, relatively shallow level here. But if you're interested in getting your hands

dirty there is the AI systems project which basically follows the course and gives you

interesting and of course challenging problems to implement.

We look back at this so in search, the first chapter we let you search for connections

in the Indian railway system. We have a big data set there and then you can see whether

you can go from Hyderabad to Bangalore with less than three stops in five hours or something

like this and that gets computationally interesting.

For the game playing we will kind of not do Kala but Chinese checkers which is this six-sided

star-shaped game which makes some of the assumptions like we only have two players and everything

is on a rectangle wrong and you have to think again.

For CSP I'm not sure that's the latest one but we kind of had a data set that basically

was the scheduling problem of scheduling all of techfac which you can work on. In knowledge

and reasoning we build theorem provers or use them from the shelf and those kind of

things and of course in planning there's a planning system that we look at.