2 min readThe search for new particles: Is AI ready to discover the unknown?

CERN, the worlds’s largest and most respected centers for scientific research is contemplating the use of artificial intelligence to aid in the discovery of new particles.

CERN, the worlds’s largest and most respected centers for scientific research is contemplating the use of artificial intelligence to aid in the discovery of new particles.

Some Important Highlights

  • Due to the huge amount of data being produced at CERN, physicists have been forced to winnow their results in order to manage it.
  • Scientists are now concerned that they are inadvertently deleting evidences for new physics.
  • Physicists are now proposing the use of a deep neural network to for data reduction.
  • Typical deep neural networks will not work in CERN because they are unable to feed the machine the actual parameters of an unknown particle. Instead, the machine will need to seek out variations in data that might point to a new particle.
  • Researchers call this strategy, “bump hunting”, which was the technique used to find the Higgs Boson particle.
  • Machines will begin the search through a strategy called “weakly supervised learning” where it studies how currently known particles have been discovered and use the strategy it learns to find the unknown.
  • Traditional researches require physicists to create an assumption of what a phenomena involving an unknown particle would look like. This process alone would take one year to complete.
  • It will take 20 years for physicists to search for an unknown particle that decays into two known particles of the same or different type.
  • There is currently no strategy which can be used to search for an unknown particle that decays into either two lighter unknown particles of the same type

Why It Matters

CERN’s contribution to human knowledge is staggering. They have discovered what the universe is made of, how it began, and how fundamental particles interact with each other to form the world we now live in.

Using AI to analyze their VAST data speaks of untold discoveries. But if you have been following this blog, they you probably know that AI programs have biases, and AI developers are still grappling with the alignment challenge. How sure are we that the anomaly found by the AI is genuine physics and not just a product of an old frame of mind?

This is a particularly important question to answer for CERN since their institution has the capacity to redefine, not just the sciences, but human life in general.

[contentcards url=”https://www.quantamagazine.org/how-artificial-intelligence-can-supercharge-the-search-for-new-particles-20180723/” target=”_blank”]

Leave a Reply

Your email address will not be published. Required fields are marked *