The Big Bang theory is the leading hypothesis for the formation of the universe and everything in it. But when the particles that make up the universe were first generated, they did not have mass.
In 2012, physicists at the European Organization for Nuclear Research — known as CERN and located in Geneva — officially discovered the Higgs boson particle, which explained why some particles today have mass. How Higgs boson particles interact with each other, though, is still an unanswered question — one that may help us understand the energy field that explains “why anything exists at all, really,” wrote Assistant Professor of Physics Loukas Gouskos in an email to The Herald.
Gouskos, an experimental particle physicist at Brown, uses artificial intelligence to study the building blocks of matter, including elementary particles like the Higgs boson. Recently, he received an $850,000 Early Career Award from the Department of Energy to bring his work to CERN’s Large Hadron Collider — the world’s largest particle accelerator, which launches protons together at almost the speed of light.
The work at CERN focuses on “the smallest building blocks of nature, what builds us, what builds our universe (and) how the universe works,” said Greg Landsberg, a professor of physics who has worked at CERN. “We are trying to answer the very fundamental existential questions about where we came from and where we are going.”
Landsberg likened particle accelerators like the LHC to powerful microscopes for biologists, allowing scientists to see very tiny things.
While physicists have studied the Higgs particle for years, how the particle interacts with itself — called self-coupling — remains unclear. Gouskos wrote to The Herald that measuring this interaction would reveal insights about the universe’s stability and its evolution following the Big Bang.
With the DOE grant, Gouskos plans to build machine learning tools that filter out background noise that makes it hard to analyze LHC data.
Lazar Novakovic GS, a physics graduate student on Gouskos’s team, called the LHC the “holy grail of large datasets.” But because so much data is produced by the particle collisions, reading it is an “enormous challenge.”
Background interference makes finding Higgs particle interactions in the LHC a “needle-in-a-haystack problem,” Gouskos wrote, something that could be aided by machine learning tools.
“The challenge is that the process we need to observe, two Higgs bosons produced together, is extraordinarily rare, about a thousand times rarer than producing a single Higgs,” Gouskos wrote. “That’s where AI comes in.”
Though Gouskos hopes to find the evidence of “two Higgs bosons being produced together, which would give us our first real handle on the self-coupling,” he believes a more realistic outcome is a foundation for expanded measurement capabilities at the LHC. The AI tools developed from this project could also translate to fields with similarly “noisy” data structures, such as medical imaging and climate science, he added.
Novakovic explained that as data sets become larger, algorithms must be more robust. “I believe the methods we’re developing will play a key role in the next generation of discoveries at the LHC,” he wrote.
Professor of Physics Antal Jevicki called Gouskos “one of the leaders in introducing AI and machine learning into this data analysis.”
According to Jevicki, the development of the internet was propelled by CERN’s need to communicate over long distances, and he believes that similar innovations could happen with AI at CERN.
“It might be the most promising direction of his work,” he added.
According to Professor of Physics Ulrich Heintz, a new discovery in this field of research “could mean a paradigm shift in our understanding of the universe.”
“It is possible that we are standing on the threshold of a new era in physics,” Heintz added, calling Gouskos “one of the young rising stars in the field.”
Gouskos emphasized that science requires a team of collaborators to produce innovations.
“Particle physics is fundamentally a team sport,” he wrote. “The experiments are too big and the questions too hard for any one group to tackle alone, and a lot of what makes a project like this work is the people you get to think alongside.”
“If we succeed, we’ll have built the tools that make a measurement of one of the most fundamental numbers in physics possible, and learned something about the fate of the universe in the process,” Gouskos added.
Angel Lopez is a senior staff writer covering Science and Research. He’s a first-year student from Tyler, Texas and planning to study neuroscience and literary arts. In his free time, you can find him playing ping pong, listening to music, or reading.
