Specialists from Sandia National Laboratories have developed an advanced method that makes it possible to determine the probability of failures in the operation of quantum computing systems even before the processes start. The creators themselves compare it to trying to predict whether an old jukebox will play a selected tune before a visitor drops a coin into it. Theoretically, quantum computers have the potential to solve extremely complex problems at speeds unattainable by conventional supercomputers, reports Tech Xplore. However, in practice, the operation of such devices is constantly subject to distortions due to physical inaccuracies, which makes the calculation results imprecise and significantly slows down the progress of the entire technology. At the core of the new scheme is a neural network that studies a digital representation of the intended quantum program. The algorithm accurately predicts which specific physical malfunctions might occur when executing the code on real hardware, and then transforms this information into a numerical formula to assess the chance of a successful outcome for the task. This mechanism is trained on a dataset that includes information on both successful and failed runs, which reduces the need for constant use of expensive equipment. This achievement will help researchers understand in advance which calculations can be effectively carried out with existing capabilities and which are doomed to fail. Scientists are confident that the proposed approach will significantly accelerate the pace of creating the next wave of computing devices. Excluding obviously unpromising research will save significant time and material assets, directing efforts toward the most promising development vectors.
About the Author
