Good morning everyone.

Before we start with a recap, one announcement.

Next Thursday, that is June the 13th, there is going to be a teaching assessment poll

here.

That is part of me trying to get a teaching certificate and the way that this is going

to work is that at 10.15, three people from the University Center for Education, Dietetics,

something I have no idea what that is called in English, are going to come here.

Then I am going to leave the room and for about half an hour they are going to interview

you.

The whole point of this exercise, by the way we are also going to turn off the cameras

and all that, and the entire point of this exercise is that they are going to talk to

you and you can voice any concerns, criticism, anything that pisses you off, anything that

you like about my teaching style, whatever, to your heart's content.

Everything you say will be completely anonymized.

I have no idea of finding out who says what.

They are not even going to tell me what you said specifically, they are going to go over

everything afterwards and make sure that everything is perfectly anonymous before they inform

me of anything.

I should mention that I am not going to be graded on the results or anything, so if you

like me, you don't have to be careful not to say bad things because otherwise I might

fail.

If you hate me, there is no point in making up shit just to punish me or whatever, that

is not going to affect me whatsoever.

I am only interested in the actual results of the whole thing and that is why everything

is perfectly anonymous.

I mentioned the cameras will be turned off.

I would also ask you to take that seriously and if at all possible, be here in person.

That is mostly concerned to the people who are probably not here now anyway.

For those of you who usually watch this on video, if you can be here in person, I would

very much appreciate it if you actually were.

The more people there are, the more data that we have and the more representative the results

are going to be.

For those who cannot be here in person at all, which affects probably quite a number

of you, again probably not the people in here but the people who watch this, there is also

going to be a Zoom link and at the same time they are going to do the exact same thing

on Zoom.

I am going to publish that Zoom link on Stutton probably today.

Any questions about that?

I hope to see as many of you as possible next week Thursday.

Yes, you have a question.

The teaching assessment poll is on Thursday.

There will be a quiz on Tuesday anyway.

So Thursday, June 13th, please be here.

So recap, partially observable Markov decision procedures, in short, POMDPs.

So the point now, as opposed to the standard Markov decision procedures, is that we assume

we are in a not fully observable environment, which means we have to integrate a sensor

model so that we have any kind of inclination about what the environment might look like

at all.

So we add a sensor model which we assume has the sensor Markov property and also stationary.

And then the entire trick behind trying to solve this, because obviously we now can't