People are made of cells, cells are made  of molecules, molecules are made of atoms,   atoms are made of electrons and proton and neutrons, and those last two are made up of three quarks.

Simple stuff, right?

All except for that last part.

Protons and neutrons are   actually made of many, many quarks that just happen to look like three quarks when you look at them in a   particular way.

And even then, sometimes they’re  made of 5 quarks - including the charm quark The humble proton may seem simple enough,  and they’re certainly common.

Most of the   visible matter in the universe comes from  protons, either as lonely hydrogen nuclei   or bound with neutrons into the nuclei  of the various elements of the periodic   table.

You’d think by now we’d have a pretty  good idea of the properties of the proton.

But actually, their interiors remain a profound  mystery.

It wasn’t until the late 60s that we even   confirmed out that protons were not themselves  elementary particles, but consistent of 3   quarks.

Now, thanks to AI we have evidence that  suggests they may be made of five quarks at least sometime.

Two of them heavier than the proton itself.

To make sense of this we need  to understand how physicists probe the   smallest scales of nature.

And that means  understanding the physics of scattering.

Every time you open your eyes you’re performing  a scattering experiment.

Photons from some light   source like the Sun bounce - or scatter -  off objects in the world around you into   your particle detectors—aka your eyes.

Your  analysis computer—aka your brain—then builds   up a colour map of the world around you  based on the properties of those photons.

Well, it’s also possible to  “see” the subatomic world if   we do a different type of scattering experiment.

In general, the higher the energy  the scattering particle has, the   smaller the object you can see.

To see subatomic  scales, it’s better not to use photons at all,   but instead to use particles of matter.

That's  how Earnest Rutherford first discovered the   nucleus.

He shot a beam of alpha particles  - which we now know to be helium nuclei - at   a thin gold foil.

He found that, while  most particles passed straight through,   a small fraction scattered off something very small and very dense.

That’s how he figured out that the   atom is mostly empty space, with most of its  mass concentrated in a tiny central nucleus.

Since Rutherford’s 1911 experiment we’ve  gotten much better at doing this.

For example,   we now have electron microscopes which  shoot electrons into a sample and measure   how they’re deflected.

Sophisticated software  then reconstructs the structure of the sample,   just as our brain reconstructs  the structure of the world.

Electron scattering can be used to study  the structure of anything larger than an   electron.

That sounds pretty useful, because  electrons have no size—they’re point-like.

Actually not so fast—electrons are quantum  objects and so have a wavelength that gets   shorter the more energy the electron has.

So really I should say that electrons can   be used to study anything that’s larger than the  electron’s wavelength, which depends on energy.

So if you want to study the interior of a  proton you need a pretty high energy electron.

Below a certain energy, the electron will just  bounce off the proton as a whole.

But with enough   energy the electron will punch into a proton and  then scatter off the proton's internal parts.

I should add that such an electron is energetic  enough to destroy the proton.

Which sounds bad,   but it’s actually okay - sometimes you need  to break something to see what it’s made of.

And that’s what physicists at the Stanford Linear  Accelerator Center—SLAC—managed to do in the   50s and 60s.

They accelerated electrons to higher energies than had previously been achieved and   slammed them into protons, then watched what came  out the other side.

In 1968 they published their   results.

From those scattering products and the  deflection of the electrons, the SLAC physicists   deduced that protons must be made of three  point-like particles.

These particles matched   the properties of the so-called quarks—theoretical  components of nucleons proposed by Murry Gell-Mann   and George Zweig earlier that decade.

And that’s  how we get the proton model that you’re probably   familiar with—two up and one down quark, with  three different charge colours bound to each other   by gluons carrying the strong nuclear force,  as we’ve talked about in previous episodes.

And same for the neutron except it's two down and one up.

I'll be talking about protons from now on, but everything I say probably applies to neutrons too.

While the SLAC electron beam was  the most energetic of its time,   we quickly figured out how to make more and  more powerful accelerators.

And as we did so,   we started to see something strange.

Remember I said that higher energy   scattering allows us to see smaller  things and finer resolution.

