International researchers suggest quantum computers boost machine learning
Computer-assisted, or machine learning is a method of artificial intelligence which can search for patterns, repeated in large databases, and use them for self-training during the course of solving various problems. This technology is applied for music and facial recognition, machine-aided translations, medical diagnostics and so on. Due to its ability to adapt to the new data, machine learning by far exceeds human skills, although some tasks are still unsolvable for artificial intelligence (AI).
Using quantum computers offers new unique opportunities which might outperform the best known classical algorithms applied in machine learning. To remove any constraints on machine learning, quantum computers could be used. The basic foundation of quantum computers is not bits elements which can only take one of two values («zero» or «one»), but qubits which can be in a superposition and take both values simultaneously.
As a result, a quantum computer can simultaneously perform many calculations and solve tasks that are beyond the reach of conventional modern computers, for example, model complex physical and chemical processes in pharmaceutics, or create quantum encrypting.
The article, recently published in the prominent research journal Nature, was written by scientists from Skoltech, Castelldefels Institute of Photonic Sciences (Spain), Max-Planck Institute of Quantum Optics (Germany), Waterloo University (Canada), Research Branch of Microsoft, and the Massachusetts Institute of Technology (MIT, USA). The authors laid out the details on the prospects for utilizing quantum computers in machine learning.
According to the scientists, machine learning will not only accelerate the retrieval (search) rate to levels unattainable by conventional computers but it will also do those tasks which cannot be done both by either humans or conventional computers.