The following content has been provided by the University of Erlangen-Nürnberg.

So, welcome everyone. This is a special lecture class on a very modern topic, machine learning for physicists, I call it.

This is a topic that I myself have only been learning for the past one and a half years.

It's also a topic in which computer scientists are currently making great progress

and where also physicists are starting to take note and to try to apply it to a particular physical problem.

So the plan for today is I will give you a short introduction about what this is about in general.

And I warn you, I decided to focus on the most promising machine learning technique that exists today, which is neural network,

specifically for deep neural networks, and my lectures are supposed to be very practical and hands-on,

which will mean that until next Thursday you will all install a certain package,

which we will then use together in order to actually program these things.

It turns out in the end it's actually not very difficult and it's much more fun if you can do it yourself.

So there are instructions on the website, I will say a little bit later on this as well.

There are instructions on the website on how to set it up and also next time we will be there to answer any questions

and help you with the installation.

Okay, so regarding the schedule, I fear this is sort of fixed.

I already took a schedule between 6 and 8 so that it will not collide with hopefully any other lectures or tutorials as far as possible,

and otherwise the lecture schedule is fixed by the fact that I myself am travelling every second week,

so nevertheless I try to give you 14 hours of interest.

Yeah, there's a question.

Would it be possible to do the lecture on Monday at 5 already?

No, the reason is that there's the physics colloquium that runs between 4 and 5 and usually runs until 5.30 or so,

and so it's really not possible.

And Thursdays, from like 6 from, I don't know, from 6, but early in the afternoon?

For me it would probably be possible on Thursdays to go down to 5, say,

because sometimes we have this Max Planck Distinguished Lecturer Series which runs between 3 and 4 and then runs until 4.30 or so.

So for me it would be possible, but then it usually also again collides with other lectures.

For me, I don't know someone else's series more than that.

Okay, so I would prefer to keep it because you understand, I mean there are so many people it would be difficult,

but also since there are people who are even abroad but would like to hear a little bit about it,

we are doing some low-level filming, not very high quality, but at least there is something.

I also promised to put the slides online, so this is what I can promise.

Any other questions? I will come back to the schedule and the organization and the homework,

but for now maybe let's just start.

So what is this poem about?

If I take a very mechanistic view of a human, it's someone who receives some input just by listening or by watching,

and then takes certain actions, which is the output, so to speak, speaks, moves and touches and so on.

But the device that really is responsible for all this is of course the brain, as shown here,

and the question is what does the brain consist of, how does this really work on the microscopic level?

And since about 100 years people have begun to understand how that works on the microscopic level,

this is actually a drawing from about 100 years ago by the great Spanish neuroscientist Ramon B. Caral,

who also was a kind of an artist and he drew many of these neurons that you can see if you take brain tissue

and zoom in with a microscope, which they were able to do at the time.

And even at the time they then understood that you have these neurons which are simply cells,

but they are not the usual kind of cell, but they have this branching network of little tubes.

And actually there's always one kind of tube, the so-called axon, which we now know sends a signal from this neuron to other neurons,

and there are the dendrites that receive signals from other axons.

So we really have a network of 100 billion or so neurons in the brain,

and the way the signals are transmitted is actually by electrical pulses,

so if you don't know that, it's very interesting, your brain works on electricity,

and really not just by DC voltages also, but really with electrical pulses.