As already told, I would like to speak about a topic which sounds quite novel.

So this is about a new approach to possibly design new computers which are inspired by

some biological aspects.

And in fact, it's a mathematics talk.

So I will mainly speak about the modeling question where we model these kinds of semiconductor

devices by drift-to-fusion equations and then I've explained to you some mathematical tools

and I end by some numerical experiments.

So the motivation is that you know that the computer processors or the basic elements,

the semiconductor transistors are becoming smaller and smaller, but this seems to reach

some physical limit.

So nowadays we have so-called 3 nanometers technology.

So this does not mean that the basic devices are 3 nanometers large.

This is more for commercial reasons.

The effective length of each device is about 48 nanometers.

But 48 nanometers means that these are less than 500 atoms.

And with 500 atomons, I think you believe that we are reaching physical limits.

So the question is what can we do?

So can we invent maybe some computers based on another technology which can outperform

what we have now?

And one idea is so-called neuromorphic computing.

So that means that it's inspired by how the brain is working or the synapses and the connections

between the synapses.

And then the question is can we do this on a semiconductor level?

And the idea is very simple.

So somehow the synapse here, which is here you have a zoom, is replaced by some layered

material with different devices like this one here.

And then the idea is that the signal through a synapse is replaced by the flux through

the semiconductor device.

And either it's going through or it's not going through.

And this is a concept which is not completely new.

So it's already suggested in the 1990s.

And the basic element which is supposed to solve these issues or to have these features

is the so-called memristor, which is just a new word, which means it's a nonlinear resistor

with memory.

So like in a synapse which is storing the information in some sense, it's not every

signal is transferred, but just it sees when a lot of signals are coming then the synapse

is thinking, okay, that's an important thought.

So I should send the signal to the other neurons.

And this is also made in these memristors that you can store some kind of information

in it.

So here's maybe another picture on how these analogies is working.

So you have here some neurons which are connected by an axon.

And at the end of these axons, at the dendrites, you have synapses.

The synapse, in fact, is something which works like an ion channel.

So you have some kind of ions, calcium, calium, and other ions which are sitting here and

which can pass through an ion channel depending on the voltage which is applied to and depending

on the ion.

So each channel only transmits certain kind of ions.

And this may be replaced, so this here is an ion channel, may be replaced by some electric

circuits involving some devices.