Theory suggests quantum computers should be exponentially faster on some learning tasks than classical machines

Bob Yirka is a research scientist at Phys.org.

A group of researchers affiliated with multiple institutions in the U.S. and Australia have developed a theory suggesting that quantum computers should be more efficient than classical machines. The group described their theory and results when they were tested on a quantum computer. The idea of combining quantum computing with machine learning to provide a new level of computer-based learning systems was outlined in a Perspective piece by Vedran Dunjko.

A machine learning system makes informed guesses about new data. Quantum computing uses sub-atomic particles to represent qubits as a means for conducting applications many times faster than classical computers. The idea of using quantum computers to run machine-learning applications was considered by the researchers.

To find out if the idea could be done, and if the results would be better than those achieved on classical computers, the researchers devised a machine learning task that would be repeated many times over. They came up with theories about how a quantum system could be used for such experiments. They discovered that a quantum computer could do it better than a classical system. They found a reduction in the amount of experiments needed to learn a concept to be four orders of magnitude lower than for classical systems. Researchers built a system and tested it on a quantum computer to confirm their theory.

If a real-word quantum computer is ever developed, it may be capable of leaning new things on a nearly unimaginable scale.

More information: Hsin-Yuan Huang et al, Quantum advantage in learning from experiments, Science (2022). DOI: 10.1126/science.abn7293

The quantum system is unraveled by quantum learning. There is a science.abp9885

Journal information: Science

There is a science network.