Well,   as we powered up our accelerators that’s exactly  what we saw—the detailed guts of the proton,   and it turned out to be much more complicated  than just three quarks bound to each other.

The interior of the proton was revealed to be  a complex cluster of energy—a dense network   of gluons that are constantly transforming  into pairs of virtual quarks and antiquarks,   which quickly annihilate each other, turning  back into gluons.

We call this the quark sea,   and this constantly shifting, flickering  mess is a stormy ocean indeed.

But because the quantum properties of  charge, spin, and colour must be conserved,   there is order within the chaos.

If you took all  the contents of the quark sea in one instant,   you could list all the quarks and all the  antiquarks, and see that they cancel each   other out except for two up quarks and one down  quark.

These “valence” quarks will constantly   exchange energy and colour charge via the quark  sea and may even be annihilated by a virtual   particle or antiparticle, but are instantly  replaced by that virtual particle’s partner.

These valence quarks are what the first SLAC  experiments detected.

But as we increased the   electron beam’s energy, those electrons started  to interact with the transient entities of the   quark sea.

They scattered off gluons and virtual  quarks, revealing that complex inner structure.

The higher the energy of the electrons, the  more fine detail we saw.

We saw the virtual   up and down quarks of the quark sea, the gluon  network.

But also some weird stuff.

For example,   around 1% of the time there was evidence of  a charm quark inside the proton.

Well that’s   odd.

The charm quark weighs more than  the entire proton—36% more in fact.

It’s like opening a 1 kg box of I dunno,  apples, and finding a 1.3 kg melon inside.

Fortunately there’s a perfectly plausible  explanation in the case of the proton.

It turns   out that the particles we discover after smashing  an electron into a proton aren’t necessarily   inside the proton to start with.

That’s because  the energy carried by the electron can create   brand new particles.

Remember, E=mc^2, which  tells us that mass and energy can be converted   into each other.

As the ingoing electron transfers  its substantial kinetic energy into the quark sea,   new particle-antiparticle pairs can be created  that were not part of the proton to start with.

As the energy of the electron beam increases,  we get a scattering signal from finer and finer   structure.

But this signal is increasingly muddied  with new particles created in the collision.

We call the particles that are inside the proton  to start with intrinsic particles, and they   include the valence quarks and the virtual quarks  and gluons of the quark sea.

On the other hand,   we call particles created in the collision itself  extrinsic particles.

After a particle collision   we see both intrinsic and extrinsic particles,  as well as the decay products of those particles   in cases where the outgoing particle is too  unstable to reach the detector before falling   apart.

So if the ingoing electron has enough  energy, perhaps it could produce charm quarks   that are detected in some cases.

If the charm  quarks are extrinsic then there’s no problem.

But even early on, there was some evidence  of intrinsic charm quarks.

There was the hint   of the presence of charm quarks even in cases  where the electron didn’t carry enough energy   to produce one.

So if the charm quark is too  massive to be produced by the incoming electron,   and is too massive to be a regular part  of the entire proton, how can it exist?

Well it’s not quite as simple a conundrum as I  make it sound.

As you reduce the energy of   the collision, the probability of finding  an extrinsic charm quark also goes down,   but only reaches zero at zero energy, which  means when no collision is happening.

In these early experiments charm quarks were appearing only slightly more often at low energies than expected from calculations.

and so the evidence for intrinsic charm quarks was weak.

Let’s talk about how those calculations and how they're done to see why this uncertainty exists.

The interior of the proton is the realm of  the strong nuclear force, which is described   by quantum chromodynamics.

With its multiple  different charge types and different charge   carriers, QCD is far more complicated than say the single-charged electrodynamics.

One weird thing   about “QCD” is that the calculations are easier at  high energy than at low energy because it’s easier   to implement a favorite hack of quantum mechanics  called perturbation theory.

That means it’s   easier to calculate the result of a high-energy  scattering event than a low one.

It’s extra hard   to properly calculate the state of the interior  of the proton in the absence of a collision.

