The approach taken by Pan and his team from the University of Science and Technology of China overcomes a major hurdle to atom-based quantum computing, according to a paper published last week in the peer-reviewed Physical Review Letters.
The researchers designed an artificial intelligence system capable of arranging more than 2,000 rubidium atoms – each serving as a qubit, the two-state basic unit of quantum computing – into perfect patterns in a mere 60,000th of a second, it said.
The milestone array was hailed by the paper’s reviewers as “a significant leap forward in computational efficiency and experimental feasibility within atom-related quantum physics”, according to a press release on the university’s website.
Three main ways to build a quantum computer have emerged since the concept was first envisioned in the 1980s, with the atom-based approach considered especially promising.
Unlike the alternatives, which use superconducting circuits or trapped ions as qubits, neutral atoms are more stable and easier to control in large numbers. However, atom-based systems have so far been limited to arrays of just a few hundred.
In an atom-based quantum computer, the atoms are held in place by focused laser beams called optical tweezers, which manipulate their energy levels and link them to perform calculations.
