Applying machine learning to the universe's mysteries

January 30, 2018 by Glenn Roberts Jr., Lawrence Berkeley National Laboratory
The colored lines represent calculated particle tracks from particle collisions occurring within Brookhaven National Laboratory's STAR detector at the Relativistic Heavy Ion Collider, and an illustration of a digital brain. The yellow-red glow at center shows a hydrodynamic simulation of quark-gluon plasma created in particle collisions. Credit: Berkeley Lab

Computers can beat chess champions, simulate star explosions, and forecast global climate. We are even teaching them to be infallible problem-solvers and fast learners.

And now, physicists at the Department of Energy's Lawrence Berkeley National Laboratory (Berkeley Lab) and their collaborators have demonstrated that computers are ready to tackle the universe's greatest mysteries. The team fed thousands of images from simulated high-energy particle collisions to train computer networks to identify important features.

The researchers programmed powerful arrays known as neural networks to serve as a sort of hivelike digital brain in analyzing and interpreting the images of the simulated particle debris left over from the collisions. During this test run the researchers found that the neural networks had up to a 95 percent success rate in recognizing important features in a sampling of about 18,000 images.

The study was published Jan. 15 in the journal Nature Communications.

The next step will be to apply the same machine learning process to actual experimental data.

Powerful machine learning algorithms allow these networks to improve in their analysis as they process more images. The underlying technology is used in facial recognition and other types of image-based object recognition applications.

The images used in this study - relevant to particle-collider nuclear physics experiments at Brookhaven National Laboratory's Relativistic Heavy Ion Collider and CERN's Large Hadron Collider - recreate the conditions of a subatomic particle "soup," which is a superhot fluid state known as the quark-gluon plasma believed to exist just millionths of a second after the birth of the universe. Berkeley Lab physicists participate in experiments at both of these sites.

"We are trying to learn about the most important properties of the quark-gluon plasma," said Xin-Nian Wang, a nuclear physicist in the Nuclear Science Division at Berkeley Lab who is a member of the team. Some of these properties are so short-lived and occur at such tiny scales that they remain shrouded in mystery.

In experiments, nuclear physicists use particle colliders to smash together heavy nuclei, like gold or lead atoms that are stripped of electrons. These collisions are believed to liberate particles inside the atoms' nuclei, forming a fleeting, subatomic-scale fireball that breaks down even protons and neutrons into a free-floating form of their typically bound-up building blocks: quarks and gluons.

Researchers hope that by learning the precise conditions under which this quark-gluon plasma forms, such as how much energy is packed in, and its temperature and pressure as it transitions into a fluid state, they will gain new insights about its component particles of matter and their properties, and about the universe's formative stages.

But exacting measurements of these properties - the so-called "equation of state" involved as matter changes from one phase to another in these collisions - have proven challenging. The initial conditions in the experiments can influence the outcome, so it's challenging to extract equation-of-state measurements that are independent of these conditions.

The diagram at left, which maps out particle distribution in a simulated high-energy heavy-ion collision, includes details on particle momentum and angles. Thousands of these images were used to train and test a neural network to identify important features in the images. At right, a neural network used the collection of images to created this "importance map" - the lighter colors represent areas that are considered more relevant to identify equation of state for the quark-gluon matter created in particle collisions. Credit: Berkeley Lab

"In the nuclear physics community, the holy grail is to see phase transitions in these high-energy interactions, and then determine the equation of state from the experimental data," Wang said. "This is the most important property of the we have yet to learn from experiments."

Researchers also seek insight about the fundamental forces that govern the interactions between quarks and gluons, what physicists refer to as quantum chromodynamics.

Long-Gang Pang, the lead author of the latest study and a Berkeley Lab-affiliated postdoctoral researcher at UC Berkeley, said that in 2016, while he was a postdoctoral fellow at the Frankfurt Institute for Advanced Studies, he became interested in the potential for artificial intelligence (AI) to help solve challenging science problems.

He saw that one form of AI, known as a deep convolutional neural network - with architecture inspired by the image-handling processes in animal brains - appeared to be a good fit for analyzing science-related images.

"These networks can recognize patterns and evaluate board positions and selected movements in the game of Go," Pang said. "We thought, 'If we have some visual scientific data, maybe we can get an abstract concept or valuable physical information from this.'"

Wang added, "With this type of machine learning, we are trying to identify a certain pattern or correlation of patterns that is a unique signature of the equation of state." So after training, the network can pinpoint on its own the portions of and correlations in an image, if any exist, that are most relevant to the problem scientists are trying to solve.

Accumulation of data needed for the analysis can be very computationally intensive, Pang said, and in some cases it took about a full day of computing time to create just one image. When researchers employed an array of GPUs that work in parallel - GPUs are graphics processing units that were first created to enhance video game effects and have since exploded into a variety of uses - they cut that time down to about 20 minutes per image.

They used computing resources at Berkeley Lab's National Energy Research Scientific Computing Center (NERSC) in their study, with most of the computing work focused at GPU clusters at GSI in Germany and Central China Normal University in China.

A benefit of using sophisticated neural networks, the researchers noted, is that they can identify features that weren't even sought in the initial experiment, like finding a needle in a haystack when you weren't even looking for it. And they can extract useful details even from fuzzy images.

"Even if you have low resolution, you can still get some important information," Pang said.

Discussions are already underway to apply the machine learning tools to data from actual heavy-ion collision experiments, and the simulated results should be helpful in training to interpret the real data.

"There will be many applications for this in high-energy particle physics," Wang said, beyond particle-collider experiments.

