The scientific team published the results of its “Demonstration of quantum advantage in machine learning” study in the Nature Quantum Information journal, IBM said Wednesday.
Researchers found that a quantum algorithm could identify an unknown string of bits using up to 100 times fewer computational steps than the classical or non-quantum algorithm.
Raytheon BBN programmed a black box designed to produce a string of bits with a hidden pattern for classical and quantum computations.
IBM said the classical computer examined the bits one by one and gave little information on the hidden string while the quantum computer used an algorithm to extract information that a classical algorithm cannot read out.
The algorithm used for the black box experiment can be accessed on IBM’s Quantum Experience website.