Improve data analysis for the Large Hadron Collider – ScienceDaily

"The methods we have developed greatly enhance our discovery potential for new physics at the LHC," says Kyle Cranmer, professor of physics and lead author of the article in the journal. Letters of physical examination.

Located in the CERN laboratory near Geneva, Switzerland, the LHC is exploring a new frontier of high energy physics and could reveal the origin of the mass of fundamental particles, source of the illusory dark matter that fills the universe, and even additional dimensions of space.

In 2012, data collected by the LHC confirmed the existence of the Higgs boson, a subatomic particle that plays a key role in our understanding of the universe. The following year, Peter Higgs and François Englert received the Nobel Prize in Physics in recognition of their work in developing the theory of what is now called the Higgs Field, which gives the mass of elementary particles.

NYU researchers, including Cranmer, had searched for evidence of the Higgs boson using data collected by the LHC, developed statistical tools, and a methodology to claim the discovery of the new particle.

The new methods described in the Letters of physical examination The paper offers the possibility of making additional discoveries.

"In many scientific fields, the simulations provide the best descriptions of a complex phenomenon, but they are difficult to use in the context of data analysis," says Cranmer, also a faculty member of the Center. for Data Science from NYU. "The techniques we have developed build a bridge allowing us to exploit these very accurate simulations in the context of data analysis."

In physics, this challenge is often daunting.

For example, notes Cranmer, it's easy to simulate the break in a game of billiards with bouncing balls and rails. However, it is much more difficult to look at the final position of the balls to infer how far and at what angle the ball was initially hit.

"Although we often think of pencils and papers or a table full of equations, modern physics often requires detailed computer simulations," he adds. "These simulations can be very accurate, but they do not immediately analyze the data.

"Machine learning excels at finding patterns in data, and this ability can be used to summarize simulated data providing the modern equivalent of a blackboard filled with equations."

The other authors of the paper are: Johann Brehmer, postdoctoral fellow at the New York University Data Science Center, Gilles Louppe, Moore-Sloan data scientist at NYU, and professor at the University of New York Liège and Juan Pavez, a PhD student at the University of Santa Maria in Chile.

"The artificial intelligence revolution is leading to scientific breakthroughs," says Cranmer. "Multidisciplinary teams – like the one that brought together physics, data science, and informatics – contribute to this."

The research was funded in part by grants from the National Science Foundation (ACI-1450310 and PHY-1505463).