For this reason it's kinda straight foward to explain all the extrinsic quarks that we see at high energies, but   it is very difficult to explain the behavior of the  three intrinsic quarks we see at low energies, and   it's even harder to explain why we seem to find  these charmed quark at those same low energies Although it’s hard to model the interior of the  proton at low energy, our intuition still tells   us that it shouldn’t contain things more massive  than the proton itself—like the charm quark.

But there is actually a way to get quantum  chromodynamics to give us intrinsic charm quarks.

We do that by taking advantage of the Heisenberg  uncertainty principle, which, among other things  tells us that we can borrow energy from out of  nowhere to create massive particle-antiparticle,   as long as those particles  vanish again very quickly.

The more massive the particles, the  shorter they’re allowed to live.

For this reason, it should be possible  to generate a charm-anticharm quark pair,   even inside a proton, as long as it  exists only for a very short duration.

Due to the large mass of the charm quark pair,  their momentary appearance is a bigger deal   for the proton’s inner structure than is the  continuous flickering of virtual up and down   quarks.

In that instance of their appearance, the proton looks more like a 5-quark particle than a 3-quark particle.

However I wouldn’t go so far as to say that  the proton briefly more massive.

The   proton’s measured mass comes from the average of  its internal energy over the time of measurement.

So if the charm quarks only exist  for a tiny fraction of the time,   then they contribute only a fraction of  their own hefty mass to the proton’s mass.

I just gave a crude outline of the theory  of intrinsic charm, proposed by   Stanley Brodsky in 1980 to explain the weak evidence of charm quarks in protons.

But proving this   was challenging due to the difficulty of  modeling low-energy particle collisions.

One issue is that there are typically many  different models of the proton interior   that can lead to the prediction of the same  scattering results.

So, you might come up   with a model that includes intrinsic charm  that gives a perfect match for your average   particle output over many collisions.

But you  may have gotten that model right just by chance, and if   you analyzed even more collisions you'd realize  the correct model is actually totally different.

Without testing all possible models, it’s  impossible to know whether a model of intrinsic   charm was really better at predicting the outcome  of these collisions, versus some unknown model   that doesn’t include intrinsic charm.

So  now you have a better idea of why there’s so much uncertainty about the  existence of charm quarks in the proton  —it just wasn’t possible to sift through  all of the possible models to be sure   that intrinsic charm was needed to  explain the scattering experiments.

That is, until Artificial Intelligence  came along.

This new tool allowed   scientists in the NNPDF collaboration to do  something that was previously impossible:   Instead of testing one model, they could test  hundreds or thousands of models of the proton interior at the same time.

They trained a neural network to analyze  nearly 30 years of proton collisions,   not constrained by a single model, but in the  limit of all possible models.

The network was   then able to create new models that approach  the scattering data as closely as possible.

This was incredible.

Instead of a few scientists  trying a few models every couple of years,   they could have this machine test thousands  of models in just a couple of days,   and while a scientists may be biased to trying to  to find evidence in favor of their favorite model,   the network just wants to find the best  answer, even if it contradicts intrinsic charm.

And now we get to the punchline.

This  Neural Network was able to find a model   with intrinsic charm that predicts the  data much better than any previous model.

Now currently this is still tentative.

The team reports a 3-sigma result which means a 1 in 1000 chance they found the wrong model in favor of intrinsic charm just by random chance.

Of course the gold standard for claiming a   victory is 5-sigma, which means a 1 in a million chance the result came from an unlucky streak.

That sounds like a high standard, but  with so many different experiments   happening around the world, we do get a lot  of 3-sigma results that end up being wrong.

So will this one pan out in the end?

Will  we find the charm quark in the proton?

The   melon in the apple box?

Well we’ll need to blast a lot more protons to find out, but we may also   get closer to our answer by improving this shiny  new tool of machine learning.

That technology is   developing at a rapid pace, as it’s increasingly  used across all areas of experimental physics.

Computers can come up with and test models  much faster than any physicist, which means the   physicists can get on with the more interesting  work of interpreting the successful models.

Wherever we land on the question  of intrinsic proton charm,   there’s something charming about  this cooperation between artificial   and natural intelligences working towards the common goal of deciphering the inner workings of space time