Explore further: 'Littlest' quark-gluon plasma revealed by physicists using Large Hadron Collider

More information: Long-Gang Pang et al, An equation-of-state-meter of quantum chromodynamics transition from deep learning, Nature Communications (2018). DOI: 10.1038/s41467-017-02726-3

Related Stories

A closer look at the perfect fluid

October 2, 2014

By combining data from two high-energy accelerators, nuclear scientists have refined the measurement of a remarkable property of exotic matter known as quark-gluon plasma. The findings reveal new aspects of the ultra-hot, ...

Recommended for you

16 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

Hyperfuzzy
1 / 5 (2) Jan 30, 2018
Silly Wabbits! Just see it in your head, There only exist the Field and the Field centers. The rest is elementary; if however, computationally massive. Can't see the trees because of the forest! By the way, give up trying to define Theoretical Physics, a Theory begins with a provable Truth! Charge exists! The rest is just nonsense. Reality is not a Fairy tell, there is no beginning no end! An infinite set of states! I'd worry where are the bigger conglomerations and already know every element's spectral response of a given state. Speed, at that level, is only relative to its nearby object! But speed, relative to distant body, shifts in both directions. In an infinite universe nothing has mass, or the Field and Response to Field, there is nothing at the center; it can travel through itself an the right currents.
Hyperfuzzy
1 / 5 (2) Jan 30, 2018
NOTHING PUBLISHED IS DEFINED WITH AT LEAST AXIOM. I CHOOSE CHARGE EXISTS. Charge is the center of its Field. Nothing more. The motion is defined. Coulomb, Maxwell, et. al.; ABSOLUTELY.

Why try to publish a paper based upon reality to this audience. Its not about Truth, its money! Oh yeah, the number of deep thinkers are minimized by the costs of higher education and prestige. That's nonsense! So why can'y mankind live in peace with his environment. You don't have to live with me, not optimal; however, everyone wants what you want. Simple, We are 1!
Hyperfuzzy
1 / 5 (2) Jan 30, 2018
A paper where no paper is needed 'cause, duh.., we know that, since you put it that way! A paper, this is an effort of the entire community. We can manipulate matter! Modulated Field Response, is all that's needed; Its in Design! What band where a signal is delivered: see everything, Medical, ...
Hyperfuzzy
1 / 5 (2) Jan 30, 2018
This method is nuts! Arithmetically I see the focus on high speed collisions; but, think we're smart enough to know what, it is, precisely! You trying to do it within the range of your sensors and Maximum speeds? Why?

If u want a weapon, OK. Go for it. My charges are immortal and infinite, within an infinite space; therefore P(I)again > 0! So let's just create a conscious generation. We have work to do!
Hyperfuzzy
1 / 5 (2) Jan 30, 2018
I know. This World should not exist! The probability of a society designed around a single Thief, Money! It's an absurdity!
Hyperfuzzy
1 / 5 (2) Jan 30, 2018
There u go P(?:q) = 1;i.e. Anything can be predicted set(q) using only q! That is the rules of Maxwell is all that is required and priori knowledge Lambda_Emitted/Measured_Period!
Hyperfuzzy
1 / 5 (4) Jan 30, 2018
Forgive my diatribe, I thought it was important to know the meaning of nonsense!
andyf
5 / 5 (7) Jan 30, 2018
Hyperfuzzy is a demented plonker. See above.

Hyperfuzzy
1 / 5 (3) Jan 30, 2018
Hyperfuzzy is a demented plonker. See above.


There are no particles, only the field centers and the associated wrinkles due to her motion.. The field is a ghost, it only affect's other field centers, this dance is eternal! To live on this planet, in America, as a Black Man, USN 10 yrs, MSEE, all Technology! Hands on. I challenge the above!

They are unique, charge, its field is unique as is every other charge. So superposition of any state may be applied, intellectually speaking. This a set of overlays, its motion's wrinkles are defined by Maxwell and Coulomb, i.e. the initial conditions at any distribution of velocities. But they will not be random, from that moment on.
Hyperfuzzy
1 / 5 (3) Jan 30, 2018
Do we even know our fundamentals? Do I have to have a dissertation upon the subject, why one should be bigger than zero?

Objectively, no, what is one?
Whydening Gyre
5 / 5 (2) Jan 30, 2018
I see we have the HF show happening here....
Merrit
not rated yet Jan 30, 2018
About time they begin using machine learning. Now, they just need to apply it to cosmology.
Hyperfuzzy
not rated yet Jan 30, 2018
About time they begin using machine learning. Now, they just need to apply it to cosmology.

1st, the experimenters must learn, this high power experiments will only end in nuclear disaster, a runaway creation of matter as anti-matter. The occurring oscillations are felt by all neighbors and so on. We die. As it should be. Idiot planet!
Merrit
5 / 5 (3) Jan 31, 2018
With your logic the universe would already be destroyed. The universe produces much higher energy levels in events such as supernova then our punny accelerators.
Hyperfuzzy
not rated yet Feb 01, 2018
With your logic the universe would already be destroyed. The universe produces much higher energy levels in events such as supernova then our punny accelerators.

The universe is infinite, charge always complies, infinite number of indestructible charges centers, only the fields exist; the wrinkles within the field give us light! We are only a part of a vastly changing immortal object!
Hyperfuzzy
not rated yet Feb 05, 2018
With your logic the universe would already be destroyed. The universe produces much higher energy levels in events such as supernova then our punny accelerators.

The universe is infinite, charge always complies, infinite number of indestructible charges centers, only the fields exist; the wrinkles within the field give us light! We are only a part of a vastly changing immortal object!

However, the shimmers are very useful; even more so, in tight conditions!

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.