Okay, so welcome back everybody. So we had an exciting lecture so far, deep learning.

We explored many different applications. So we constructed deeper and more deeper networks

and we were playing games and we could solve so many different things that you may want

to ask your question. So where is the limit? How far can we go? And interestingly, we recently

established a collaboration with neuroscientists and they are very interested in deep learning

because we somehow approach very similar problems. So they look at the brain and they try to

figure out how it works and now they realize that we have these very deep networks and they

kind of work really well. But at some point we also get into this problem that we don't

really understand why they work very well. So at this point we decided, hey, maybe it's

a nice idea to establish a collaboration and this is why I have Patrick Krauss today as

a guest. So he's trained in physics but is now working in neuroscience. And at this point

we also were asking the question, so this is now obviously not a part of your exam preparation,

but I thought it maybe is a very interesting last lecture in a deep learning lecture where

we asked this question, will there ever be consciousness in the machine? Can machines

get conscious? And for that reason I think it's quite nice for us to understand what

is actually consciousness, can we understand what consciousness is and what are typical

problems when you think about consciousness. And this is why I brought Patrick today and

he will give a short introduction into also into the philosophical problems that appear

here. And then later we will see whether this applies to machine learning or whether it

applies not. So I hope you have fun in this lecture. And yeah, Patrick, please come onto

the stage and hand over the microphone. There you go. Okay.

Yeah, so thank you, Andi, for the kind introduction. So as you already said, my name is Patrick

Krauss and I'm working in the University Hospital in Erlangen in the field of neuroscience.

And now I will give you a short overview, a short introduction into neuroscience and

into how the brain works and into the questions that arises when we are challenging the final

question whether will machines ever be consciousness. So first of all, let's start with some

philosophical questions here. So but no worries, I don't want to bore you with the history

of philosophy of mind of the last two and a half thousand years. Instead, I will focus

very briefly on three nowadays thoughts to this question. And when we are dealing with

the question whether a machine is consciousness or not, we have to decide is there actually

a good test to try to ask a machine whether it has conscious feelings or not. And actually,

the only test so far we have is the so-called Turing test, which was proposed by Alan Turing,

a very famous computer scientist and crypto analyst. And he proposed this test in a seminal

paper in the 1960s. And in this paper, he didn't call it of course Turing test, but

instead he called it the imitation game, which you might have already heard about. So there

are many variants of this Turing test, but the basic idea is always the same. So suppose

you have an artificial intelligence system, which you want to test whether it is conscious

or whether it at least has some kind of general intelligence. And then you perform the Turing

test and it goes like this. So you have human participants that have a conversation with

some unknown partner. And this conversation can be either text-based or a chat, for example,

or it can also be in a more sophisticated variant of the Turing test, also be speech-based

of course. And then the human participants have a conversation with different unknown

partners and then they have to decide whether their unknown partner was another human or

a machine actually. And if sufficiently a number of these participants decide then that

the partner which is a machine, which they actually don't know, and they decide that

it must be a human, then the machine passed the Turing test and then we state, okay, this

machine shows some kind of general intelligence or even consciousness. But of course, there

are a lot of criticisms to this test and one famous counter-argument to this test is the

thought experiment proposed by John Searle in the 1980s. So John Searle was a philosopher

dealing mainly with the philosophy of mind. And his counter-argument was the so-called