Researchers developed RIDDLE, a method that scans particle-collider data for events that don't match the expected background. Instead of guessing what a new particle might look like, it flags oddities on its own — and it performed well on simulated collision events.
Most particle searches only test predictions — they look for something someone already imagined. A method that flags genuinely unexpected events opens the door to discoveries no theory anticipated. Because it leaves the expected background shape largely undistorted, it also reduces the risk of mistaking analysis quirks for a discovery.
✦ The next new particle might be found by software that wasn't told what to look for.