Story perspectives

New Photonic Machine Learning Outshines Traditional Models

10/19/2025

42 5

Summary

Story summary

Researchers from Imperial College London and Durham University developed a continuous-variable photonic extreme learning machine, QELM, to advance data processing in high-energy physics.
It encodes data in light and uses a fixed optical setup for rapid training, enabling top-jet tagging and Higgs boson identification.
Measurements show the QELM outperforms traditional machine learning models while operating at low optical power and room temperature, supporting ultra-fast data selection in future collider experiments.

1 of 1

Story summary

Researchers from Imperial College London and Durham University developed a continuous-variable photonic extreme learning machine, QELM, to advance data processing in high-energy physics.
It encodes data in light and uses a fixed optical setup for rapid training, enabling top-jet tagging and Higgs boson identification.
Measurements show the QELM outperforms traditional machine learning models while operating at low optical power and room temperature, supporting ultra-fast data selection in future collider experiments.