Communications of the ACM
A Quantum Trick with Photons Gives ML a Speed Boost
Valeria Saggio and her colleagues at the University of Vienna in Austria have added a quantum twist to accelerate reinforcement learning.
Credit: Just_Super/Getty Images
Researchers at Austria's University of Vienna have accelerated machine learning using a quantum trick involving photons.
The experiment entailed having a photon traveling through a waveguide and ending up in one of four possible states, with artificial intelligence (AI) ensuring the photon ended up in one particular state.
In a classical setup, the AI would only be able to move the photon to one specific state at a time, and would be rewarded for making a correct guess.
For the quantum iteration, the AI could place the photon in a superposition of multiple states, and narrow to the correct answer before making a final, classical guess at the goal state.
The new approach expedited the AI's learning time 63%, from 270 guesses to 100.
From New Scientist
View Full Article
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA
No entries found