Uta.

I wanted to ask, I don't know if I missed it or if it just isn't online when the examination

date is.

We haven't set one, but probably it's a good idea to do that.

We're good for planning by now.

Right.

Yeah.

Let's talk about that but let me first.

Why aren't you.

Come on.

No, not yet. Sorry for being so late but turned out that my.

I noticed a couple of problems so I had to recompile and then I ran into lots of errors, and very, very annoying.

There we go.

The job speak 510, go to the Bluetooth menu.

Very good. Now we're.

Now we're talking.

So, yeah.

Happy New Year 2024.

The plan.

So, the plan for the rest of the semester is to complete fragment four today and tomorrow so I think we're going to get through things this week.

Otherwise we'll take longer.

So, I'm going to essentially cover two kinds of topics.

The first one is called dynamic logic or discourse representation theory.

And then we'll get through with what happens if you have texts that are longer than one sentence.

And it turns out surprisingly that first order logic is actually not very well suited for that application, or higher logic or whatever you want.

So, I'm going to kind of do something new there and get out a couple of features that are

So I completely knew.

Then I would like to get at a completely different way of dealing with semantics. Right now we've kind of been doing very shout very small coverage one sentence to the by then maybe three four sentences.

But what you do if you want to have some context from big corpora.

Think of a million times papers. How do we deal with those kind of things.

That is something we want to show you ways of doing that better not machine learning based.

So not the standard stuff symbolic essentially scaling symbolic methods to larger corpora.

And I think that'll keep us occupied.

If it doesn't.

We can always look at things like believing and propositional attitudes and those kinds of things.

So but today I would like to

go back to fragment for remind you of what we've been doing so that everybody

is in 2000.

Others.

What talk.

Remember

what we're doing with is the method of fragments.

We have kind of this very irregularly shaped blob

of language called English or German or Chinese or whatever.

And the way we're doing this is by defining

regular fragments we call them

via a grammar which actually makes this fragment of English into a formal language.

And the idea is that we kind of want to take

ever increasing fragments that kind of parcel out

the kind of important parts of English so that we actually look at small fragments to study certain